1 2013-10-20 Jan Hubicka <jh@suse.cz>
3 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
4 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
5 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
7 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
9 * config/mips/mips.h (ISA_HAS_WSBH): Define.
10 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
12 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
14 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
17 * system.h: Undef m_slot.
19 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
22 all elements for both endian flavors.
24 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
27 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
28 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
29 and SI_REG for 64bit SYSV ABI targets.
31 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
33 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
34 to multi_reg_return. Clarify that we are skipping USEs of multiple
35 return registers. Use bool type where appropriate.
37 2013-10-18 Jan Hubicka <jh@suse.cz>
39 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
41 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
42 (X86_TUNE_PUSH_MEMORY): Likewise.
43 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
44 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
45 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
47 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
48 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
50 (ix86_option_override_internal): Update to use tune features instead
53 2013-10-18 Cong Hou <congh@google.com>
55 PR tree-optimization/58508
56 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
57 statement that contains data refs with zero-step.
59 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
61 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
62 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
63 * gimple-low.c (gimple_check_call_arg,
64 gimple_check_call_matching_types): Move to cgraph.c.
65 * gimple-low.h: Remove prototype.
66 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
67 Relocate from gimple-low.c.
68 * cgraph.h: Add prototype. Don't include basic-block.h.
69 * gimplify.c: Add gimple-low to include list.
70 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
71 * tree-eh.c: Add gimple-low to include list.
72 * tree-nested.c: Likewise.
73 * cfgexpand.c: Add tree-ssa-address.h to include list.
75 * gimple-fold.c: Likewise.
76 * gimple-ssa-strength-reduction.c: Likewise.
77 * trans-mem.c: Likewise.
78 * tree-mudflap.c: Likewise.
79 * tree-ssa-loop-ivopts.c: Likewise.
80 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
81 (degenerate_phi_result): Move to tree-phinodes.c.
82 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
83 * tree-ssa-threadedge.c: Likewise.
84 * tree-vrp.c: Likewise.
85 * tree-phinodes.c (degenerate_phi_result): Relocate here.
86 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
87 * tree-phinodes.h (degenerate_phi_result): Add prototype.
88 * tree-ssa-copy.c: Include tree-ssa-dom.h.
89 * tree-ssa-forwprop.c: Likewise.
90 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
91 pass_data_cleanup_cfg_post_optimizing,
92 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
93 * tree-optimize.c: Delete File.
94 * graphite.c: Include tree-cfgcleanup.h.
96 * tree-cfg.c: Likewise.
97 * tree-profile.c: Likewise.
98 * tree-ssa-dse.c: Likewise.
99 * tree-ssa-loop-ivcanon.c: Likewise.
100 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
101 * tree-outof-ssa.c: Include sbitmap.h.
102 * tree-ssa-live.c: Likewise.
103 * tree-ssa-propagate.c: Likewise.
104 * tree-ssa-structalias.c: Likewise.
105 * tree-stdarg.c: Likewise.
106 * Makefile.in (OBJS): Delete tree-optimize.o.
107 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
108 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
109 * varasm.c: Include basic-block.h.
110 * cfgloop.h: Include function.h instead of basic-block.h
111 (bb_loop_depth): Move to cfgloop.c.
112 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
114 2013-10-18 Teresa Johnson <tejohnson@google.com>
116 * predict.c (probably_never_executed): Compare frequency-based
117 count to number of training runs.
118 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
120 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
122 * config/arm/arm.c (cortexa9_extra_costs): New table.
123 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
125 2013-10-18 Jeff Law <law@redhat.com>
127 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
129 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
130 (register_jump_thread): Add "registered_jump_thread" debug
132 * dbgcnt.def (registered_jump_thread): New debug counter.
134 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
136 * config/rs6000/rs6000.c: Include cgraph.h.
138 2013-10-18 Teresa Johnson <tejohnson@google.com>
140 * tree-ssa-tail-merge.c (replace_block_by): Update edge
141 weights during merging.
143 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
145 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
146 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
147 * tree-ssa-operands.h: Remove prototype.
148 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
149 * gimple.c (virtual_operand_p): Relocate from gimple.c.
150 * gimple.h: Add prototype.
151 * gimple-ssa.h: Include tree-ssa-operands.h.
152 * tree-dump.c: Add tree-cfg.h to include list.
153 * tree-ssa-alias.c: Add ipa-reference.h to include list.
154 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
155 * config/i386/i386.c: Don't include tree-flow.h.
156 * config/rs6000/rs6000.c: Likewise.
158 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
160 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
162 2013-10-18 Richard Biener <rguenther@suse.de>
164 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
165 or TYPE_UNSIGNED of integral types.
166 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
167 NULL_TREE for zero-precision integral types.
169 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
171 * config/aarch64/arm_neon.h
172 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
174 2013-10-17 Sriraman Tallam <tmsriram@google.com>
176 * opth-gen.awk: Define target_flags_explicit.
178 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
180 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
181 fields to the reg_addr array that describes the valid addressing
182 mode for any register, general purpose registers, floating point
183 registers, and Altivec registers.
184 (FIRST_RELOAD_REG_CLASS): Likewise.
185 (LAST_RELOAD_REG_CLASS): Likewise.
186 (struct reload_reg_map_type): Likewise.
187 (reload_reg_map_type): Likewise.
188 (RELOAD_REG_VALID): Likewise.
189 (RELOAD_REG_MULTIPLE): Likewise.
190 (RELOAD_REG_INDEXED): Likewise.
191 (RELOAD_REG_OFFSET): Likewise.
192 (RELOAD_REG_PRE_INCDEC): Likewise.
193 (RELOAD_REG_PRE_MODIFY): Likewise.
194 (reg_addr): Likewise.
195 (mode_supports_pre_incdec_p): New helper functions to say whether
196 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
197 (mode_supports_pre_modify_p): Likewise.
198 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
199 print the valid address mode bits for each mode.
200 (rs6000_debug_print_mode): Likewise.
201 (rs6000_debug_reg_global): Likewise.
202 (rs6000_setup_reg_addr_masks): New function to set up the address
203 mask bits for each type.
204 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
205 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
206 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
207 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
208 PRE_MODIFY are supported.
209 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
210 registers, instead of {src,dest}_av_p.
211 (rs6000_print_options_internal): Tweak the debug output slightly.
213 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
215 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
218 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
220 * tree-flow.h (struct omp_region): Move to omp-low.c.
221 Remove omp_ prototypes and variables.
222 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
223 (copy_var_decl): Relocate prototype from tree-flow.h.
224 * gimple.c (copy_var_decl): Relocate from omp-low.c.
225 * tree.h: Move prototype to omp-low.h.
226 * omp-low.h: New File. Relocate prototypes here.
227 * omp-low.c (struct omp_region): Make local here.
228 (root_omp_region): Make static.
229 (copy_var_decl) Move to gimple.c.
230 (new_omp_region): Make static.
231 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
232 * tree-cfg.c: Include omp-low.h.
233 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
234 * gimplify.c: Include omp-low.h.
235 * tree-parloops.c: Likewise.
237 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
239 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
240 operands are in memory, prefer to force non-matched operand 1 to
243 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
246 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
247 restrict TImode addresses to single indirect registers if both
248 -mquad-memory and -mvsx-timode are used.
249 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
250 we should emit load/store quad. Remove using %y for quad memory
253 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
254 constraints to allow load/store quad on machines where TImode is
255 not allowed in VSX registers. Use 'n' instead of 'F' constraint
256 for TImode to load integer constants.
258 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
260 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
262 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
264 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
265 handling of STACK_REG.
267 2013-10-17 Richard Biener <rguenther@suse.de>
269 PR tree-optimization/58143
270 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
272 (rewrite_to_defined_overflow): Likewise.
273 (move_computations_dom_walker::before_dom): Rewrite stmts
274 with undefined signed overflow that are not always executed
275 into unsigned arithmetic.
277 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
280 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
281 explicit isa flag to be an options variable, instead of using
282 global_options_set. Remove define from rs6000.h.
283 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
285 * config/rs6000/rs6000.c (rs6000_option_override_internal):
286 Initialize rs6000_isa_flags_explicit.
287 (rs6000_function_specific_save): Add gcc_options* parameter, so
288 that the powerpc builds after the 2013-10-15 changes.
289 (rs6000_function_specific_restore): Likewise.
291 2013-10-16 DJ Delorie <dj@redhat.com>
293 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
294 op is a REG before checking REGNO.
295 (rl78_alloc_physical_registers): Verify pattern is a SET before
298 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
300 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
302 (vec_unpacks_lo_v4sf): Likewise.
303 (vec_unpacks_float_hi_v4si): Likewise.
304 (vec_unpacks_float_lo_v4si): Likewise.
305 (vec_unpacku_float_hi_v4si): Likewise.
306 (vec_unpacku_float_lo_v4si): Likewise.
308 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
310 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
311 (vsx_concat_v2sf): Likewise.
313 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
315 * config/aarch64/aarch64.md
316 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
318 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
320 PR tree-optimization/58697
321 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
322 estimated_loop_iterations_int.
323 (max_stmt_executions_int): Call get_max_loop_iterations_int.
324 (get_max_loop_iterations_int): New. HWINT version of
325 get_max_loop_iterations.
326 * cfgloop.h: Add prototypes.
327 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
328 * loop-unroll.c (decide_peel_once_rolling): Call
329 get_estimated_loop_iterations_int.
330 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
331 * tree-ssa-loop-niter.h: Tweak prototypes.
333 2013-10-16 David Malcolm <dmalcolm@redhat.com>
335 * gengtype-parse.c (struct_field_seq): Ignore access-control
336 keywords ("public:" etc).
338 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
340 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
341 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
343 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
345 * config/arm/arm.opt (mlra): New option.
346 * config/arm/arm.c (arm_lra_p): New function.
347 (TARGET_LRA_P): Define.
349 2013-10-16 Paulo Matos <pmatos@broadcom.com>
351 * tree-core.h (tree_code_name): Remove.
352 * tree.h (get_tree_code_name): New prototype.
353 * tree.c (tree_code_name): Make static.
354 (get_tree_code_name): New function.
355 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
356 tree_class_check_failed, tree_range_check_failed,
357 tree_not_class_check_failed, omp_clause_check_failed,
358 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
359 wrapper get_tree_code_name instead of calling tree_code_name directly.
360 * tree-vrp.c (dump_asserts_for): Likewise.
361 * tree-dump.c (dequeue_and_dump): Likewise.
362 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
363 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
364 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
365 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
366 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
367 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
368 dump_gimple_omp_for): Likewise.
369 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
370 * tree-ssa-pre.c (print_pre_expr): Likewise.
371 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
372 * print-tree.c (print_node_brief, print_node): Likewise.
373 * gimple.c (gimple_check_failed): Likewise.
374 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
375 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
376 * config/mep/mep.c (mep_validate_vliw): Likewise.
377 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
378 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
380 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
382 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
385 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
387 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
388 (MULTILIB_MATCHES): Add multilib for -march=v8.
390 2013-10-15 Sriraman Tallam <tmsriram@google.com>
393 * optc-save-gen.awk: Add extra parameter to the save and restore
395 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
396 * tree.c (build_optimization_node): New parameter. Add extra parameter
397 to call to cl_optimization_save.
398 (build_target_option_node): New parameter. Add extra parameter
399 to call to cl_target_option_save.
400 * tree.h (build_optimization_node): New parameter.
401 (build_target_option_node): New parameter.
402 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
403 build_optimization_node and build_target_option_node.
404 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
405 (handle_pragma_push_options): Ditto.
406 * toplev.c (process_options): Ditto.
407 * opts.c (init_options_struct): Check for opts_set non-null.
408 * target.def (target_option.save): New parameter.
409 (target_option.restore): New parameter.
411 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
412 (ix86_pragma_target_parse): Ditto.
413 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
415 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
416 to build_optimization_node and build_target_option_node.
417 (rs6000_valid_attribute_p): Ditto.
418 (rs6000_pragma_target_parse): Ditto.
419 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
421 * config/i386/i386.h:
422 TARGET_64BIT_P: New Macro
423 TARGET_MMX_P: New Macro.
424 TARGET_3DNOW_P: New Macro.
425 TARGET_3DNOW_A_P: New Macro.
426 TARGET_SSE_P: New Macro.
427 TARGET_SSE2_P: New Macro.
428 TARGET_SSE3_P: New Macro.
429 TARGET_SSSE3_P: New Macro.
430 TARGET_SSE4_1_P: New Macro.
431 TARGET_SSE4_2_P: New Macro.
432 TARGET_AVX_P: New Macro.
433 TARGET_AVX2_P: New Macro.
434 TARGET_AVX512F_P: New Macro.
435 TARGET_AVX512PF_P: New Macro.
436 TARGET_AVX512ER_P: New Macro.
437 TARGET_AVX512CD_P: New Macro.
438 TARGET_FMA_P: New Macro.
439 TARGET_SSE4A_P: New Macro.
440 TARGET_FMA4_P: New Macro.
441 TARGET_XOP_P: New Macro.
442 TARGET_LWP_P: New Macro.
443 TARGET_ABM_P: New Macro.
444 TARGET_BMI_P: New Macro.
445 TARGET_BMI2_P: New Macro.
446 TARGET_LZCNT_P: New Macro.
447 TARGET_TBM_P: New Macro.
448 TARGET_POPCNT_P: New Macro.
449 TARGET_SAHF_P: New Macro.
450 TARGET_MOVBE_P: New Macro.
451 TARGET_CRC32_P: New Macro.
452 TARGET_AES_P: New Macro.
453 TARGET_PCLMUL_P: New Macro.
454 TARGET_CMPXCHG16B_P: New Macro.
455 TARGET_FSGSBASE_P: New Macro.
456 TARGET_RDRND_P: New Macro.
457 TARGET_F16C_P: New Macro.
458 TARGET_RTM_P: New Macro.
459 TARGET_HLE_P: New Macro.
460 TARGET_RDSEED_P: New Macro.
461 TARGET_PRFCHW_P: New Macro.
462 TARGET_ADX_P: New Macro.
463 TARGET_FXSR_P: New Macro.
464 TARGET_XSAVE_P: New Macro.
465 TARGET_XSAVEOPT_P: New Macro.
466 TARGET_LP64_P: New Macro.
467 TARGET_X32_P: New Macro.
468 TARGET_FPMATH_DEFAULT_P: New Macro.
469 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
470 * config/i386/i386.c (ix86_option_override_internal): New parameters.
472 Change ix86_tune_string to access opts->x_ix86_tune_string.
473 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
474 Change ix86_arch_string to access opts->x_ix86_arch_string.
475 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
476 Change ix86_pmode to access opts->x_ix86_pmode.
477 Change ix86_abi to access opts->x_ix86_abi.
478 Change ix86_cmodel to access opts->x_ix86_cmodel.
479 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
480 Change ix86_isa_flags_explicit to access
481 opts->x_ix86_isa_flags_explicit.
482 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
483 Change ix86_regparm to access opts->x_ix86_regparm.
484 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
485 Change ix86_preferred_stack_boundary_arg to access
486 opts->x_ix86_preferred_stack_boundary_arg.
487 Change ix86_force_align_arg_pointer to access
488 opts->x_ix86_force_align_arg_pointer.
489 Change ix86_incoming_stack_boundar_arg to access
490 opts->x_ix86_incoming_stack_boundar_arg.
491 Change ix86_fpmath to access opts->x_ix86_fpmath.
492 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
493 Change ix86_recip_name to access opts->x_ix86_recip_name.
494 Change ix86_stack_protector_guard to access
495 opts->x_ix86_stack_protector_guard.
496 Change ix86_tune_memcpy_strategy to access
497 opts->x_ix86_tune_memcpy_strategy.
498 Change ix86_tune_memset_strategy to access
499 opts->x_ix86_tune_memset_strategy.
500 Change global_options to access opts.
501 Change global_options_set to access opts_set.
502 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
503 Change TARGET_MMX to TARGET_MMX_P (opts->...).
504 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
505 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
506 Change TARGET_SSE to TARGET_SSE_P (opts->...).
507 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
508 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
509 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
510 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
511 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
512 Change TARGET_AVX to TARGET_AVX_P (opts->...).
513 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
514 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
515 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
516 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
517 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
518 Change TARGET_FMA to TARGET_FMA_P (opts->...).
519 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
520 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
521 Change TARGET_XOP to TARGET_XOP_P (opts->...).
522 Change TARGET_LWP to TARGET_LWP_P (opts->...).
523 Change TARGET_ABM to TARGET_ABM_P (opts->...).
524 Change TARGET_BMI to TARGET_BMI_P (opts->...).
525 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
526 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
527 Change TARGET_TBM to TARGET_TBM_P (opts->...).
528 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
529 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
530 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
531 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
532 Change TARGET_AES to TARGET_AES_P (opts->...).
533 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
534 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
535 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
536 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
537 Change TARGET_F16C to TARGET_F16C_P (opts->...).
538 Change TARGET_RTM to TARGET_RTM_P (opts->...).
539 Change TARGET_HLE to TARGET_HLE_P (opts->...).
540 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
541 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
542 Change TARGET_ADX to TARGET_ADX_P (opts->...).
543 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
544 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
545 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
546 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
547 Change TARGET_X32 to TARGET_X32_P (opts->...).
548 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
549 Change TARGET_FLOAT_RETURNS_IN_80387 to
550 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
551 (ix86_function_specific_save): New parameter. Use opts-> fields
552 to replace global fields.
553 (ix86_function_specific_restore): Ditto.
554 (ix86_valid_target_attribute_inner_p): New parameters.
556 Fix call to ix86_handle_option and set_option.
557 (ix86_valid_target_attribute_tree): New parameters.
558 Change global_options to access opts.
559 Change global_options_set to access opts_set.
560 Fix call to ix86_valid_target_attribute_inner_p.
561 Change ix86_tune_string to access opts->x_ix86_tune_string.
562 Change ix86_arch_string to access opts->x_ix86_arch_string.
563 Change ix86_fpmath to access opts->x_ix86_fpmath
564 Fix call to ix86_option_override_internal.
565 Fix call to ix86_add_new_builtins.
566 Fix calls to build_optimization_node and build_target_option_node.
567 (ix86_valid_target_attribute_p): Remove access to global_options.
568 Use new gcc_options structure func_options.
569 Fix call to ix86_valid_target_attribute_tree.
570 Fix call to build_optimization_node.
571 (get_builtin_code_for_version): Fix call to
572 ix86_valid_target_attribute_tree.
574 2013-10-15 David Malcolm <dmalcolm@redhat.com>
576 * Makefile.in (PICFLAG): New.
577 (enable_host_shared): New.
578 (INTERNAL_CFLAGS): Use PICFLAG.
579 (LIBIBERTY): Use pic build of libiberty.a if configured with
580 --enable-host-shared.
581 * configure.ac: Add --enable-host-shared, setting up new
583 * configure: Regenerate.
584 * doc/install.texi (--enable-shared): Add note contrasting it with ...
585 (--enable-host-shared): New option.
587 2013-10-15 Richard Biener <rguenther@suse.de>
589 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
590 for built-in functions.
592 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
594 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
595 (optimize_range_tests_1): New function,
596 extracted from optimize_range_tests.
597 (optimize_range_tests_xor): Similarly.
598 (optimize_range_tests_diff): New function.
599 (optimize_range_tests): Use optimize_range_tests_1.
601 2013-10-15 Cong Hou <congh@google.com>
603 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
604 requirement of the reduction pattern so that one operand of the
605 reduction operation can come from outside of the loop.
607 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
609 * config/arm/neon-schedgen.ml: Remove.
610 * config/arm/cortex-a9-neon.md: Remove comment regarding
613 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
615 * config/arm/types: Remove old neon types.
617 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
619 * config/arm/cortex-a7.md
620 (cortex_a7_neon_type): New.
621 (cortex_a7_neon_mul): Update for new types.
622 (cortex_a7_neon_mla): Likewise.
623 (cortex_a7_neon): Likewise.
625 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
627 * config/arm/cortex-a15-neon.md
628 (cortex_a15_neon_type): New,
630 (cortex_a15_neon_int_1): Remove.
631 (cortex_a15_neon_int_2): Likewise.
632 (cortex_a15_neon_int_3): Likewise.
633 (cortex_a15_neon_int_4): Likewise.
634 (cortex_a15_neon_int_5): Likewise.
635 (cortex_a15_neon_vqneg_vqabs): Likewise.
636 (cortex_a15_neon_vmov): Likewise.
637 (cortex_a15_neon_vaba): Likewise.
638 (cortex_a15_neon_vaba_qqq): Likewise.
639 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
640 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
641 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
643 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
644 (cortex_a15_neon_mla_qqq_8_16): Likewise.
645 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
646 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
647 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
648 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
649 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
650 (cortex_a15_neon_shift_1): Likewise.
651 (cortex_a15_neon_shift_2): Likewise.
652 (cortex_a15_neon_shift_3): Likewise.
653 (cortex_a15_neon_vshl_ddd): Likewise.
654 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
655 (cortex_a15_neon_vsra_vrsra): Likewise.
656 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
657 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
658 (cortex_a15_neon_bp_3cycle): Likewise.
659 (cortex_a15_neon_ldm_2): Likewise.
660 (cortex_a15_neon_stm_2): Likewise.
661 (cortex_a15_neon_mcr): Likewise.
662 (cortex_a15_neon_mrc): Likewise.
663 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
664 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
665 (cortex_a15_neon_fp_vmul_ddd): Likewise.
666 (cortex_a15_neon_fp_vmul_qqd): Likewise.
667 (cortex_a15_neon_fp_vmla_ddd): Likewise.
668 (cortex_a15_neon_fp_vmla_qqq): Likewise.
669 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
670 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
671 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
672 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
673 (cortex_a15_neon_bp_simple): Likewise.
674 (cortex_a15_neon_bp_2cycle): Likewise.
675 (cortex_a15_neon_bp_3cycle): Likewise.
676 (cortex_a15_neon_vld1_1_2_regs): Likewise.
677 (cortex_a15_neon_vld1_3_4_regs): Likewise.
678 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
679 (cortex_a15_neon_vld2_4_regs): Likewise.
680 (cortex_a15_neon_vld3_vld4): Likewise.
681 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
682 (cortex_a15_neon_vst1_3_4_regs): Likewise.
683 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
684 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
685 (cortex_a15_neon_vst3_vst4): Rename to...
686 (cortex_a15_neon_vst4): This, update for new attributes.
687 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
688 (cortex_a15_neon_vld3_vld4_lane): Likewise.
689 (cortex_a15_neon_vst1_vst2_lane): Likewise.
690 (cortex_a15_neon_vst3_vst4_lane): Likewise.
691 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
692 (cortex_a15_neon_ldm_2): Likewise.
693 (cortex_a15_neon_stm_2): Likewise.
694 (cortex_a15_neon_mcr): Likewise.
695 (cortex_a15_neon_mcr_2_mcrr): Likewise.
696 (cortex_a15_neon_mrc): Likewise.
697 (cortex_a15_neon_mrrc): Likewise.
699 (cortex_a15_neon_abd): New.
700 (cortex_a15_neon_abd_q): Likewise.
701 (cortex_a15_neon_aba): Likewise.
702 (cortex_a15_neon_aba_q): Likewise.
703 (cortex_a15_neon_acc): Likewise.
704 (cortex_a15_neon_acc_q): Likewise.
705 (cortex_a15_neon_arith_basic): Likewise.
706 (cortex_a15_neon_arith_complex): Likewise.
707 (cortex_a15_neon_multiply): Likewise.
708 (cortex_a15_neon_multiply_q): Likewise.
709 (cortex_a15_neon_mla): Likewise.
710 (cortex_a15_neon_mla_q): Likewise.
711 (cortex_a15_neon_sat_mla_long): Likewise.
712 (cortex_a15_neon_shift_acc): Likewise.
713 (cortex_a15_neon_shift_imm_basic): Likewise.
714 (cortex_a15_neon_shift_imm_complex): Likewise.
715 (cortex_a15_neon_shift_reg_basic): Likewise.
716 (cortex_a15_neon_shift_reg_basic_q): Likewise.
717 (cortex_a15_neon_shift_reg_complex): Likewise.
718 (cortex_a15_neon_shift_reg_complex_q): Likewise.
719 (cortex_a15_neon_fp_negabs): Likewise
720 (cortex_a15_neon_fp_arith): Likewise
721 (cortex_a15_neon_fp_arith_q): Likewise
722 (cortex_a15_neon_fp_cvt_int): Likewise
723 (cortex_a15_neon_fp_cvt_int_q): Likewise
724 (cortex_a15_neon_fp_cvt_16): Likewise
725 (cortex_a15_neon_fp_mul): Likewise
726 (cortex_a15_neon_fp_mul_q): Likewise
727 (cortex_a15_neon_fp_mla): Likewise
728 (cortex_a15_neon_fp_mla_q): Likewise
729 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
730 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
731 (cortex_a15_neon_bitops): Likewise.
732 (cortex_a15_neon_bitops_q): Likewise.
733 (cortex_a15_neon_from_gp): Likewise.
734 (cortex_a15_neon_from_gp_q): Likewise.
735 (cortex_a15_neon_tbl3_tbl4): Likewise.
736 (cortex_a15_neon_zip_q): Likewise.
737 (cortex_a15_neon_to_gp): Likewise.
738 (cortex_a15_neon_load_a): Likewise.
739 (cortex_a15_neon_load_b): Likewise.
740 (cortex_a15_neon_load_c): Likewise.
741 (cortex_a15_neon_load_d): Likewise.
742 (cortex_a15_neon_load_e): Likewise.
743 (cortex_a15_neon_load_f): Likewise.
744 (cortex_a15_neon_store_a): Likewise.
745 (cortex_a15_neon_store_b): Likewise.
746 (cortex_a15_neon_store_c): Likewise.
747 (cortex_a15_neon_store_d): Likewise.
748 (cortex_a15_neon_store_e): Likewise.
749 (cortex_a15_neon_store_f): Likewise.
750 (cortex_a15_neon_store_g): Likewise.
751 (cortex_a15_neon_store_h): Likewise.
752 (cortex_a15_vfp_to_from_gp): Likewise.
754 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
756 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
758 (cortex_a9_neon_vshl_ddd): Remove.
759 (cortex_a9_neon_vst3_vst4): Likewise.
760 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
762 (cortex_a9_neon_bit_ops_q): New.
764 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
765 (cortex_a9_neon_int_2): Likewise.
766 (cortex_a9_neon_int_3): Likewise.
767 (cortex_a9_neon_int_4): Likewise.
768 (cortex_a9_neon_int_5): Likewise.
769 (cortex_a9_neon_vqneg_vqabs): Likewise.
770 (cortex_a9_neon_vmov): Likewise.
771 (cortex_a9_neon_vaba): Likewise.
772 (cortex_a9_neon_vaba_qqq): Likewise.
773 (cortex_a9_neon_shift_1): Likewise.
774 (cortex_a9_neon_shift_2): Likewise.
775 (cortex_a9_neon_shift_3): Likewise.
776 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
777 (cortex_a9_neon_vsra_vrsra): Likewise.
778 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
779 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
780 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
782 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
783 (cortex_a9_neon_mla_qqq_8_16): Likewise.
784 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
786 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
787 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
788 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
789 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
790 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
791 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
792 (cortex_a9_neon_fp_vsum): Likewise.
793 (cortex_a9_neon_fp_vmul_ddd): Likewise.
794 (cortex_a9_neon_fp_vmul_qqd): Likewise.
795 (cortex_a9_neon_fp_vmla_ddd): Likewise.
796 (cortex_a9_neon_fp_vmla_qqq): Likewise.
797 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
798 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
799 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
800 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
801 (cortex_a9_neon_bp_simple): Likewise.
802 (cortex_a9_neon_bp_2cycle): Likewise.
803 (cortex_a9_neon_bp_3cycle): Likewise.
804 (cortex_a9_neon_ldr): Likewise.
805 (cortex_a9_neon_str): Likewise.
806 (cortex_a9_neon_vld1_1_2_regs): Likewise.
807 (cortex_a9_neon_vld1_3_4_regs): Likewise.
808 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
809 (cortex_a9_neon_vld2_4_regs): Likewise.
810 (cortex_a9_neon_vld3_vld4): Likewise.
811 (cortex_a9_neon_vld1_vld2_lane): Likewise.
812 (cortex_a9_neon_vld3_vld4_lane): Likewise.
813 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
814 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
815 (cortex_a9_neon_vst1_3_4_regs): Likewise.
816 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
817 (cortex_a9_neon_vst1_vst2_lane): Likewise.
818 (cortex_a9_neon_vst3_vst4_lane): Likewise.
819 (cortex_a9_neon_mcr): Likewise.
820 (cortex_a9_neon_mcr_2_mcrr): Likewise.
821 (cortex_a9_neon_mrc): Likewise.
822 (cortex_a9_neon_mrrc): Likewise.
824 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
826 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
828 (cortex_a8_neon_vshl_ddd): Remove.
829 (cortex_a8_neon_vst3_vst4): Likewise.
830 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
832 (cortex_a8_neon_bit_ops_q): New.
834 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
835 (cortex_a8_neon_int_2): Likewise..
836 (cortex_a8_neon_int_3): Likewise.
837 (cortex_a8_neon_int_5): Likewise.
838 (cortex_a8_neon_vqneg_vqabs): Likewise.
839 (cortex_a8_neon_int_4): Likewise.
840 (cortex_a8_neon_vaba): Likewise.
841 (cortex_a8_neon_vaba_qqq): Likewise.
842 (cortex_a8_neon_shift_1): Likewise.
843 (cortex_a8_neon_shift_2): Likewise.
844 (cortex_a8_neon_shift_3): Likewise.
845 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
846 (cortex_a8_neon_vsra_vrsra): Likewise.
847 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
848 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
849 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
851 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
852 (cortex_a8_neon_mla_qqq_8_16): Likewise.
853 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
855 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
856 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
857 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
858 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
859 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
860 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
861 (cortex_a8_neon_fp_vsum): Likewise.
862 (cortex_a8_neon_fp_vmul_ddd): Likewise.
863 (cortex_a8_neon_fp_vmul_qqd): Likewise.
864 (cortex_a8_neon_fp_vmla_ddd): Likewise.
865 (cortex_a8_neon_fp_vmla_qqq): Likewise.
866 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
867 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
868 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
869 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
870 (cortex_a8_neon_bp_simple): Likewise.
871 (cortex_a8_neon_bp_2cycle): Likewise.
872 (cortex_a8_neon_bp_3cycle): Likewise.
873 (cortex_a8_neon_ldr): Likewise.
874 (cortex_a8_neon_str): Likewise.
875 (cortex_a8_neon_vld1_1_2_regs): Likewise.
876 (cortex_a8_neon_vld1_3_4_regs): Likewise.
877 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
878 (cortex_a8_neon_vld2_4_regs): Likewise.
879 (cortex_a8_neon_vld3_vld4): Likewise.
880 (cortex_a8_neon_vld1_vld2_lane): Likewise.
881 (cortex_a8_neon_vld3_vld4_lane): Likewise.
882 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
883 (cortex_a8_neon_vst1_3_4_regs): Likewise.
884 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
885 (cortex_a8_neon_vst1_vst2_lane): Likewise.
886 (cortex_a8_neon_vst3_vst4_lane): Likewise.
887 (cortex_a8_neon_mcr): Likewise.
888 (cortex_a8_neon_mcr_2_mcrr): Likewise.
889 (cortex_a8_neon_mrc): Likewise.
890 (cortex_a8_neon_mrrc): Likewise.
892 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
894 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
896 * config/aarch64/aarch64-simd.md (neon_type): Remove.
897 (aarch64_simd_dup<mode>): Add "type" attribute.
898 (aarch64_dup_lane<mode>): Likewise.
899 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
900 (*aarch64_simd_mov<mode>): Likewise.
901 (aarch64_simd_mov_from_<mode>low): Likewise.
902 (aarch64_simd_mov_from_<mode>high): Likewise.
903 (orn<mode>3): Likewise.
904 (bic<mode>3): Likewise.
905 (add<mode>3): Likewise.
906 (sub<mode>3): Likewise.
907 (mul<mode>3): Likewise.
908 (*aarch64_mul3_elt<mode>): Likewise.
909 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
910 (*aarch64_mul3_elt_to_128df): Likewise.
911 (*aarch64_mul3_elt_to_64v2df): Likewise.
912 (neg<mode>2): Likewise.
913 (abs<mode>2): Likewise.
914 (abd<mode>_3): Likewise.
915 (aba<mode>_3): Likewise.
916 (fabd<mode>_3): Likewise.
917 (*fabd_scalar<mode>3): Likewise.
918 (and<mode>3): Likewise.
919 (ior<mode>3): Likewise.
920 (xor<mode>3): Likewise.
921 (one_cmpl<mode>2): Likewise.
922 (aarch64_simd_vec_set<mode>): Likewise.
923 (aarch64_simd_lshr<mode>): Likewise.
924 (aarch64_simd_ashr<mode>): Likewise.
925 (aarch64_simd_imm_shl<mode>): Likewise.
926 (aarch64_simd_reg_sshl<mode): Likewise.
927 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
928 (aarch64_simd_reg_shl<mode>_signed): Likewise.
929 (aarch64_simd_vec_setv2di): Likewise.
930 (aarch64_simd_vec_set<mode>): Likewise.
931 (aarch64_mla<mode>): Likewise.
932 (*aarch64_mla_elt<mode>): Likewise.
933 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
934 (aarch64_mls<mode>): Likewise.
935 (*aarch64_mls_elt<mode>): Likewise.
936 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
937 (<su><maxmin><mode>3): Likewise.
938 (move_lo_quad_<mode>): Likewise.
939 (aarch64_simd_move_hi_quad_<mode>): Likewise.
940 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
941 (vec_pack_trunc_<mode>): Likewise.
942 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
943 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
944 (*aarch64_<su>mlal_lo<mode>): Likewise.
945 (*aarch64_<su>mlal_hi<mode>): Likewise.
946 (*aarch64_<su>mlsl_lo<mode>): Likewise.
947 (*aarch64_<su>mlsl_hi<mode>): Likewise.
948 (*aarch64_<su>mlal<mode>): Likewise.
949 (*aarch64_<su>mlsl<mode>): Likewise.
950 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
951 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
952 (add<mode>3): Likewise.
953 (sub<mode>3): Likewise.
954 (mul<mode>3): Likewise.
955 (div<mode>3): Likewise.
956 (neg<mode>2): Likewise.
957 (abs<mode>2): Likewise.
958 (fma<mode>4): Likewise.
959 (*aarch64_fma4_elt<mode>): Likewise.
960 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
961 (*aarch64_fma4_elt_to_128df): Likewise.
962 (*aarch64_fma4_elt_to_64v2df): Likewise.
963 (fnma<mode>4): Likewise.
964 (*aarch64_fnma4_elt<mode>): Likewise.
965 (*aarch64_fnma4_elt_<vswap_width_name><mode>
966 (*aarch64_fnma4_elt_to_128df): Likewise.
967 (*aarch64_fnma4_elt_to_64v2df): Likewise.
968 (<frint_pattern><mode>2): Likewise.
969 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
970 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
971 (vec_unpacks_lo_v4sf): Likewise.
972 (aarch64_float_extend_lo_v2df): Likewise.
973 (vec_unpacks_hi_v4sf): Likewise.
974 (aarch64_float_truncate_lo_v2sf): Likewise.
975 (aarch64_float_truncate_hi_v4sf): Likewise.
976 (aarch64_vmls<mode>): Likewise.
977 (<su><maxmin><mode>3): Likewise.
978 (<maxmin_uns><mode>3): Likewise.
979 (reduc_<sur>plus_<mode>): Likewise.
980 (reduc_<sur>plus_v2di): Likewise.
981 (reduc_<sur>plus_v2si): Likewise.
982 (reduc_<sur>plus_<mode>): Likewise.
983 (aarch64_addpv4sf): Likewise.
984 (clz<mode>2): Likewise.
985 (reduc_<maxmin_uns>_<mode>): Likewise.
986 (reduc_<maxmin_uns>_v2di): Likewise.
987 (reduc_<maxmin_uns>_v2si): Likewise.
988 (reduc_<maxmin_uns>_<mode>): Likewise.
989 (reduc_<maxmin_uns>_v4sf): Likewise.
990 (aarch64_simd_bsl<mode>_internal): Likewise.
991 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
992 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
993 (aarch64_get_lane<mode>): Likewise.
994 (*aarch64_combinez<mode>): Likewise.
995 (aarch64_combine<mode>): Likewise.
996 (aarch64_simd_combine<mode>): Likewise.
997 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
998 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
999 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
1000 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
1001 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
1002 (aarch64_<sur>h<addsub><mode>): Likewise.
1003 (aarch64_<sur><addsub>hn<mode>): Likewise.
1004 (aarch64_<sur><addsub>hn2<mode>): Likewise.
1005 (aarch64_pmul<mode>): Likewise.
1006 (aarch64_<su_optab><optab><mode>): Likewise.
1007 (aarch64_<sur>qadd<mode>): Likewise.
1008 (aarch64_sqmovun<mode>): Likewise.
1009 (aarch64_<sur>qmovn<mode>): Likewise.
1010 (aarch64_s<optab><mode>): Likewise.
1011 (aarch64_sq<r>dmulh<mode>): Likewise.
1012 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1013 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
1014 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1015 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
1016 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1017 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1018 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
1019 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
1020 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
1021 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
1022 (aarch64_sqdmull<mode>): Likewise.
1023 (aarch64_sqdmull_lane<mode>_internal): Likewise.
1024 (aarch64_sqdmull_n<mode>): Likewise.
1025 (aarch64_sqdmull2<mode>_internal): Likewise.
1026 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
1027 (aarch64_sqdmull2_n<mode>_internal): Likewise.
1028 (aarch64_<sur>shl<mode>): Likewise.
1029 (aarch64_<sur>q<r>shl<mode>
1030 (aarch64_<sur>shll_n<mode>): Likewise.
1031 (aarch64_<sur>shll2_n<mode>): Likewise.
1032 (aarch64_<sur>shr_n<mode>): Likewise.
1033 (aarch64_<sur>sra_n<mode>): Likewise.
1034 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
1035 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
1036 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
1037 (aarch64_cm<optab><mode>): Likewise.
1038 (aarch64_cm<optab>di): Likewise.
1039 (aarch64_cm<optab><mode>): Likewise.
1040 (aarch64_cm<optab>di): Likewise.
1041 (aarch64_cmtst<mode>): Likewise.
1042 (aarch64_cmtstdi): Likewise.
1043 (aarch64_cm<optab><mode>): Likewise.
1044 (*aarch64_fac<optab><mode>): Likewise.
1045 (aarch64_addp<mode>): Likewise.
1046 (aarch64_addpdi): Likewise.
1047 (sqrt<mode>2): Likewise.
1048 (vec_load_lanesoi<mode>): Likewise.
1049 (vec_store_lanesoi<mode>): Likewise.
1050 (vec_load_lanesci<mode>): Likewise.
1051 (vec_store_lanesci<mode>): Likewise.
1052 (vec_load_lanesxi<mode>): Likewise.
1053 (vec_store_lanesxi<mode>): Likewise.
1054 (*aarch64_mov<mode>): Likewise.
1055 (aarch64_ld2<mode>_dreg): Likewise.
1056 (aarch64_ld2<mode>_dreg): Likewise.
1057 (aarch64_ld3<mode>_dreg): Likewise.
1058 (aarch64_ld3<mode>_dreg): Likewise.
1059 (aarch64_ld4<mode>_dreg): Likewise.
1060 (aarch64_ld4<mode>_dreg): Likewise.
1061 (aarch64_tbl1<mode>): Likewise.
1062 (aarch64_tbl2v16qi): Likewise.
1063 (aarch64_combinev16qi): Likewise.
1064 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
1065 (aarch64_st2<mode>_dreg): Likewise.
1066 (aarch64_st2<mode>_dreg): Likewise.
1067 (aarch64_st3<mode>_dreg): Likewise.
1068 (aarch64_st3<mode>_dreg): Likewise.
1069 (aarch64_st4<mode>_dreg): Likewise.
1070 (aarch64_st4<mode>_dreg): Likewise.
1071 (*aarch64_simd_ld1r<mode>): Likewise.
1072 (aarch64_frecpe<mode>): Likewise.
1073 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
1074 (aarch64_frecps<mode>): Likewise.
1076 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1078 * config/arm/iterators.md (V_elem_ch): New.
1080 (VQH_type): Likewise.
1081 * config/arm/arm.md (is_neon_type): New.
1082 (conds): Use is_neon_type.
1083 (anddi3_insn): Update type attribute.
1084 (xordi3_insn): Likewise.
1085 (one_cmpldi2): Likewise.
1086 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
1087 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
1088 (*movmisalign<mode>_neon_store): Likewise.
1089 (*movmisalign<mode>_neon_load): Likewise.
1090 (vec_set<mode>_internal): Likewise.
1091 (vec_set<mode>_internal): Likewise.
1092 (vec_setv2di_internal): Likewise.
1093 (vec_extract<mode>): Likewise.
1094 (vec_extract<mode>): Likewise.
1095 (vec_extractv2di): Likewise.
1096 (*add<mode>3_neon): Likewise.
1097 (adddi3_neon): Likewise.
1098 (*sub<mode>3_neon): Likewise.
1099 (subdi3_neon): Likewise.
1100 (fma<VCVTF:mode>4): Likewise.
1101 (fma<VCVTF:mode>4_intrinsic): Likewise.
1102 (*fmsub<VCVTF:mode>4): Likewise.
1103 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
1104 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
1105 (ior<mode>3): Likewise.
1106 (and<mode>3): Likewise.
1107 (orn<mode>3_neon): Likewise.
1108 (orndi3_neon): Likewise.
1109 (bic<mode>3_neon): Likewise.
1110 (bicdi3_neon): Likewise.
1111 (xor<mode>3): Likewise.
1112 (one_cmpl<mode>2): Likewise.
1113 (abs<mode>2): Likewise.
1114 (neg<mode>2): Likewise.
1115 (negdi2_neon): Likewise.
1116 (*umin<mode>3_neon): Likewise.
1117 (*umax<mode>3_neon): Likewise.
1118 (*smin<mode>3_neon): Likewise.
1119 (*smax<mode>3_neon): Likewise.
1120 (vashl<mode>3): Likewise.
1121 (vashr<mode>3_imm): Likewise.
1122 (vlshr<mode>3_imm): Likewise.
1123 (ashl<mode>3_signed): Likewise.
1124 (ashl<mode>3_unsigned): Likewise.
1125 (neon_load_count): Likewise.
1126 (ashldi3_neon_noclobber): Likewise.
1127 (ashldi3_neon): Likewise.
1128 (signed_shift_di3_neon): Likewise.
1129 (unsigned_shift_di3_neon): Likewise.
1130 (ashrdi3_neon_imm_noclobber): Likewise.
1131 (lshrdi3_neon_imm_noclobber): Likewise.
1132 (<shift>di3_neon): Likewise.
1133 (widen_ssum<mode>3): Likewise.
1134 (widen_usum<mode>3): Likewise.
1135 (quad_halves_<code>v4si): Likewise.
1136 (quad_halves_<code>v4sf): Likewise.
1137 (quad_halves_<code>v8hi): Likewise.
1138 (quad_halves_<code>v16qi): Likewise.
1139 (reduc_splus_v2di): Likewise.
1140 (neon_vpadd_internal<mode>): Likewise.
1141 (neon_vpsmin<mode>): Likewise.
1142 (neon_vpsmax<mode>): Likewise.
1143 (neon_vpumin<mode>): Likewise.
1144 (neon_vpumax<mode>): Likewise.
1145 (*ss_add<mode>_neon): Likewise.
1146 (*us_add<mode>_neon): Likewise.
1147 (*ss_sub<mode>_neon): Likewise.
1148 (*us_sub<mode>_neon): Likewise.
1149 (neon_vadd<mode>_unspec): Likewise.
1150 (neon_vaddl<mode>): Likewise.
1151 (neon_vaddw<mode>): Likewise.
1152 (neon_vhadd<mode>): Likewise.
1153 (neon_vqadd<mode>): Likewise.
1154 (neon_vaddhn<mode>): Likewise.
1155 (neon_vmul<mode>): Likewise.
1156 (neon_vfms<VCVTF:mode>): Likewise.
1157 (neon_vmlal<mode>): Likewise.
1158 (neon_vmls<mode>): Likewise.
1159 (neon_vmlsl<mode>): Likewise.
1160 (neon_vqdmulh<mode>): Likewise.
1161 (neon_vqdmlal<mode>): Likewise.
1162 (neon_vqdmlsl<mode>): Likewise.
1163 (neon_vmull<mode>): Likewise.
1164 (neon_vqdmull<mode>): Likewise.
1165 (neon_vsub<mode>_unspec): Likewise.
1166 (neon_vsubl<mode>): Likewise.
1167 (neon_vsubw<mode>): Likewise.
1168 (neon_vqsub<mode>): Likewise.
1169 (neon_vhsub<mode>): Likewise.
1170 (neon_vsubhn<mode>): Likewise.
1171 (neon_vceq<mode>): Likewise.
1172 (neon_vcge<mode>): Likewise.
1173 (neon_vcgeu<mode>): Likewise.
1174 (neon_vcgt<mode>): Likewise.
1175 (neon_vcgtu<mode>): Likewise.
1176 (neon_vcle<mode>): Likewise.
1177 (neon_vclt<mode>): Likewise.
1178 (neon_vcage<mode>): Likewise.
1179 (neon_vcagt<mode>): Likewise.
1180 (neon_vtst<mode>): Likewise.
1181 (neon_vabd<mode>): Likewise.
1182 (neon_vabdl<mode>): Likewise.
1183 (neon_vaba<mode>): Likewise.
1184 (neon_vabal<mode>): Likewise.
1185 (neon_vmax<mode>): Likewise.
1186 (neon_vmin<mode>): Likewise.
1187 (neon_vpaddl<mode>): Likewise.
1188 (neon_vpadal<mode>): Likewise.
1189 (neon_vpmax<mode>): Likewise.
1190 (neon_vpmin<mode>): Likewise.
1191 (neon_vrecps<mode>): Likewise.
1192 (neon_vrsqrts<mode>): Likewise.
1193 (neon_vqabs<mode>): Likewise.
1194 (neon_vqneg<mode>): Likewise.
1195 (neon_vcls<mode>): Likewise.
1196 (clz<mode>2): Likewise.
1197 (popcount<mode>2): Likewise.
1198 (neon_vrecpe<mode>): Likewise.
1199 (neon_vrsqrte<mode>): Likewise.
1200 (neon_vget_lane<mode>_sext_internal): Likewise.
1201 (neon_vget_lane<mode>_zext_internal): Likewise.
1202 (neon_vdup_n<mode>): Likewise.
1203 (neon_vdup_n<mode>): Likewise.
1204 (neon_vdup_nv2di): Likewise.
1205 (neon_vdup_lane<mode>_interal): Likewise.
1206 (*neon_vswp<mode>): Likewise.
1207 (neon_vcombine<mode>): Likewise.
1208 (float<mode><V_cvtto>2): Likewise.
1209 (floatuns<mode><V_cvtto>2): Likewise.
1210 (fix_trunc<mode><V_cvtto>2): Likewise.
1211 (fixuns_trunc<mode><V_cvtto>2
1212 (neon_vcvt<mode>): Likewise.
1213 (neon_vcvt<mode>): Likewise.
1214 (neon_vcvtv4sfv4hf): Likewise.
1215 (neon_vcvtv4hfv4sf): Likewise.
1216 (neon_vcvt_n<mode>): Likewise.
1217 (neon_vcvt_n<mode>): Likewise.
1218 (neon_vmovn<mode>): Likewise.
1219 (neon_vqmovn<mode>): Likewise.
1220 (neon_vqmovun<mode>): Likewise.
1221 (neon_vmovl<mode>): Likewise.
1222 (neon_vmul_lane<mode>): Likewise.
1223 (neon_vmul_lane<mode>): Likewise.
1224 (neon_vmull_lane<mode>): Likewise.
1225 (neon_vqdmull_lane<mode>): Likewise.
1226 (neon_vqdmulh_lane<mode>): Likewise.
1227 (neon_vqdmulh_lane<mode>): Likewise.
1228 (neon_vmla_lane<mode>): Likewise.
1229 (neon_vmla_lane<mode>): Likewise.
1230 (neon_vmlal_lane<mode>): Likewise.
1231 (neon_vqdmlal_lane<mode>): Likewise.
1232 (neon_vmls_lane<mode>): Likewise.
1233 (neon_vmls_lane<mode>): Likewise.
1234 (neon_vmlsl_lane<mode>): Likewise.
1235 (neon_vqdmlsl_lane<mode>): Likewise.
1236 (neon_vext<mode>): Likewise.
1237 (neon_vrev64<mode>): Likewise.
1238 (neon_vrev32<mode>): Likewise.
1239 (neon_vrev16<mode>): Likewise.
1240 (neon_vbsl<mode>_internal): Likewise.
1241 (neon_vshl<mode>): Likewise.
1242 (neon_vqshl<mode>): Likewise.
1243 (neon_vshr_n<mode>): Likewise.
1244 (neon_vshrn_n<mode>): Likewise.
1245 (neon_vqshrn_n<mode>): Likewise.
1246 (neon_vqshrun_n<mode>): Likewise.
1247 (neon_vshl_n<mode>): Likewise.
1248 (neon_vqshl_n<mode>): Likewise.
1249 (neon_vqshlu_n<mode>): Likewise.
1250 (neon_vshll_n<mode>): Likewise.
1251 (neon_vsra_n<mode>): Likewise.
1252 (neon_vsri_n<mode>): Likewise.
1253 (neon_vsli_n<mode>): Likewise.
1254 (neon_vtbl1v8qi): Likewise.
1255 (neon_vtbl2v8qi): Likewise.
1256 (neon_vtbl3v8qi): Likewise.
1257 (neon_vtbl4v8qi): Likewise.
1258 (neon_vtbl1v16qi): Likewise.
1259 (neon_vtbl2v16qi): Likewise.
1260 (neon_vcombinev16qi): Likewise.
1261 (neon_vtbx1v8qi): Likewise.
1262 (neon_vtbx2v8qi): Likewise.
1263 (neon_vtbx3v8qi): Likewise.
1264 (neon_vtbx4v8qi): Likewise.
1265 (*neon_vtrn<mode>_insn): Likewise.
1266 (*neon_vzip<mode>_insn): Likewise.
1267 (*neon_vuzp<mode>_insn): Likewise.
1268 (neon_vld1<mode>): Likewise.
1269 (neon_vld1_lane<mode>): Likewise.
1270 (neon_vld1_lane<mode>): Likewise.
1271 (neon_vld1_dup<mode>): Likewise.
1272 (neon_vld1_dup<mode>): Likewise.
1273 (neon_vld1_dupv2di): Likewise.
1274 (neon_vst1<mode>): Likewise.
1275 (neon_vst1_lane<mode>): Likewise.
1276 (neon_vst1_lane<mode>): Likewise.
1277 (neon_vld2<mode>): Likewise.
1278 (neon_vld2<mode>): Likewise.
1279 (neon_vld2_lane<mode>): Likewise.
1280 (neon_vld2_lane<mode>): Likewise.
1281 (neon_vld2_dup<mode>): Likewise.
1282 (neon_vst2<mode>): Likewise.
1283 (neon_vst2<mode>): Likewise.
1284 (neon_vst2_lane<mode>): Likewise.
1285 (neon_vst2_lane<mode>): Likewise.
1286 (neon_vld3<mode>): Likewise.
1287 (neon_vld3qa<mode>): Likewise.
1288 (neon_vld3qb<mode>): Likewise.
1289 (neon_vld3_lane<mode>): Likewise.
1290 (neon_vld3_lane<mode>): Likewise.
1291 (neon_vld3_dup<mode>): Likewise.
1292 (neon_vst3<mode>): Likewise.
1293 (neon_vst3qa<mode>): Likewise.
1294 (neon_vst3qb<mode>): Likewise.
1295 (neon_vst3_lane<mode>): Likewise.
1296 (neon_vst3_lane<mode>): Likewise.
1297 (neon_vld4<mode>): Likewise.
1298 (neon_vld4qa<mode>): Likewise.
1299 (neon_vld4qb<mode>): Likewise.
1300 (neon_vld4_lane<mode>): Likewise.
1301 (neon_vld4_lane<mode>): Likewise.
1302 (neon_vld4_dup<mode>): Likewise.
1303 (neon_vst4<mode>): Likewise.
1304 (neon_vst4qa<mode>): Likewise.
1305 (neon_vst4qb<mode>): Likewise.
1306 (neon_vst4_lane<mode>): Likewise.
1307 (neon_vst4_lane<mode>): Likewise.
1308 (neon_vec_unpack<US>_lo_<mode>): Likewise.
1309 (neon_vec_unpack<US>_hi_<mode>): Likewise.
1310 (neon_vec_<US>mult_lo_<mode>): Likewise.
1311 (neon_vec_<US>mult_hi_<mode>): Likewise.
1312 (neon_vec_<US>shiftl_<mode>): Likewise.
1313 (neon_unpack<US>_<mode>): Likewise.
1314 (neon_vec_<US>mult_<mode>): Likewise.
1315 (vec_pack_trunc_<mode>): Likewise.
1316 (neon_vec_pack_trunc_<mode>): Likewise.
1317 (neon_vabd<mode>_2): Likewise.
1318 (neon_vabd<mode>_3): Likewise.
1320 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1322 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
1323 (load_pair): Update type attribute.
1324 (store_pair): Update type attribute.
1325 * config/aarch64/iterators.md (q): New.
1327 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1329 * config/arm/types.md: Add new types for Neon insns.
1331 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1332 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1333 Sergey Lega <sergey.s.lega@intel.com>
1334 Anna Tikhonova <anna.tikhonova@intel.com>
1335 Ilya Tocar <ilya.tocar@intel.com>
1336 Andrey Turetskiy <andrey.turetskiy@intel.com>
1337 Ilya Verbin <ilya.verbin@intel.com>
1338 Kirill Yukhin <kirill.yukhin@intel.com>
1339 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1341 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
1342 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
1343 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
1345 (srcp14<mode>): Ditto.
1346 (rsqrt14<mode>): Ditto.
1347 (rsqrt14<mode>): Ditto.
1348 (avx512f_vmscalef<mode>): Ditto.
1349 (avx512f_scalef<mode>): Ditto.
1350 (avx512f_getexp<mode>): Ditto.
1351 (avx512f_sgetexp<mode>): Ditto.
1352 (avx512f_fixupimm<mode>): Ditto.
1353 (avx512f_sfixupimm<mode>): Ditto.
1354 (avx512f_rndscale<mode>): Ditto.
1355 (*avx512er_exp2<mode>): Ditto.
1356 (*avx512er_rcp28<mode>): Ditto.
1357 (avx512er_rsqrt28<mode>): Ditto.
1358 (avx512f_getmant<mode>): Ditto.
1359 (avx512f_getmant<mode>): Ditto.
1360 (avx512f_rndscale<mode>): Fix formatting.
1362 2013-10-15 Martin Jambor <mjambor@suse.cz>
1364 * ipa-utils.h (ipa_edge_within_scc): Declare.
1365 * ipa-cp.c (edge_within_scc): Moved...
1366 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
1368 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1369 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1370 Sergey Lega <sergey.s.lega@intel.com>
1371 Anna Tikhonova <anna.tikhonova@intel.com>
1372 Ilya Tocar <ilya.tocar@intel.com>
1373 Andrey Turetskiy <andrey.turetskiy@intel.com>
1374 Ilya Verbin <ilya.verbin@intel.com>
1375 Kirill Yukhin <kirill.yukhin@intel.com>
1376 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1378 * config/i386/predicates.md (const_8_to_15_operand): New.
1379 (const_16_to_31_operand): Ditto.
1380 * config/i386/sse.md (V8FI): New.
1382 (reduc_splus_v8df): Ditto.
1383 (reduc_splus_v16sf): Ditto.
1384 (avx512f_vextract<shuffletype>32x4_1): Ditto.
1385 (vec_extract_hi_<mode>): Ditto.
1386 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
1387 (vec_set_lo_<mode>): Ditto.
1388 (vec_set_hi_<mode>): Ditto.
1389 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
1390 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
1391 (avx512f_pshufd_1): Ditto.
1392 (avx512f_broadcast<mode>): Ditto.
1393 (avx512f_broadcast<mode>): Ditto.
1394 (define_split): Split vec_extract_lo into move.
1395 (ssequartermode): Ditto.
1396 (ssedoublemode): Extened with wider modes.
1397 (vec_extract_lo_<mode>): Ditto.
1399 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1400 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1401 Sergey Lega <sergey.s.lega@intel.com>
1402 Anna Tikhonova <anna.tikhonova@intel.com>
1403 Ilya Tocar <ilya.tocar@intel.com>
1404 Andrey Turetskiy <andrey.turetskiy@intel.com>
1405 Ilya Verbin <ilya.verbin@intel.com>
1406 Kirill Yukhin <kirill.yukhin@intel.com>
1407 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1409 * config/i386/predicates.md (register_or_constm1_operand): New.
1410 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
1411 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
1412 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
1413 UNSPEC_SCATTER_PREFETCH
1415 (avx512f_ucmp<mode>3): Ditto.
1416 (avx512f_vternlog<mode>): Ditto.
1417 (avx512f_align<mode>): Ditto.
1418 (<shift_insn><mode>3): Ditto.
1419 (avx512f_<rotate>v<mode>): Ditto.
1420 (avx512f_<rotate><mode>): Ditto.
1421 (avx512f_eq<mode>3): Ditto.
1422 (avx512f_eq<mode>3_1): Ditto.
1423 (avx512f_gt<mode>3): Ditto.
1424 (avx512f_testm<mode>3): Ditto.
1425 (avx512f_testnm<mode>3): Ditto.
1426 (avx512pf_gatherpf<mode>): Ditto.
1427 (*avx512pf_gatherpf<mode>_mask): Ditto.
1428 (*avx512pf_gatherpf<mode>): Ditto.
1429 (avx512pf_scatterpf<mode>): Ditto.
1430 (*avx512pf_scatterpf<mode>_mask): Ditto.
1431 (*avx512pf_scatterpf<mode>): Ditto.
1432 (avx512f_vec_dup_gpr<mode>): Ditto.
1433 (clz<mode>2): Ditto.
1434 (conflict<mode>): Ditto.
1435 (REDUC_SMINMAX_MODE): Extened with wider modes.
1436 (reduc_<code>_<mode>): Ditto.
1437 (vlshr<mode>3): Ditto.
1438 (vashl<mode>3): Ditto.
1440 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1441 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1442 Sergey Lega <sergey.s.lega@intel.com>
1443 Anna Tikhonova <anna.tikhonova@intel.com>
1444 Ilya Tocar <ilya.tocar@intel.com>
1445 Andrey Turetskiy <andrey.turetskiy@intel.com>
1446 Ilya Verbin <ilya.verbin@intel.com>
1447 Kirill Yukhin <kirill.yukhin@intel.com>
1448 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1450 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
1453 (avx512fmaskmode): Ditto.
1454 (bcstscalarsuff): Ditto.
1455 (avx512f_blendm<mode>): Ditto.
1456 (cmp_imm_predicate): Ditto.
1457 (avx512f_cmp<mode>3): Ditto.
1458 (avx512f_vec_dup<mode>): Ditto.
1459 (avx512f_vec_dup_mem<mode>): Ditto.
1460 (avx512f_vpermi2var<mode>3): Ditto.
1461 (avx512f_vpermt2var<mode>3): Ditto.
1462 (vec_init<mode>): Ditto.
1463 (avx512f_gathersi<mode>): Ditto.
1464 (*avx512f_gathersi<mode>): Ditto.
1465 (*avx512f_gathersi<mode>_2): Ditto.
1466 (avx512f_gatherdi<mode>): Ditto.
1467 (*avx512f_gatherdi<mode>): Ditto.
1468 (*avx512f_gatherdi<mode>_2): Ditto.
1469 (avx512f_scattersi<mode>): Ditto.
1470 (*avx512f_scattersi<mode>): Ditto.
1471 (avx512f_scatterdi<mode>): Ditto.
1472 (*avx512f_scatterdi<mode>): Ditto.
1473 (sseintprefix): Extened with wider modes.
1474 (VEC_GATHER_IDXSI): Ditto.
1475 (VEC_GATHER_IDXDI): Ditto.
1476 (VEC_GATHER_SRCDI): Ditto.
1478 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
1479 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1481 * config/arm/t-aprofile: New file.
1482 * config.gcc: Handle --with-multilib-list option.
1484 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
1486 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
1489 2013-10-15 Richard Biener <rguenther@suse.de>
1491 * tree-loop-distribution.c (build_empty_rdg): Inline into
1493 (rdg_flag_vertex): Inline into single user.
1494 (rdg_flag_vertex_and_dependent): Likewise.
1495 (build_rdg_partition_for_vertex): Remove processed bitmap.
1496 (rdg_build_partitions): Simplify.
1498 2013-10-15 Richard Biener <rguenther@suse.de>
1500 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
1501 Restructure forwarding through conversions and copies to
1502 avoid performing copy-propagation the wrong way. Adjust
1503 recursion invocations.
1504 (forward_propagate_addr_expr): Add argument stating if we
1505 are recursing from a single-use.
1506 (ssa_forward_propagate_and_combine): Adjust.
1508 2013-10-14 David Malcolm <dmalcolm@redhat.com>
1510 * dumpfile.h (gcc::dump_manager): New class, to hold state
1511 relating to dumpfile management.
1512 (get_dump_file_name): Remove in favor of method of dump_manager.
1513 (dump_initialized_p): Likewise.
1514 (dump_start): Likewise.
1515 (dump_finish): Likewise.
1516 (dump_switch_p): Likewise.
1517 (dump_register): Likewise.
1518 (get_dump_file_info): Likewise.
1519 * context.c (gcc::context::context): Construct the dump_manager
1521 * context.h (gcc::context::get_dumps): New.
1522 (gcc::context::m_dumps): New.
1523 * coverage.c (coverage_init): Port to dump_manager API.
1524 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
1525 (extra_dump_files_in_use): Likewise.
1526 (extra_dump_files_alloced): Likewise.
1527 (gcc::dump_manager::dump_manager): New.
1528 (dump_register): Convert to...
1529 (gcc::dump_manager::dump_register): ...method, replacing
1530 function-static next_dump with m_next_dump field.
1531 (get_dump_file_info): Convert to...
1532 (gcc::dump_manager::get_dump_file_info): ...method.
1533 (get_dump_file_name): Convert to...
1534 (gcc::dump_manager::get_dump_file_name): ...method.
1535 (dump_start): Convert to...
1536 (gcc::dump_manager::dump_start): ...method.
1537 (dump_finish): Convert to...
1538 (gcc::dump_manager::dump_finish): ...method.
1539 (dump_begin): Replace body with...
1540 (gcc::dump_manager::dump_begin): ...new method.
1541 (dump_phase_enabled_p): Convert to...
1542 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1543 (dump_phase_enabled_p): Convert to...
1544 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1545 (dump_initialized_p): Convert to...
1546 (gcc::dump_manager::dump_initialized_p): ...method.
1547 (dump_flag_name): Replace body with...
1548 (gcc::dump_manager::dump_flag_name): ...new method.
1549 (dump_enable_all): Convert to...
1550 (gcc::dump_manager::dump_enable_all): ...new method.
1551 (opt_info_enable_passes): Convert to...
1552 (gcc::dump_manager::opt_info_enable_passes): ...new method.
1553 (dump_switch_p_1): Convert to...
1554 (gcc::dump_manager::dump_switch_p_1): ...new method.
1555 (dump_switch_p): Convert to...
1556 (gcc::dump_manager::dump_switch_p): ...new method.
1557 (opt_info_switch_p): Port to dump_manager API.
1558 (enable_rtl_dump_file): Likewise.
1559 * opts-global.c (handle_common_deferred_options): Port to new
1561 * passes.c (pass_manager::finish_optimization_passes): Likewise.
1562 (pass_manager::register_one_dump_file): Likewise.
1563 (pass_manager::register_pass): Likewise.
1564 (pass_init_dump_file): Likewise.
1565 (pass_fini_dump_file): Likewise.
1566 * statistics.c (statistics_early_init): Likewise.
1568 2013-10-14 Richard Biener <rguenther@suse.de>
1570 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
1571 iterative_hash_canonical_type, gimple_canonical_type_hash,
1572 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
1573 gimple_register_canonical_type, print_gimple_types_stats,
1574 free_gimple_type_tables): Move to lto/lto.c
1575 (gt-gimple.h): Do not include.
1576 * gimple.h (gimple_register_canonical_type,
1577 print_gimple_types_stats, free_gimple_type_tables): Remove.
1578 * Makefile.in (GTFILES): Remove gimple.c.
1580 2013-10-14 Travis Snoozy <quandary@remstate.com>
1583 * config/msp430/msp430.c (msp430_option_override): Correct thinko
1584 scanning for msp430x targets.
1586 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
1589 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
1590 (registers_ok_for_ldd_peep): Move around.
1591 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
1592 * config/sparc/sparc.md (widening peepholes): Use it.
1594 2013-10-14 Richard Biener <rguenther@suse.de>
1598 * gimple.c (iterative_hash_canonical_type): Make sure to
1599 record the hash into the correct hashtable slot.
1601 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
1603 PR rtl-optimization/58662
1604 * combine.c (try_combine): Take into account death nodes on I2 when
1605 splitting a PARALLEL of two independent SETs. Fix dump message.
1607 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
1610 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
1611 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
1612 * config/sh/t-sh (sh_treg_combine.o): New entry.
1613 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
1614 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
1615 (register_sh_passes): New function. Register sh_treg_combine pass.
1616 (sh_option_override): Invoke it.
1617 (sh_canonicalize_comparison): Handle op0_preserve_value.
1618 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
1619 opportunities. Canonicalize branch condition.
1620 (nott): Allow only if pseudos can be created for non-SH2A.
1622 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
1625 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
1626 (ix86_expand_movmem): Replace copy_addr_to_reg with
1627 ix86_copy_addr_to_reg.
1628 (ix86_expand_setmem): Likewise.
1630 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
1632 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
1633 provided by ix86_fp_compare_mode instead of CCFPUmode.
1635 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
1637 * config/aarch64/arm_neon.h
1638 (vtbx<1,3>_<psu>8): Fix register constriants.
1640 2013-10-11 Jeff Law <law@redhat.com>
1642 PR tree-optimization/58640
1643 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
1644 threading paths that cross over two loop entry points.
1646 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1648 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
1649 handle vector float as well.
1650 (*vsx_le_perm_load_v4si): Likewise.
1651 (*vsx_le_perm_store_v2di): Likewise.
1652 (*vsx_le_perm_store_v4si): Likewise.
1654 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1656 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
1657 directly to circumvent subtract from splat{31} workaround.
1658 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
1660 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
1661 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
1662 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
1663 (altivec_vperm_<mode>): Convert to define_insn_and_split to
1664 separate big and little endian logic.
1665 (*altivec_vperm_<mode>_internal): New define_insn.
1666 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
1667 separate big and little endian logic.
1668 (*altivec_vperm_<mode>_uns_internal): New define_insn.
1669 (vec_permv16qi): Add little endian logic.
1671 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
1673 * doc/extend.texi (returns_nonnull): Remove arguments.
1675 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1676 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1677 Sergey Lega <sergey.s.lega@intel.com>
1678 Anna Tikhonova <anna.tikhonova@intel.com>
1679 Ilya Tocar <ilya.tocar@intel.com>
1680 Andrey Turetskiy <andrey.turetskiy@intel.com>
1681 Ilya Verbin <ilya.verbin@intel.com>
1682 Kirill Yukhin <kirill.yukhin@intel.com>
1683 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1685 * config/i386/sse.md (VI48F_256_512): New.
1686 (avx2_permvar<mode>): Change to ...
1687 (<avx2_avx512f>_permvar<mode>): This.
1689 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1690 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1691 Sergey Lega <sergey.s.lega@intel.com>
1692 Anna Tikhonova <anna.tikhonova@intel.com>
1693 Ilya Tocar <ilya.tocar@intel.com>
1694 Andrey Turetskiy <andrey.turetskiy@intel.com>
1695 Ilya Verbin <ilya.verbin@intel.com>
1696 Kirill Yukhin <kirill.yukhin@intel.com>
1697 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1699 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
1700 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
1701 * config/i386/sse.md (VI4_AVX): New.
1702 (sf2simodelower): Ditto.
1703 (sse2_cvtps2dq): Change to ...
1704 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
1706 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1707 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1708 Sergey Lega <sergey.s.lega@intel.com>
1709 Anna Tikhonova <anna.tikhonova@intel.com>
1710 Ilya Tocar <ilya.tocar@intel.com>
1711 Andrey Turetskiy <andrey.turetskiy@intel.com>
1712 Ilya Verbin <ilya.verbin@intel.com>
1713 Kirill Yukhin <kirill.yukhin@intel.com>
1714 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1716 * config/i386/sse.md (V_512): New.
1718 (vcond<V_512:mode><VF_512:mode>): Ditto.
1719 (vcond<V_512:mode><VI_512:mode>): Ditto.
1720 (vcondu<V_512:mode><VI_512:mode>): Ditto.
1722 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1723 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1724 Sergey Lega <sergey.s.lega@intel.com>
1725 Anna Tikhonova <anna.tikhonova@intel.com>
1726 Ilya Tocar <ilya.tocar@intel.com>
1727 Andrey Turetskiy <andrey.turetskiy@intel.com>
1728 Ilya Verbin <ilya.verbin@intel.com>
1729 Kirill Yukhin <kirill.yukhin@intel.com>
1730 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1732 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
1733 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
1735 (fma<mode>4): Removed condition.
1736 (fms<mode>4): Ditto.
1737 (fnma<mode>4): Ditto.
1738 (fnms<mode>4): Ditto.
1739 (fma4i_fmadd_<mode>): Ditto.
1740 (*fma_fmadd_<mode>): Ditto.
1741 (*fma_fmsub_<mode>): Ditto.
1742 (*fma_fnmadd_<mode>): Ditto.
1743 (*fma_fnmsub_<mode>): Ditto.
1744 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
1745 (*fma_fmaddsub_<mode>): Ditto.
1746 (*fma_fmsubadd_<mode>): Ditto.
1747 (*fmai_fmadd_<mode>): Ditto.
1748 (*fmai_fmsub_<mode>): Ditto.
1749 (*fmai_fnmadd_<mode>): Ditto.
1750 (*fmai_fnmsub_<mode>): Ditto.
1752 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1753 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1754 Sergey Lega <sergey.s.lega@intel.com>
1755 Anna Tikhonova <anna.tikhonova@intel.com>
1756 Ilya Tocar <ilya.tocar@intel.com>
1757 Andrey Turetskiy <andrey.turetskiy@intel.com>
1758 Ilya Verbin <ilya.verbin@intel.com>
1759 Kirill Yukhin <kirill.yukhin@intel.com>
1760 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1762 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
1763 (VI124_256): Changed to ...
1764 (VI124_256_48_512): This.
1765 (ssepackmode): Extended with wider modes.
1766 (<code><mode>3): Changed iterator.
1767 (*avx2_<code><mode>3): Ditto.
1768 (vec_pack_trunc_<mode>): Ditto.
1770 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1771 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1772 Sergey Lega <sergey.s.lega@intel.com>
1773 Anna Tikhonova <anna.tikhonova@intel.com>
1774 Ilya Tocar <ilya.tocar@intel.com>
1775 Andrey Turetskiy <andrey.turetskiy@intel.com>
1776 Ilya Verbin <ilya.verbin@intel.com>
1777 Kirill Yukhin <kirill.yukhin@intel.com>
1778 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1780 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
1781 (VI8F_256_512): Ditto.
1782 (abs<mode>2): Changed iterator.
1783 (avx2_perm<mode>): Changed to ...
1784 (<avx2_avx512f>_perm<mode>): This.
1785 (avx2_perm<mode>_1): Changed to ...
1786 (<avx2_avx512f>_perm<mode>_1): This.
1788 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1789 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1790 Sergey Lega <sergey.s.lega@intel.com>
1791 Anna Tikhonova <anna.tikhonova@intel.com>
1792 Ilya Tocar <ilya.tocar@intel.com>
1793 Andrey Turetskiy <andrey.turetskiy@intel.com>
1794 Ilya Verbin <ilya.verbin@intel.com>
1795 Kirill Yukhin <kirill.yukhin@intel.com>
1796 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1798 * config/i386/sse.md (VI48_AVX512F): New.
1799 (VI48_AVX2): Changed to ...
1800 (VI48_AVX2_48_AVX512F): This.
1801 (avx2_ashrv<mode>): Changed to ...
1802 (<avx2_avx512f>_ashrv<mode>): This.
1803 (avx2_<shift_insn>v<mode>): Changed to ...
1804 (<avx2_avx512f>_<shift_insn>v<mode>): This.
1806 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1807 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1808 Sergey Lega <sergey.s.lega@intel.com>
1809 Anna Tikhonova <anna.tikhonova@intel.com>
1810 Ilya Tocar <ilya.tocar@intel.com>
1811 Andrey Turetskiy <andrey.turetskiy@intel.com>
1812 Ilya Verbin <ilya.verbin@intel.com>
1813 Kirill Yukhin <kirill.yukhin@intel.com>
1814 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1816 * config/i386/sse.md (VI4_AVX512F): New.
1817 (VI8_AVX2_AVX512F): Ditto.
1818 (mul<mode>3): Extended with wider modes.
1819 (*<sse4_1_avx2>_mul<mode>3): Ditto.
1820 (mul<mode>3): Ditto.
1821 (vec_widen_<s>mult_odd_<mode>): Ditto.
1823 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1824 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1825 Sergey Lega <sergey.s.lega@intel.com>
1826 Anna Tikhonova <anna.tikhonova@intel.com>
1827 Ilya Tocar <ilya.tocar@intel.com>
1828 Andrey Turetskiy <andrey.turetskiy@intel.com>
1829 Ilya Verbin <ilya.verbin@intel.com>
1830 Kirill Yukhin <kirill.yukhin@intel.com>
1831 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1833 * config/i386/sse.md (VI2_AVX512F): New.
1834 (VI124_AVX512F): Ditto.
1835 (sseunpackmode): Extended with wider modes.
1836 (sseunpackfltmode): Ditto.
1837 (vec_unpacks_float_hi_<mode>): Ditto.
1838 (vec_unpacks_float_lo_<mode>): Ditto.
1839 (vec_unpacku_float_hi_<mode>): Ditto.
1840 (vec_unpacku_float_lo_<mode>): Ditto.
1841 (vec_unpacks_lo_<mode>): Ditto.
1842 (vec_unpacks_hi_<mode>): Ditto.
1843 (vec_unpacku_lo_<mode>): Ditto.
1844 (vec_unpacku_hi_<mode>): Ditto.
1846 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1847 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1848 Sergey Lega <sergey.s.lega@intel.com>
1849 Anna Tikhonova <anna.tikhonova@intel.com>
1850 Ilya Tocar <ilya.tocar@intel.com>
1851 Andrey Turetskiy <andrey.turetskiy@intel.com>
1852 Ilya Verbin <ilya.verbin@intel.com>
1853 Kirill Yukhin <kirill.yukhin@intel.com>
1854 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1856 * config/i386/i386.md (multdiv): New.
1857 (multdiv_mnemonic): Ditto.
1858 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
1859 (<sse>_vm<multdiv_mnemonic><mode>3): This.
1860 (<sse>_vmdiv<mode>3): Removed.
1862 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1863 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1864 Sergey Lega <sergey.s.lega@intel.com>
1865 Anna Tikhonova <anna.tikhonova@intel.com>
1866 Ilya Tocar <ilya.tocar@intel.com>
1867 Andrey Turetskiy <andrey.turetskiy@intel.com>
1868 Ilya Verbin <ilya.verbin@intel.com>
1869 Kirill Yukhin <kirill.yukhin@intel.com>
1870 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1872 * config/i386/sse.md (V): Extended with wider modes.
1874 (ssehalfvecmode): Ditto.
1876 (ssepackfltmode): Ditto.
1877 (avx_vec_concat<mode>): Ditto.
1878 (V_256_512): New iterator.
1879 (VF2_512_256): Ditto.
1880 (si2dfmode): New attribute.
1881 (si2dfmodelower): Ditto.
1883 (concat_tg_mode): Ditto.
1884 (floatv4siv4df2): Changed to ...
1885 (float<si2dfmodelower><mode>2): This.
1886 (avx_cvtps2pd256): Changed to ...
1887 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
1888 (vec_pack_trunc_v4df): Changed to ...
1889 (vec_pack_trunc_<mode>): This.
1890 (avx_vpermil<mode>): Changed to ...
1891 (<sse2_avx_avx512f>_vpermil<mode>): This.
1892 (<fixsuffix>fix_truncv8dfv8si2): New.
1893 (vec_pack_sfix_trunc_v8df): Ditto.
1894 (avx512f_rndscale<mode>): Ditto.
1895 (avx512f_roundpd512): Ditto.
1896 (vec_pack_ufix_trunc_<mode>): Updated iterator.
1898 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1899 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1900 Sergey Lega <sergey.s.lega@intel.com>
1901 Anna Tikhonova <anna.tikhonova@intel.com>
1902 Ilya Tocar <ilya.tocar@intel.com>
1903 Andrey Turetskiy <andrey.turetskiy@intel.com>
1904 Ilya Verbin <ilya.verbin@intel.com>
1905 Kirill Yukhin <kirill.yukhin@intel.com>
1906 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1908 * config/i386/i386.md (any_fix): New iterator.
1909 (fixsuffix): New attribute.
1910 * config/i386/sse.md (VF1): Extened with wider modes.
1914 (sseintvecmodelower): Ditto.
1915 (ssescalarmode): Ditto.
1916 (ssescalarnum): Ditto.
1918 (ssexmmmode): Ditto.
1919 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
1920 (<sse>_rcp<mode>2): Change iterator.
1921 (rsqrt<mode>2): Ditto.
1922 (<sse>_rsqrt<mode>2): Ditto.
1923 (avx2_vec_dup<mode>): Ditto.
1924 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
1925 (round<mode>2_sfix): Ditto.
1926 (avx2_pbroadcast<mode>): Ditto.
1927 (*andnot<mode>3): Handle XI mode.
1928 (*<code><mode>3): Ditto.
1929 (AVXTOSSEMODE): Removed.
1930 (avx_vpermil<mode>): Changed to ...
1931 (<sse2_avx_avx512f>_vpermil<mode>): This.
1933 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1934 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1935 Sergey Lega <sergey.s.lega@intel.com>
1936 Anna Tikhonova <anna.tikhonova@intel.com>
1937 Ilya Tocar <ilya.tocar@intel.com>
1938 Andrey Turetskiy <andrey.turetskiy@intel.com>
1939 Ilya Verbin <ilya.verbin@intel.com>
1940 Kirill Yukhin <kirill.yukhin@intel.com>
1941 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1943 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
1944 (<sse>_comi): Ditto.
1945 (<sse>_ucomi): Ditto.
1946 (sse_cvtss2siq_2): Ditto.
1947 (sse2_cvtsd2si): Ditto.
1948 (sse2_cvtsd2siq): Ditto.
1949 (sse2_cvttsd2si): Ditto.
1950 (sse2_cvttsd2siq): Ditto.
1951 (<shift_insn><mode>3): Ditto.
1952 (sse2_cvtsi2sdq): Update constraint and prefix.
1953 (sse_cvtsi2ss): Update prefix.
1954 (sse_cvtsi2ssq): Ditto.
1956 2013-10-11 Jakub Jelinek <jakub@redhat.com>
1958 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
1959 ignore internal calls.
1961 2013-10-11 Richard Biener <rguenther@suse.de>
1963 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
1964 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
1965 TREE_OVERFLOW is set.
1967 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
1969 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
1971 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
1972 * gimple.h: Likewise.
1974 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
1976 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
1979 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
1981 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
1982 * configure: Regenerate.
1984 2013-10-11 Jakub Jelinek <jakub@redhat.com>
1986 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
1987 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
1988 formatting fixes, use pp_colon instead of pp_character (..., ':'),
1989 similarly pp_right_paren.
1990 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
1991 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
1992 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
1993 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
1995 (omp_declare_simd_clauses_equal,
1996 omp_remove_redundant_declare_simd_attrs): New functions.
1997 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
1998 (walk_tree_1): Handle new OpenMP 4.0 clauses.
1999 * tree.h (OMP_LOOP_CHECK): Define.
2000 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
2001 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
2002 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
2003 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
2004 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
2005 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
2006 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
2007 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
2008 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
2009 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
2010 OMP_CLAUSE_SIMDLEN_EXPR): Define.
2011 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
2012 (omp_remove_redundant_declare_simd_attrs): New prototype.
2013 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
2014 GIMPLE_OMP_TEAMS): New codes.
2015 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
2016 * omp-low.c (struct omp_context): Add cancel_label and cancellable
2018 (target_nesting_level): New variable.
2019 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
2020 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
2021 for collapse > 1 static schedule unless ordered.
2022 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
2023 (determine_parallel_type): Adjust get_ws_args_for caller.
2024 (install_var_field): Handle mask & 4 for double indirection.
2025 (scan_sharing_clauses): Ignore shared clause on teams construct.
2026 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
2027 (create_omp_child_function): If inside target or declare target
2028 constructs, set "omp declare target" attribute on the child function.
2029 (find_combined_for): New function.
2030 (scan_omp_parallel): Handle combined loops.
2031 (scan_omp_target, scan_omp_teams): New functions.
2032 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
2033 restrictions and set ctx->cancellable for cancellable constructs.
2034 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
2035 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
2036 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
2037 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
2038 (omp_clause_aligned_alignment): New function.
2039 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
2040 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
2041 on teams constructs. Handle user defined reductions and new
2043 (lower_reduction_clauses): Don't set placeholder to address of ref
2044 if it has already the right type.
2045 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
2046 (expand_parallel_call): Use the new non-_start suffixed builtins,
2047 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
2048 and GOMP_parallel_end after the call.
2049 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
2050 (expand_omp_for_init_counts): Handle combined loops.
2051 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
2053 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
2054 end of cancellable loops.
2055 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
2056 Likewise. Handle collapse > 1 loops.
2057 (expand_omp_simd): Handle combined loops.
2058 (expand_omp_for): Add inner_stmt argument, adjust callers of
2059 expand_omp_for* functions, use expand_omp_for_static*chunk even
2060 for collapse > 1 unless ordered.
2061 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
2062 of cancellable sections.
2063 (expand_omp_single): Remove need_barrier variable, just rely on
2064 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
2065 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
2066 (expand_omp_atomic_load, expand_omp_atomic_store,
2067 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
2068 (expand_omp_target): New function.
2069 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
2070 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
2071 (build_omp_regions_1): Immediately close region for
2072 GF_OMP_TARGET_KIND_UPDATE.
2073 (maybe_add_implicit_barrier_cancel): New function.
2074 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
2076 (lower_omp_single): Likewise. Add clobber after the barrier.
2077 (lower_omp_taskgroup): New function.
2078 (lower_omp_for): Handle combined loops. Adjust
2079 lower_rec_input_clauses caller. Handle cancellation.
2080 (lower_depend_clauses): New function.
2081 (lower_omp_taskreg): Lower depend clauses. Adjust
2082 lower_rec_input_clauses caller. Add clobber after the call. Handle
2084 (lower_omp_target, lower_omp_teams): New functions.
2085 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
2086 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
2087 and GOMP_cancellation_point calls.
2088 (lower_omp): Fold stmts inside of target region.
2089 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
2090 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2091 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
2092 (BT_FN_VOID_OMPFN_PTR_UINT,
2093 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
2094 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
2095 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
2096 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
2097 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
2098 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
2099 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
2100 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
2101 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
2102 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
2103 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
2104 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
2105 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
2106 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
2107 BUILT_IN_GOMP_PARALLEL_END.
2108 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
2109 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2110 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
2111 GF_OMP_FOR_KIND_DISTRIBUTE.
2112 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
2113 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
2114 (dump_gimple_omp_return): Print lhs if it has any.
2115 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
2116 gimple_omp_atomic_seq_cst_p.
2117 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
2118 and GIMPLE_OMP_TEAMS.
2119 * langhooks.c (lhd_omp_mappable_type): New function.
2120 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
2121 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
2123 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
2124 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
2125 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
2126 (struct gimplify_omp_ctx): Add combined_loop field.
2127 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
2128 on stmts inside of target region.
2129 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
2130 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
2131 ORT_TARGET and ORT_TARGET_DATA.
2132 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
2134 (omp_notice_threadprivate_variable): Complain about threadprivate
2135 variables in target region.
2136 (omp_notice_variable): Complain about vars with non-mappable type
2137 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
2138 (omp_check_private): Ignore ORT_TARGET* regions.
2139 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
2140 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
2141 (find_combined_omp_for): New function.
2142 (gimplify_omp_for): Handle gimplification of combined loops.
2143 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
2145 (gimplify_omp_target_update): New function.
2146 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
2147 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
2148 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
2149 (gimplify_body): If fndecl has "omp declare target" attribute, add
2150 implicit ORT_TARGET context around it.
2151 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
2152 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
2153 * tree-nested.c (convert_nonlocal_reference_stmt,
2154 convert_local_reference_stmt, convert_gimple_call): Handle
2155 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2156 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
2157 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
2158 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
2159 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
2160 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
2161 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
2162 BUILT_IN_GOMP_LOOP_END_CANCEL,
2163 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
2164 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
2165 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
2166 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
2167 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
2168 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
2169 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
2170 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
2171 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
2172 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
2173 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
2174 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
2175 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
2176 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
2177 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
2178 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2179 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
2180 gimple_build_omp_teams): New functions.
2181 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
2182 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
2183 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
2184 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2185 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
2186 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
2187 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
2188 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
2189 GF_OMP_ATOMIC_SEQ_CST): New.
2190 (gimple_build_omp_taskgroup, gimple_build_omp_target,
2191 gimple_build_omp_teams): New prototypes.
2192 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
2193 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2194 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
2195 GIMPLE_OMP_SINGLE as end of range.
2196 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
2197 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
2198 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
2199 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
2200 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
2201 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
2202 gimple_omp_target_kind, gimple_omp_target_set_kind,
2203 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
2204 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
2205 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
2206 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
2207 gimple_omp_teams_set_clauses): New inlines.
2208 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
2209 and GIMPLE_OMP_TASKGROUP.
2210 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
2211 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
2212 enum omp_clause_proc_bind_kind): New.
2213 (union omp_clause_subcode): Add depend_kind, map_kind and
2214 proc_bind_kind fields.
2215 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
2216 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2217 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
2218 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
2219 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
2221 2013-10-10 Teresa Johnson <tejohnson@google.com>
2223 * predict.c (tree_estimate_probability): Add new parameter
2224 for estimate_bb_frequencies.
2225 (estimate_bb_frequencies): Add new parameter to force estimation.
2226 (rebuild_frequencies): When max frequency in function is small,
2227 recompute counts from frequencies.
2228 * predict.h (estimate_bb_frequencies): New parameter.
2230 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2232 * ipa-inline.c (ipa_inline): Fix leak of "order" when
2233 optimizations are disabled.
2235 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2237 * coverage.c (coverage_finish): Fix leak of da_file_name.
2239 2013-10-10 Jan Hubicka <jh@suse.cz>
2241 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
2242 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
2245 2013-10-10 Jakub Jelinek <jakub@redhat.com>
2248 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
2249 if any labels are in FALLTHRU_BB, use a special label emitted
2250 immediately after the asm goto insn rather than label_rtx
2252 (expand_asm_stmt): Adjust caller.
2253 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
2254 edge if the last insn in predecessor is a jump with single successor,
2255 but it isn't simplejump_p.
2257 2013-10-10 Richard Biener <rguenther@suse.de>
2259 PR tree-optimization/58656
2260 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
2262 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2264 * gimplify.c: Include expr.h and tm_p.h for targets with special
2265 va-arg padding requirements.
2267 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2269 * tree-flow.h: Move some prototypes to gimple.h.
2270 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
2271 * gimple.h: Relocate some prototypes from tree-flow.h
2272 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
2274 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
2275 (build_va_arg_indirect_ref): Relocate and make static.
2276 (std_gimplify_va_arg_expr): Relocate here.
2277 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
2278 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
2280 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2282 * doc/md.texi: Document the mnemonic attribute.
2284 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2287 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
2288 (set_attr_alternative x ...) when searching for user defined
2291 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2293 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
2295 2013-10-09 Easwaran Raman <eraman@google.com>
2297 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
2298 * cfgexpand.c (defer_stack_allocation): ...use here
2299 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
2301 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
2303 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
2304 (operand_equal_for_value_replacement): New function, extracted from
2305 value_replacement and enhanced to catch more cases.
2306 (value_replacement): Use operand_equal_for_value_replacement.
2308 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2310 * loop-doloop.c (doloop_modify, doloop_optimize): Use
2311 get_max_loop_iterations.
2313 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2315 * config/arm/aarch-common.c (arm_early_load_addr_dep):
2316 Place comment above function.
2318 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2320 * tree-flow.h: Remove all remaining prototypes, enums and structs that
2321 are not related to tree-cfg.c.
2322 * tree-ssa-address.h: New file. Relocate prototypes.
2323 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
2324 (addr_for_mem_ref): New. Combine call to get_address_description and
2325 return addr_for_mem_ref.
2326 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
2327 * tree-ssa-live.h: Adjust prototypes.
2328 * passes.c: Include tree-ssa-live.h.
2329 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
2330 * graphite.c (graphite_transform_loops): Make static.
2331 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2332 make_pass_graphite, pass_data_graphite_transforms,
2333 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
2334 * ipa-pure-const.c (warn_function_noreturn): Make static.
2335 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2336 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2337 Relocate from tree-cfg.c
2338 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
2340 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2341 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2342 Move to ipa-pure-const.c.
2343 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
2344 Relocate from tree-optimize.c.
2345 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
2346 make_pass_fixup_cfg): Move to tree-cfg.c.
2347 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
2348 Relocate some prototypes.
2349 * tree-data-ref.h (tree_check_data_deps) Add prototype.
2350 * tree-dump.c (dump_function_to_file): Remove prototype.
2351 Add tree-flow.h to the include file.
2352 * tree-dump.h: Remove prototype.
2353 * tree-parloops.h: New File. Add prototypes.
2354 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
2355 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
2356 from tree-ssa-loop.c.
2357 * tree-predcom.c (run_tree_predictive_commoning,
2358 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
2359 Relocate here from tree-ssa-loop.c.
2360 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
2361 ssa_name_values.release ().
2362 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
2363 (ssa_name_values): Relocate from tree-flow.h.
2364 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
2365 * tree-ssa-loop.c (run_tree_predictive_commoning,
2366 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
2367 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2368 make_pass_graphite, pass_data_graphite_transforms,
2369 make_pass_graphite_transforms, gate_tree_parallelize_loops,
2370 tree_parallelize_loops, pass_data_parallelize_loops,
2371 make_pass_parallelize_loops): Move to other files.
2372 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
2374 * tree.h: Remove prototypes from tree-address.c.
2376 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2378 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
2379 (struct int_tree_map): Move to tree-hasher.h
2380 (SCALE, LABEL, PERCENT): Move to gimple.h
2381 * tree-flow-inline.h: Delete. Move functions to other files.
2382 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
2383 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
2384 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
2385 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
2386 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
2387 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
2388 * tree-hasher.h: Don't include tree-flow.h.
2389 (struct int_tree_map): Relocate from tree-flow.h.
2390 * tree-sra.c (contains_view_convert_expr_p): Relocate from
2391 tree-flow-inline.h and make static.
2392 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
2394 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
2395 tree-flow-inline.h and make static.
2396 * tree.h (is_global_var, may_be_aliased): Relocate from
2398 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
2399 * value-prof.c: No longer include tree-flow-inline.h.
2400 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
2402 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2404 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
2405 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
2406 (enum move_pos): Move to tree-ssa-loop-im.h
2407 * cfgloop.h: Move some prototypes.
2408 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
2409 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
2410 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
2411 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
2412 (loop_containing_stmt): Relocate from tree-flow-inline.h.
2413 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
2414 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
2415 (enum move_pos): Relocate here.
2416 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
2418 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
2419 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
2420 make_pass_lim): Relocate here from tree-ssa-loop.c.
2421 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
2423 (loop_edge_to_cancel, unloop_loops): Make static.
2424 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2425 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
2426 (tree_complete_unroll, gate_tree_complete_unroll,
2427 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
2428 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2429 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
2430 * tree-ssa-loop-ivopts.c: Remove local prototypes.
2431 (stmt_invariant_in_loop_p): Remove unused function.
2432 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
2433 * tree-ssa-loop-manip.h: New file. Add prototypes.
2434 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
2435 (gcov_type_to_double_int): Move to cfgloop.h.
2436 (double_int_cmp, bound_index,
2437 estimate_numbers_of_iterations_loop): Make static.
2438 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
2439 (max_loop_iterations): Factor out get_max_loop_iterations.
2440 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
2442 * tree-ssa-loop-niter.h: New file. Add prototypes.
2443 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
2444 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
2445 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
2446 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
2447 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
2448 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
2449 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
2450 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
2451 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
2452 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
2453 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2454 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
2455 pass_data_complete_unroll, make_pass_complete_unroll,
2456 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2457 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
2458 tree-ssa-loop-ivcanon.c.
2459 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
2460 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
2461 tree-ssa-loop-prefetch.c.
2462 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
2464 (get_lsm_tmp_name): Relocate and add suffix parameter.
2465 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
2466 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
2467 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
2468 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
2469 (get_estimated_loop_iterations): Factor out accessor from
2470 estimated_loop_iterations in tree-ssa-loop-niter.c.
2471 (get_max_loop_iterations): Factor out accessor from
2472 _max_loop_iterations in tree-ssa-niter.c.
2473 * loop-unroll.c (decide_unroll_constant_iterations,
2474 decide_unroll_runtime_iterations, decide_peel_simple,
2475 decide_unroll_stupid): Use new get_* accessors.
2477 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
2479 PR tree-optimization/20318
2480 * doc/extend.texi (returns_nonnull): New function attribute.
2481 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
2483 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
2484 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
2486 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
2489 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
2490 false if both components are bitfields.
2492 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2494 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
2495 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
2496 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
2497 * config/aarch64/aarch64.h
2498 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
2499 * config/aarch64/aarch64-simd-builtins.def
2500 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
2502 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2504 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
2506 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2508 * config/aarch64/arm_neon.h (vdiv_f64): Added.
2510 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2512 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
2513 (vneg_f64): New intrinsic.
2514 (vneg_s8): Asm replaced with C.
2515 (vneg_s16): Likewise.
2516 (vneg_s32): Likewise.
2517 (vneg_s64): New intrinsic.
2518 (vnegq_f32): Asm replaced with C.
2519 (vnegq_f64): Likewise.
2520 (vnegq_s8): Likewise.
2521 (vnegq_s16): Likewise.
2522 (vnegq_s32): Likewise.
2523 (vnegq_s64): Likewise.
2525 2013-10-09 Renlin Li <Renlin.Li@arm.com>
2527 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
2529 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2531 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
2532 packed stack special handling.
2533 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
2534 back to fixed stack slots for FPRs saved due to stdarg.
2536 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2538 * config/s390/s390.c (s390_frame_info): Restructure function.
2540 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2542 * config/s390/s390.c (struct s390_frame_layout): New field
2544 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
2545 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
2546 regs_ever_clobbered to char*.
2547 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
2548 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
2549 (s390_register_info): Enable FPR save slots. Move/Copy some
2550 functionality into ...
2551 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
2552 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
2554 (s390_frame_info): Do gpr slot allocation here now. stdarg does
2555 not imply a stack frame.
2556 (s390_init_frame_layout): Remove variable clobbered_regs.
2557 (s390_update_register_info): Remove function.
2558 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
2559 cannot be used for register renaming.
2560 (s390_hard_regno_scratch_ok): New function.
2561 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
2562 (s390_initial_elimination_offset): Change offset calculation of
2563 the return address pointer.
2564 (save_gprs): Deal with only r6 being saved from the call-saved regs.
2565 (restore_gprs): Set frame related flag.
2566 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
2567 (s390_emit_prologue): Call s390_register_info instead of
2568 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
2569 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
2570 (s390_optimize_prologue): Call s390_optimize_register_info.
2571 Try to remove also FPR slot save/restore INSNs. Remove frame
2572 related flags from restore INSNs.
2574 2013-10-08 DJ Delorie <dj@redhat.com>
2576 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
2579 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
2580 avoid conflict with the MI use of %c.
2581 * config/rl78/rl78-real.md: change %c to %C throughout.
2582 * config/rl78/rl78-virt.md: Likewise.
2584 2013-10-08 Jan Hubicka <jh@suse.cz>
2586 * config/i386/i386.c (ix86_option_override_internal): Switch
2587 to SSE math for -ffast-math when target ISA supports SSE2.
2589 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
2591 * tree-flow.h: Remove some prototypes.
2592 * tree.h: Remove some protypes, add a couple.
2593 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
2594 using_eh_for_cleanups_p): Add interface routines for front ends.
2595 * tree-eh.h: New file. Add protoptyes.
2596 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
2597 (add_stmt_to_eh_lp_fn): Make static.
2598 (lower_try_finally): Use new using_eh_for_cleanups_p.
2599 * emit-rtl.c: Include tree-eh.h.
2600 * gimple.h: Include tree-eh.h.
2602 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
2604 PR tree-optimization/58480
2605 * tree-vrp.c (infer_nonnull_range): New function.
2606 (infer_value_range): Call infer_nonnull_range.
2608 2013-10-08 Dehao Chen <dehao@google.com>
2610 PR tree-optimization/58619
2611 * tree-inline.c (copy_phis_for_bb): Combine location data
2614 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2617 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
2618 RTX_FRAME_RELATED_P on INSN.
2620 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2622 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
2623 (altivec_expand_vec_perm_const): Call it.
2625 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2627 * config/rs6000/vector.md (mov<mode>): Emit permuted move
2628 sequences for LE VSX loads and stores at expand time.
2629 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
2631 * config/rs6000/rs6000.c (rs6000_const_vec): New.
2632 (rs6000_gen_le_vsx_permute): New.
2633 (rs6000_gen_le_vsx_load): New.
2634 (rs6000_gen_le_vsx_store): New.
2635 (rs6000_gen_le_vsx_move): New.
2636 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
2637 (*vsx_le_perm_load_v4si): New.
2638 (*vsx_le_perm_load_v8hi): New.
2639 (*vsx_le_perm_load_v16qi): New.
2640 (*vsx_le_perm_store_v2di): New.
2641 (*vsx_le_perm_store_v4si): New.
2642 (*vsx_le_perm_store_v8hi): New.
2643 (*vsx_le_perm_store_v16qi): New.
2644 (*vsx_xxpermdi2_le_<mode>): New.
2645 (*vsx_xxpermdi4_le_<mode>): New.
2646 (*vsx_xxpermdi8_le_V8HI): New.
2647 (*vsx_xxpermdi16_le_V16QI): New.
2648 (*vsx_lxvd2x2_le_<mode>): New.
2649 (*vsx_lxvd2x4_le_<mode>): New.
2650 (*vsx_lxvd2x8_le_V8HI): New.
2651 (*vsx_lxvd2x16_le_V16QI): New.
2652 (*vsx_stxvd2x2_le_<mode>): New.
2653 (*vsx_stxvd2x4_le_<mode>): New.
2654 (*vsx_stxvd2x8_le_V8HI): New.
2655 (*vsx_stxvd2x16_le_V16QI): New.
2657 2013-10-07 Renlin Li <Renlin.Li@arm.com>
2659 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
2661 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2663 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
2664 loop to work also for 31bit ABI.
2665 Save the stack pointer for frame_size > 0.
2667 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2669 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
2670 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
2671 constraint letters from expanders.
2672 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
2673 retry count to general_operand.
2674 ("tabort"): Give operand 0 a mode.
2675 ("tabort_1"): Add mode and constraint letter for operand 0.
2676 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
2678 2013-10-04 Jeff Law <law@redhat.com>
2680 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
2682 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
2684 (thread_across_edge): Here. Call it.
2686 2013-10-04 Cary Coutant <ccoutant@google.com>
2688 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
2689 discarding a location list expression (or a piece of one).
2691 2013-10-03 Jan Hubicka <jh@suse.cz>
2693 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
2694 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
2695 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
2697 2013-10-03 Jan Hubicka <jh@suse.cz>
2699 * config/i386/i386.c (ix86_option_override_internal): Do not enable
2700 accumulate-outgoing-args when producing unwind info.
2702 2013-10-03 Wei Mi <wmi@google.com>
2704 * lra-constraints.c (insert_move_for_subreg): New function
2705 extracted from simplify_operand_subreg.
2706 (simplify_operand_subreg): Add reload for paradoxical subreg.
2708 2013-10-03 Rong Xu <xur@google.com>
2710 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
2711 the candidate of builtin_expect such that we should fix the
2712 size/time estimation.
2713 (estimate_function_body_sizes): Do the acutally size/time fix-up
2716 2013-10-03 Rong Xu <xur@google.com>
2718 * predict.c (tree_predict_by_opcode): Get the probability
2719 for builtin_expect from param builtin_expect_probability.
2720 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
2721 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
2722 * doc/invoke.texi: Add documentation for builtin-expect-probability.
2724 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
2727 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
2728 * calls.c (alloca_call_p): Use get_callee_fndecl.
2729 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
2730 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
2732 (vrp_visit_stmt): Remove duplicated code.
2734 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
2736 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
2737 ceildf2, btruncdf2, instead of vsx_* name.
2739 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
2740 iterators to only do V2DF and V4SF here. Move the DF code to
2741 rs6000.md where it is combined with SF mode. Replace <VSv> with
2742 just 'v' since only vector operations are handled with these insns
2743 after moving the DF support to rs6000.md.
2744 (vsx_sub<mode>3): Likewise.
2745 (vsx_mul<mode>3): Likewise.
2746 (vsx_div<mode>3): Likewise.
2747 (vsx_fre<mode>2): Likewise.
2748 (vsx_neg<mode>2): Likewise.
2749 (vsx_abs<mode>2): Likewise.
2750 (vsx_nabs<mode>2): Likewise.
2751 (vsx_smax<mode>3): Likewise.
2752 (vsx_smin<mode>3): Likewise.
2753 (vsx_sqrt<mode>2): Likewise.
2754 (vsx_rsqrte<mode>2): Likewise.
2755 (vsx_fms<mode>4): Likewise.
2756 (vsx_nfma<mode>4): Likewise.
2757 (vsx_copysign<mode>3): Likewise.
2758 (vsx_btrunc<mode>2): Likewise.
2759 (vsx_floor<mode>2): Likewise.
2760 (vsx_ceil<mode>2): Likewise.
2761 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
2762 (vsx_sminsf3): Likewise.
2763 (vsx_fmadf4): Likewise.
2764 (vsx_fmsdf4): Likewise.
2765 (vsx_nfmadf4): Likewise.
2766 (vsx_nfmsdf4): Likewise.
2767 (vsx_cmpdf_internal1): Likewise.
2769 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
2770 simpler to select whether a target has SPE or traditional floating
2771 point support in iterators.
2772 (TARGET_DF_SPE): Likewise.
2773 (TARGET_SF_FPR): Likewise.
2774 (TARGET_DF_FPR): Likewise.
2775 (TARGET_SF_INSN): Macros to say whether floating point support
2776 exists for a given operation for expanders.
2777 (TARGET_DF_INSN): Likewise.
2779 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
2780 combining of SF/DF mode operations, using both traditional and VSX
2788 (abs<mode>2): Combine SF/DF modes using traditional floating point
2789 instructions. Add support for using the upper DF registers with
2790 VSX support, and SF registers with power8-vector support. Update
2791 expanders for operations supported by both the SPE and traditional
2792 floating point units.
2793 (abs<mode>2_fpr): Likewise.
2794 (nabs<mode>2): Likewise.
2795 (nabs<mode>2_fpr): Likewise.
2796 (neg<mode>2): Likewise.
2797 (neg<mode>2_fpr): Likewise.
2798 (add<mode>3): Likewise.
2799 (add<mode>3_fpr): Likewise.
2800 (sub<mode>3): Likewise.
2801 (sub<mode>3_fpr): Likewise.
2802 (mul<mode>3): Likewise.
2803 (mul<mode>3_fpr): Likewise.
2804 (div<mode>3): Likewise.
2805 (div<mode>3_fpr): Likewise.
2806 (sqrt<mode>3): Likewise.
2807 (sqrt<mode>3_fpr): Likewise.
2808 (fre<Fs>): Likewise.
2809 (rsqrt<mode>2): Likewise.
2810 (cmp<mode>_fpr): Likewise.
2811 (smax<mode>3): Likewise.
2812 (smin<mode>3): Likewise.
2813 (smax<mode>3_vsx): Likewise.
2814 (smin<mode>3_vsx): Likewise.
2815 (negsf2): Delete SF operations that are merged with DF.
2822 (fmasf4_fpr): Likewise.
2823 (fmssf4_fpr): Likewise.
2824 (nfmasf4_fpr): Likewise.
2825 (nfmssf4_fpr): Likewise.
2826 (sqrtsf2): Likewise.
2827 (rsqrtsf_internal1): Likewise.
2828 (smaxsf3): Likewise.
2829 (sminsf3): Likewise.
2830 (cmpsf_internal1): Likewise.
2831 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
2832 (negdf2): Delete DF operations that are merged with SF.
2834 (nabsdf2): Likewise.
2840 (rsqrtdf_internal1): Likewise.
2841 (fmadf4_fpr): Likewise.
2842 (fmsdf4_fpr): Likewise.
2843 (nfmadf4_fpr): Likewise.
2844 (nfmsdf4_fpr): Likewise.
2845 (sqrtdf2): Likewise.
2846 (smaxdf3): Likewise.
2847 (smindf3): Likewise.
2848 (cmpdf_internal1): Likewise.
2849 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
2850 (btrunc<mode>2): Delete separate expander, and combine with the
2851 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
2852 (btrunc<mode>2_fpr): Likewise.
2853 (ceil<mode>2): Likewise.
2854 (ceil<mode>2_fpr): Likewise.
2855 (floor<mode>2): Likewise.
2856 (floor<mode>2_fpr): Likewise.
2857 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
2858 Add support for using the upper registers with VSX and
2859 power8-vector. Move insns to be closer to the define_expands. On
2860 VSX systems, prefer the traditional form of FMA over the VSX
2861 version, since the traditional form allows the target not to
2862 overlap with the inputs.
2863 (fms<mode>4_fpr): Likewise.
2864 (nfma<mode>4_fpr): Likewise.
2865 (nfms<mode>4_fpr): Likewise.
2867 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2868 Richard Earnshaw <richard.earnshaw@arm.com>
2870 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
2871 (struct mult_cost_table): Likewise.
2872 (struct mem_cost_table): Likewise.
2873 (struct fp_cost_table): Likewise.
2874 (struct vector_cost_table): Likewise.
2875 (cpu_cost_table): Likewise.
2876 * config/arm/arm.opt (mold-rts-costs): New option.
2877 (mnew-generic-costs): Likewise.
2878 * config/arm/arm.c (generic_extra_costs): New table.
2879 (cortexa15_extra_costs): Likewise.
2880 (arm_slowmul_tune): Use NULL as new costs.
2881 (arm_fastmul_tune): Likewise.
2882 (arm_strongarm_tune): Likewise.
2883 (arm_xscale_tune): Likewise.
2884 (arm_9e_tune): Likewise.
2885 (arm_v6t2_tune): Likewise.
2886 (arm_cortex_a5_tune): Likewise.
2887 (arm_cortex_a9_tune): Likewise.
2888 (arm_v6m_tune): Likewise.
2889 (arm_fa726te_tune): Likewise.
2890 (arm_cortex_a15_tune): Use cortex15_extra_costs.
2891 (arm_cortex_tune): Use generict_extra_costs.
2892 (shifter_op_p): New function.
2893 (arm_unspec_cost): Likewise.
2894 (LIBCALL_COST): Define.
2895 (arm_new_rtx_costs): New function.
2896 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
2897 table is available. Use old costs otherwise unless mnew-generic-costs
2899 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
2900 (cpu_cost_table): Declare.
2902 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
2905 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
2906 (*subs_mul_imm_<mode>)
2907 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
2908 (*sub_<shift>_<mode>)
2909 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
2910 Remove k constraint.
2912 2013-10-03 Ian Bolton <ian.bolton@arm.com>
2914 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
2916 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
2918 2013-10-02 Teresa Johnson <tejohnson@google.com>
2920 * predict.c (probably_never_executed): New function.
2921 (probably_never_executed_bb_p): Invoke probably_never_executed.
2922 (probably_never_executed_edge_p): Ditto.
2923 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
2924 Treat profile insanities conservatively.
2926 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
2928 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
2930 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
2932 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
2933 use it. Use smaller increase for scratch. Don't increase reject
2934 for early clobber scratch.
2935 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
2936 setting eliminated regs except setting fp from hfp.
2937 (lra_eliminate): Check lra_insn_recog_data on NULL.
2939 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2942 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
2943 setting -mvsx-timode by default until the underlying problem is fixed.
2944 (RS6000_CPU, power7 defaults): Likewise.
2946 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2948 * config/x-linux (host-linux.o): Remove header dependencies.
2949 Use $(COMPILE) and $(POSTCOMPILE).
2950 * config/t-linux-android (linux-android.o): Ditto.
2952 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2954 * Makefile.in (expmed.o-warn): Remove.
2956 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2958 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
2959 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
2961 2013-10-02 Teresa Johnson <tejohnson@google.com>
2963 * dojump.c (do_jump_1): Divide probability between
2964 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
2966 2013-10-02 Tom Tromey <tromey@redhat.com>
2968 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
2970 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2972 * tree-flow.h: Remove some prototypes.
2973 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
2974 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
2975 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
2976 mark_virtual_phi_result_for_renaming): Relocate here.
2977 * tree-into-ssa.h: Add prototypes.
2978 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
2979 single_pred_before_succ_order.
2980 (blocks_in_phiopt_order): Rename and move to cfganal.c.
2981 (nonfreeing_call_p) Move to gimple.c.
2982 * cfganal.c (single_pred_before_succ_order): Move and renamed from
2984 * basic-block.h (single_pred_before_succ_order): Add prototype.
2985 * gimple.c (nonfreeing_call_p): Relocate here.
2986 * gimple.h: Add prototype.
2987 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
2988 * tree-ssa-dom.h: New file. Relocate prototypes here.
2989 * tree-ssa.h: Include tree-ssa-dom.h.
2991 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2993 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
2994 Use $(COMPILE) and $(POSTCOMPILE).
2996 * config/alpha/x-alpha (driver-alpha.o): Ditto.
2998 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3000 * tree-flow.h: Remove some prototypes.
3001 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
3002 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3003 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
3004 propagate_tree_value*): Move from here to...
3005 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
3006 propagate_tree_value*): Relocate here.
3007 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3008 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
3009 * gimple-fold.c: Remove gimple-fold.h from include list.
3010 * tree-vrp.c: Remove gimple-fold.h from include list.
3011 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
3012 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
3013 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
3014 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
3015 * sese.c: Add tree-ssa-propagate.h to include list.
3017 2013-10-02 Richard Biener <rguenther@suse.de>
3019 * tree-loop-distribution.c: Include tree-vectorizer.h for
3021 (enum partition_kind): Remove PKIND_REDUCTION.
3022 (struct partition_s): Remove has_writes member, add reduction_p member.
3023 (partition_alloc): Adjust.
3024 (partition_builtin_p): Likewise.
3025 (partition_has_writes): Remove.
3026 (partition_reduction_p): New function.
3027 (partition_merge_into): Likewise.
3028 (generate_code_for_partition): Commonize builtin partition
3030 (rdg_cannot_recompute_vertex_p): Remove.
3031 (already_processed_vertex_p): Likewise.
3032 (rdg_flag_vertex): Do not set has_writes.
3033 (classify_partition): Adjust.
3034 (rdg_build_partitions): Do not set has_writes, treat all
3035 partitions as useful.
3036 (distribute_loop): Record number of library calls generated. Adjust.
3037 (tree_loop_distribution): Report number of loops and library
3038 calls generated as opt-info.
3040 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3042 * tree-flow.h: Include new .h files. Move prototypes.
3043 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
3044 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
3045 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
3046 * tree-pretty-print.h: Add prototypes from tree-flow.h.
3047 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
3048 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
3049 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
3050 * tree.h (get_ref_base_and_extent): Move prototype out.
3051 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
3053 * gimple-low.h: New File. Add prototypes from tree-flow.h.
3054 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
3055 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
3056 * tree-scalar-evolution.c: Include tree.h.
3057 * sese.c: Include tree.h.
3058 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
3059 * dwarf2out.c: Include tree-dfa.h.
3060 * tree-chrec.c: Include tree.h.
3061 * tree-data-ref.c: Include tree.h.
3063 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
3065 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
3068 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3070 * config/t-sol2 (sol2-c.o): Remove header dependencies.
3071 Use $(COMPILE) and $(POSTCOMPILE).
3072 (sol2-cxx.o): Likewise.
3073 (sol2-stubs.o): Likewise.
3075 * config/x-solaris (host-solaris.o): Likewise.
3077 * config/sparc/t-sparc (sparc.o): Remove.
3078 (sparc-c.o): Remove header dependencies.
3079 Use $(COMPILE) and $(POSTCOMPILE).
3080 * config/sparc/x-sparc: Likewise.
3082 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
3084 * config/arc/arc-opts.h: Add 2013 to Copyright years.
3085 * config/arc/arc700.md: Likewise.
3086 * config/arc/arc-modes.def: Likewise.
3087 * config/arc/arc-simd.h: Likewise.
3088 * config/arc/t-arc-uClibc: Likewise.
3089 * config/arc/t-arc-newlib: Likewise.
3091 2013-10-02 Renlin Li <renlin.li@arm.com>
3093 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
3095 (aarch64_expand_epilogue): Likewise.
3097 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3098 Yufeng Zhang <yufeng.zhang@arm.com>
3100 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
3102 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
3103 'base_in' represent a conversion and legal_cast_p_1 holds; set
3104 'base_in' with the returned value from get_unwidened.
3106 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3108 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
3109 plus_constant is not used.
3111 2013-10-01 Wei Mi <wmi@google.com>
3113 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
3114 * config/i386/i386.md: Add define_peephole2 to
3115 break partial reg stall for cvtss2sd/cvtsd2ss.
3117 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3119 * config/arc/arc.c (pass_arc_ifcvt::clone):
3120 Update for ctxt_ -> m_ctxt change.
3122 2013-10-01 Jeff Law <law@redhat.com>
3124 * tree-ssa-threadupdate.c (struct redirection_data): Delete
3125 outgoing_edge and intermediate_edge fields. Instead store the path.
3126 (redirection_data::hash): Hash on the last edge's destination index.
3127 (redirection_data::equal): Check the entire thread path.
3128 (lookup_redirectio_data): Corresponding changes.
3129 (create_edge_and_update_destination_phis): Likewise.
3130 (thread_single_edge): Likewise.
3132 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3133 Diego Novillo <dnovillo@google.com>
3135 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
3136 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
3137 (vld32wh_insn, vld32wl_insn): Delete commented-out old
3138 versions of these patterns.
3140 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
3141 (__builtin_arc_aligned): Likewise.
3143 * config/arc/arc.md: Expand adc_0 comment stating the intended
3144 purpose and why it isn't ready.
3145 Replace commented out call_value_via_label_mixed with a
3146 plain comment about bl_s.
3148 * config/arc/arc.c (stdio.h): Don't include directly.
3149 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
3150 Remove if (1) condition.
3151 (arc_encode_section_info): Fix comment.
3153 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3155 * config/arc/arc.c (arc_conditional_register_usage):
3156 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
3157 Also set reg_alloc_order for DMA config regs.
3159 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3160 Jeremy Bennett <jeremy.bennett@embecosm.com>
3162 * doc/install.texi (--with-cpu): Mention ARC.
3163 (arc-*-elf32): New paragraph.
3164 (arc-linux-uclibc): Likewise.
3165 * doc/md.texi (Machine Constraints): Add ARC part.
3166 * doc/invoke.texi: (menu): Add ARC Options.
3167 (Machine Dependent Options) <ARC Options>: Add synopsis.
3168 (node ARC Options): Add.
3169 * doc/extend.texi (long_call / short_call attribute): Add ARC.
3170 (ARC Built-in Functions): New section defining
3171 generic ARC built-in functions.
3172 (ARC SIMD Built-in Functions): New section defining SIMD specific
3174 (Declaring Attributes of Functions): Extended
3175 description of short_call and long_call attributes for ARC and
3176 added index entries.
3178 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
3179 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
3180 Joern Rennecke <joern.rennecke@embecosm.com>
3181 Muhammad Khurram Riaz <khurram.riaz@arc.com>
3182 Brendan Kehoe <brendan@zen.org>
3183 Michael Eager <eager@eagercon.com>
3184 Simon Cook <simon.cook@embecosm.com>
3185 Jeremy Bennett <jeremy.bennett@embecosm.com>
3187 * config/arc, common/config/arc: New directories.
3189 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3190 Brendan Kehoe <brendan@zen.org>
3191 Simon Cook <simon.cook@embecosm.com>
3193 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
3195 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3197 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
3198 * tree-ssa-coalesce.h: New. Move prototype to here.
3199 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
3200 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
3201 (gimple_can_coalesce_p): Move to...
3202 * gimple.c (gimple_can_coalesce_p): Here.
3204 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3206 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
3207 (dump_decl_set): Move to gimple.c.
3208 * gimple.h: Don't include tree-ssa-operands.h.
3209 (dump_decl_set): Add prototype.
3210 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3211 Move to gimple-ssa.h.
3212 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
3213 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
3214 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
3215 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
3216 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
3218 (dump_decl_set): Relocate here.
3219 * gimple-ssa.h: New file.
3220 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3221 Relocate from gimple.h.
3222 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
3223 * tree-ssa-operands.c (swap_ssa_operands): Rename from
3224 swap_tree_operands and remove non-ssa path.
3225 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
3226 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
3228 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
3229 vect_is_simple_reduction_1): Use swap_ssa_operands.
3230 * tree-flow.h: Move various prototypes to tree-phinodes.h.
3231 (enum need_phi_state): Move to tree-into-ssa.c.
3232 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3233 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
3234 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
3235 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
3236 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3237 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3238 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3239 num_imm_uses): Move to ssa-iterators.h.
3240 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
3241 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
3243 (op_iter_done, op_iter_next_def, op_iter_next_tree,
3244 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
3245 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
3246 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
3247 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
3248 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
3249 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
3250 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
3251 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
3252 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
3253 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
3254 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
3255 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
3256 (set_phi_nodes): Move to tree-phinodes.h.
3257 * tree-ssa-operands.h (enum ssa_op_iter_type,
3258 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
3259 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
3260 (dump_decl_set): Remove prototype.
3261 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
3262 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
3263 (set_phi_nodes): Relocate from tree-flow-inline.h.
3264 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
3266 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
3267 include list. Temporarily add gimple.h to include list.
3268 * ssa-iterators.h: New file.
3269 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3270 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
3271 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
3272 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
3273 Relocate from tree-ssa-operands.h.
3274 (delink_imm_use, link_imm_use_to_list, link_imm_use,
3275 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3276 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3277 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3278 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
3279 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
3280 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
3281 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
3282 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
3283 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
3284 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
3285 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
3286 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
3287 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
3288 Relocate from tree-flow-inline.h.
3289 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
3291 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
3294 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
3295 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
3296 Insert '\t' to output template.
3297 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
3298 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
3299 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
3300 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
3302 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
3304 * doc/install.texi (Host/target specific installation notes for GCC):
3305 Put @anchor before @heading.
3306 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
3307 Use @email for email addresses.
3309 2013-10-01 Jeff Law <law@redhat.com>
3311 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
3312 a vec. Only delete the path if we create one without successfully
3313 registering a jump thread.
3314 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
3316 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
3317 (paths): New vector of jump threading paths.
3318 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
3319 (THREAD_PATH): New accessor macro for the entire thread path.
3320 (lookup_redirection_data): Get intermediate and final outgoing edge
3321 from the thread path.
3322 (create_edge_and_update_destination_phis): Copy the threading path.
3323 (ssa_fix_duplicate_block_edges): Get edges and block types from the
3324 jump threading path.
3325 (ssa_redirect_edges): Get edges and block types from the jump threading
3326 path. Free the path vector.
3327 (thread_block): Get edges from the jump threading path. Look at the
3328 entire path to see if we thread to a loop exit. If we cancel a jump
3329 thread request, then free the path vector.
3330 (thread_single_edge): Get edges and block types from the jump threading
3331 path. Free the path vector.
3332 (thread_through_loop_header): Get edges and block types from the jump
3333 threading path. Free the path vector.
3334 (mark_threaded_blocks): Iterate over the vector of paths and store
3335 the path on the appropriate edge. Get edges and block types from the
3336 jump threading path.
3337 (mark_threaded_blocks): Get edges and block types from the jump
3338 threading path. Free the path vector.
3339 (thread_through_all_blocks): Use the vector of paths rather than
3340 a vector of 3-edge sets.
3341 (register_jump_thread): Accept pointer to a path vector rather
3342 than the path vector itself. Store the path vector for later use.
3345 2013-10-01 Jakub Jelinek <jakub@redhat.com>
3346 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3349 * config/s390/s390.c (s390_split_branches): Modify check for table
3351 (s390_chunkify_start): Rearrange table jump insn check in order to
3352 deal with compare and branch insns correctly.
3354 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
3358 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
3359 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
3360 define_insn_and_split.
3361 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
3362 (shiftsi3_compare): New pattern.
3364 * config/arm/unspecs.md (UNSPEC_RRX): New.
3366 2013-10-01 Alan Modra <amodra@gmail.com>
3368 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
3369 casing inout operands.
3371 2013-10-01 Richard Biener <rguenther@suse.de>
3373 PR tree-optimization/58553
3374 * tree-loop-distribution.c (struct partition_s): Add niter member.
3375 (classify_partition): Populate niter member for the partition
3376 and properly identify whether the relevant store happens before
3377 or after the loop exit.
3378 (generate_memset_builtin): Use niter member from the partition.
3379 (generate_memcpy_builtin): Likewise.
3381 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3383 * vec.h (vec_prefix, vec): Prefix member names with "m_".
3384 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
3386 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3388 * basic-block.h (edge_list): Prefix member names with "m_".
3389 * context.h (context): Likewise.
3390 * domwalk.h (dom_walker): Likewise.
3391 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
3392 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
3393 * hash-table.h (hash_table): Likewise.
3394 * machmode.h (bit_field_mode_iterator): Likewise.
3395 * pass_manager.h (pass_list): Likewise.
3396 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
3397 * tree-pass.h (pass_data): Likewise.
3398 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
3399 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
3400 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
3401 * asan.c (pass_data_asan): Update accordingly.
3402 * cfganal.c (control_dependences::find_control_dependence): Likewise.
3403 (control_dependences::control_dependences): Likewise.
3404 (control_dependences::~control_dependences): Likewise.
3405 (control_dependences::~control_dependences): Likewise.
3406 (control_dependences::get_edges_dependent_on): Likewise.
3407 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
3408 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
3409 * context.c (gcc::context::context): Likewise.
3410 * cprop.c (pass_rtl_cprop::clone): Likewise.
3411 * domwalk.c (dom_walker::walk): Likewise.
3412 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
3413 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
3414 * mode-switching.c (pass_mode_switching::clone): Likewise.
3415 * passes.c (opt_pass::opt_pass): Likewise.
3416 (pass_manager::pass_manager): Likewise.
3417 * predict.c (pass_strip_predict_hints::clone): Likewise.
3418 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
3419 (pass_split_all_insns::clone): Likewise.
3420 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
3422 (bit_field_mode_iterator::next_mode): Likewise.
3423 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
3424 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
3425 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
3426 * tree-complex.c (pass_lower_complex::clone): Likewise.
3427 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
3428 * tree-object-size.c (pass_object_sizes::clone): Likewise.
3429 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
3430 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
3431 (pass_fold_builtins::clone): Likewise.
3432 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
3433 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
3434 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
3435 (pass_cd_dce::clone): Likewise.
3436 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
3437 (pass_phi_only_cprop::clone): Likewise.
3438 * tree-ssa-dse.c (pass_dse::clone): Likewise.
3439 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
3440 * tree-ssa-loop.c (pass_lim::clone): Likewise.
3441 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
3442 * tree-ssa-pre.c (pass_fre::clone): Likewise.
3443 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
3444 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
3445 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
3446 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
3447 * tree-vrp.c (pass_vrp::clone): Likewise.
3448 * tsan.c (pass_tsan::clone): Likewise.
3450 2013-09-30 Jakub Jelinek <jakub@redhat.com>
3453 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
3454 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
3457 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
3458 optimization, punt if sign_bit_p looked through any zero extension.
3460 2013-09-30 Teresa Johnson <tejohnson@google.com>
3462 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
3463 Update redirected out edge count in joiner case.
3464 (ssa_redirect_edges): Common the joiner and non-joiner cases
3465 so that joiner case gets profile updates.
3467 2013-09-30 Richard Biener <rguenther@suse.de>
3469 PR tree-optimization/58554
3470 * tree-loop-distribution.c (classify_partition): Require
3471 unconditionally executed stores for memcpy and memset recognition.
3472 (tree_loop_distribution): Calculate dominance info.
3474 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3476 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
3477 (NO_PROFILE_COUNTERS): Likewise.
3478 (PROFILE_HOOK): Likewise.
3479 (FUNCTION_PROFILER): Likewise.
3480 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
3482 2013-09-30 Iain Sandoe <iain@codesourcery.com>
3484 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
3485 calls and defines in TARGET_MACHO conditional.
3486 (load_macho_picbase_di): Likewise.
3487 (reload_macho_picbase): Likewise.
3488 (reload_macho_picbase_si): Likewise.
3489 (reload_macho_picbase_di): Likewise.
3490 (nonlocal_goto_receiver): Likewise.
3492 2013-09-30 Nick Clifton <nickc@redhat.com>
3494 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
3495 that use the MSP430X ISA.
3496 (msp430_option_override): Scan -mmcu command line option for any
3497 MCU name that supports the MSP430X ISA.
3498 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
3499 -mmcu options which enable the MSP430X ISA.
3501 2013-09-30 Richard Biener <rguenther@suse.de>
3504 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
3505 before looking for setjmp-like calls.
3507 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3510 * config/darwin-protos.h (machopic_get_function_picbase): New.
3511 * config/darwin.c (machopic_get_function_picbase): New.
3512 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
3513 label for a new func. (load_macho_picbase_di): Likewise.
3514 (reload_macho_picbase): New expand.
3515 (reload_macho_picbase_si): New insn.
3516 (reload_macho_picbase_di): New insn.
3517 (nonlocal_goto_receiver): New define and split.
3518 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
3519 (unspecv enum): Add UNSPECV_NLGR.
3521 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3523 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
3524 that altivec registers are correctly sized on Darwin.
3526 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3528 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
3529 darwin-driver.o): Use COMPILE and POSTCOMPILE.
3530 * config/x-darwin (host-darwin.o): Likewise.
3531 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
3532 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
3533 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
3535 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
3537 * doc/invoke.texi: Fix usage of @tie{} command.
3539 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
3541 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
3543 2013-09-28 Jan Hubicka <jh@suse.cz>
3545 * config/i386/x86-tune.def: Add documentation for each of the options;
3548 2013-09-28 Jan Hubicka <jh@suse.cz>
3550 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
3552 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
3553 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
3554 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
3555 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
3556 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
3558 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
3560 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
3561 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
3562 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
3563 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
3564 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
3565 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
3566 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
3567 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
3568 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
3569 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
3570 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
3571 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
3572 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
3573 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
3574 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
3575 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
3576 gimple.h, godump.c, graphite-clast-to-gimple.c,
3577 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
3578 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
3579 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
3580 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3581 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
3582 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
3583 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
3584 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
3585 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
3586 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
3587 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
3588 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
3589 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
3590 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
3591 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
3592 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
3593 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
3594 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
3595 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
3596 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
3597 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
3598 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
3599 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
3600 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
3601 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3602 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3603 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
3604 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
3605 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
3606 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3607 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3608 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
3609 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
3610 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3611 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
3612 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
3613 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
3614 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
3615 whitespace before "(".
3617 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
3619 * expr.h (extract_bit_field): Remove packedp parameter.
3620 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
3621 from forward declaration.
3622 (store_split_bit_field): Remove packedp arg from calls to
3623 extract_fixed_bit_field.
3624 (extract_bit_field_1): Remove packedp parameter and packedp
3625 argument from recursive calls and calls to extract_fixed_bit_field.
3626 (extract_bit_field): Remove packedp parameter and corresponding
3627 arg to extract_bit_field_1.
3628 (extract_fixed_bit_field): Remove packedp parameter. Remove code
3630 (extract_split_bit_field): Remove packedp arg from call to
3631 extract_fixed_bit_field.
3632 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
3633 (copy_blkmode_from_reg): Likewise.
3634 (copy_blkmode_to_reg): Likewise.
3635 (read_complex_part): Likewise.
3636 (store_field): Likewise.
3637 (expand_expr_real_1): Likewise.
3638 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
3639 to extract_bit_field.
3640 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
3641 call to extract_bit_field.
3642 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
3643 call to extract_bit_field.
3644 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
3645 and special packedp behavior from -fstrict-volatile-bitfields
3648 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3650 * lra-eliminations.c (init_elim_table): Guard value_p.
3652 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3654 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
3655 DFmode, DImode, and SFmode in the upper VSX registers based on the
3656 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
3657 if -mpower8-vector. Combine -mvsx-timode handling with the rest
3658 of the VSX register handling.
3660 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
3662 (zero_extendsidi2_lfiwzx): Add support for loading into the
3663 Altivec registers with -mpower8-vector. Use wu/wv constraints to
3664 only do VSX memory options on Altivec registers.
3665 (extendsidi2_lfiwax): Likewise.
3666 (extendsfdf2_fpr): Likewise.
3667 (mov<mode>_hardfloat, SF/SD modes): Likewise.
3668 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
3669 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
3670 (movdi_internal64): Likewise.
3672 2013-09-27 Xinliang David Li <davidxl@google.com>
3674 * opts.c (finish_options): Adjust parameters
3675 according to vect cost model.
3676 (common_handle_option): Set dynamic vect cost
3678 targhooks.c (default_add_stmt_cost): Compute stmt cost
3680 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
3681 Use helper function.
3682 * tree-vectorizer.h (unlimited_cost_model): New function.
3683 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
3684 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
3686 (vect_enhance_data_refs_alignment): Ditto.
3687 * flag-types.h: New enum.
3688 * common/config/i386/i386-common.c (ix86_option_init_struct):
3689 No need to initialize vect_cost_model flag.
3690 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
3693 2013-09-27 Diego Novillo <dnovillo@google.com>
3695 * gimple.h (enum ssa_mode): Remove.
3697 2013-09-27 Paulo Matos <pmatos@broadcom.com>
3699 * cfgloop.h (number_of_loops): Fix typo in check for null.
3701 2013-09-27 Jakub Jelinek <jakub@redhat.com>
3704 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
3705 are children of outermost saved_cfun's loop, and set it up to
3706 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
3707 if loop != loop0 and SESE region contains bbs that belong to loop0.
3709 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3711 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
3712 (binary_scale_code_p, get_base_term, get_index_term): New functions.
3713 (set_address_segment, set_address_base, set_address_index)
3714 (set_address_disp): Accept the argument unconditionally.
3715 (baseness): Remove must_be_base_p and must_be_index_p checks.
3716 (decompose_normal_address): Classify as much as possible in the
3719 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3721 * cse.c (count_reg_usage): Handle INT_LIST.
3722 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
3723 * reginfo.c (reg_scan_mark_refs): Likewise.
3724 * reload1.c (eliminate_regs_1): Likewise.
3726 2013-09-27 Iain Sandoe <iain@codesourcery.com>
3729 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
3732 2013-09-27 Richard Biener <rguenther@suse.de>
3734 PR tree-optimization/58459
3735 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
3736 restriction not propagating into loops.
3738 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
3740 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
3741 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
3742 * doc/tree-ssa.texi (Walking use-def chains): Delete.
3744 2013-09-26 Richard Biener <rguenther@suse.de>
3746 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
3748 2013-09-26 Richard Biener <rguenther@suse.de>
3750 * alias.h (component_uses_parent_alias_set): Rename to ...
3751 (component_uses_parent_alias_set_from): ... this.
3752 * alias.c (component_uses_parent_alias_set): Rename to ...
3753 (component_uses_parent_alias_set_from): ... this and return
3755 (reference_alias_ptr_type_1): Use the result from
3756 component_uses_parent_alias_set_from instead of stripping
3757 components one at a time.
3758 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
3760 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3762 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
3763 Move prototypes to...
3764 * tree-ssa-ter.h: New File. Move prototypes here.
3765 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
3766 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
3767 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
3768 from is_replaceable_p.
3769 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
3770 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
3771 (process_replaceable): Use ter_is_replaceable_p.
3772 (find_replaceable_in_bb): Use ter_is_replaceable_p.
3773 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
3774 newly refactored ssa_is_replaceable_p.
3775 * cfgexpand.c: Include tree-outof-ssa.h.
3776 * ssaexpand.h: Delete.
3778 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3780 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
3781 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
3783 (create_gimple_tmp): Delete.
3784 (get_expr_type, build_assign, build_type_cast): Move to...
3785 * gimple-builder.c: New File.
3786 (get_expr_type): Relocate from gimple.c.
3787 (build_assign, build_type_cast): Change to only create ssanames.
3788 * gimple.h: Move prototypes to...
3789 * gimple-builder.h: New File. Here.
3790 * tree-ssa.h: And here.
3791 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
3792 count_uses_and_derefs): Relocate from gimple.c.
3793 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
3794 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
3795 * tree-ssa-math-opts (execute_cse_reciprocals): Use
3796 gimple_replace_ssa_lhs.
3797 * asan.c: Include gimple-builder.h.
3798 * Makefile.in: Add gimple-builder.o.
3800 2013-09-26 Richard Biener <rguenther@suse.de>
3802 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
3803 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
3804 (loe_visit_block): Use gcc_checking_assert.
3805 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
3806 gimple_assign_ssa_name_copy_p.
3807 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
3810 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
3812 * config/rs6000/t-rs6000 (rs6000.o): Remove.
3813 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
3815 2013-09-26 Richard Biener <rguenther@suse.de>
3817 PR tree-optimization/58539
3818 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
3819 the fact that debug statements are not taking part in loop-closed
3822 2013-09-26 Nick Clifton <nickc@redhat.com>
3824 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
3825 time warning message.
3826 (msp430_print_operand_raw): Delete unused letter parameter.
3827 (TARGET_PRINT_OPERAND_ADDRESS): Define.
3828 (msp430_print_operand_address): New function.
3829 (msp430_print_operand): Move address printing code from here to
3831 * config/msp430/msp430.md (movsipsi2): Add comment in generated
3833 (zero_extendpsisi2): Likewise.
3834 (extendpsisi2): New pattern.
3835 (andneghi3): New pattern.
3837 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
3839 * config/aarch64/aarch64.opt (mlra): New option.
3840 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
3841 (TARGET_LRA_P): Define.
3843 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
3845 * expr.c (expand_assignment): Remove obsolete comment.
3847 2013-09-25 Jeff Law <law@redhat.com>
3849 * tree-flow.h (thread_through_all_blocks): Prototype moved into
3850 tree-ssa-threadupdate.h.
3851 (register_jump_thread): Similarly.
3852 * tree-ssa-threadupdate.h: New header file.
3853 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
3854 * tree-vrp.c: Likewise.
3855 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
3856 (thread_around_empty_blocks): Change type of path vector argument to
3857 an edge,type pair from just an edge. Initialize both elements when
3858 appending to a jump threading path. Tweak references to elements
3860 (thread_across_edge): Similarly. Release memory for the elements
3862 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
3863 (dump_jump_thread_path): New function broken out from
3864 register_jump_thread.
3865 (register_jump_thread): Use dump_jump_thread_path. Change type of
3866 path vector entries. Search the path for NULL edges and dump
3867 the path if one is found. Tweak the conversion of path to 3-edge
3868 form to use the block copy type information embedded in the path.
3870 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
3872 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
3874 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
3875 Vladimir Makarov <vmakarov@redhat.com>
3877 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
3878 from the least significant bit.
3879 (strip_address_mutations): Add bitfield operations handling.
3880 (must_be_index_p): Add shifting and rotate operations handling.
3881 (set_address_base): Use must_be_base_p predicate.
3882 (set_address_index): Use must_be_index_p predicate.
3884 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
3885 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3886 Sergey Lega <sergey.s.lega@intel.com>
3887 Anna Tikhonova <anna.tikhonova@intel.com>
3888 Ilya Tocar <ilya.tocar@intel.com>
3889 Andrey Turetskiy <andrey.turetskiy@intel.com>
3890 Ilya Verbin <ilya.verbin@intel.com>
3891 Kirill Yukhin <kirill.yukhin@intel.com>
3892 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3894 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
3896 (ix86_expand_vector_move_misalign): Support new unaligned load and
3897 stores and use new names.
3898 (CODE_FOR_sse2_storedqu): Rename to ...
3899 (CODE_FOR_sse2_storedquv16qi): ... this.
3900 (CODE_FOR_sse2_loaddqu): Rename to ...
3901 (CODE_FOR_sse2_loaddquv16qi): ... this.
3902 (CODE_FOR_avx_loaddqu256): Rename to ...
3903 (CODE_FOR_avx_loaddquv32qi): ... this.
3904 (CODE_FOR_avx_storedqu256): Rename to ...
3905 (CODE_FOR_avx_storedquv32qi): ... this.
3906 * config/i386/i386.md (fpint_logic): New.
3907 * config/i386/sse.md (VMOVE): Extend for AVX512.
3911 (VI_UNALIGNED_LOADSTORE): Ditto.
3912 (sse2_avx_avx512f): Ditto.
3913 (sse2_avx2): Extend for AVX512.
3914 (sse4_1_avx2): Ditto.
3915 (avx2_avx512f): New.
3916 (sse): Extend for AVX512.
3919 (avxsizesuffix): Ditto.
3920 (sseintvecmode): Ditto.
3922 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
3923 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
3924 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3925 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
3926 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3927 (<sse2_avx_avx512f>_storedqu<mode): ... this.
3928 (<sse>_movnt<mode>): Replace constraint "x" with "v".
3929 (STORENT_MODE): Extend for AVX512.
3930 (*absneg<mode>2): Replace constraint "x" with "v".
3931 (*mul<mode>3): Ditto.
3932 (*ieee_smin<mode>3): Ditto.
3933 (*ieee_smax<mode>3): Ditto.
3934 (avx_cmp<mode>3): Replace VF with VF_128_256.
3935 (*<sse>_maskcmp<mode>3_comm): Ditto.
3936 (<sse>_maskcmp<mode>3): Ditto.
3937 (<sse>_andnot<mode>3): Extend for AVX512.
3938 (<code><mode>3, anylogic): Replace VF with VF_128_256.
3939 (<code><mode>3, fpint_logic): New.
3940 (*<code><mode>3): Extend for AVX512.
3941 (avx512flogicsuff): New.
3942 (avx512f_<logic><mode>): Ditto.
3943 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
3945 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
3946 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
3947 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
3948 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
3949 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
3950 (xop_vpermil2<mode>3): Ditto.
3951 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
3952 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
3953 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
3954 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
3956 2013-09-25 Tom Tromey <tromey@redhat.com>
3958 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
3959 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
3960 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
3961 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
3962 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
3963 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
3964 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
3965 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
3966 (GRAPHITE_HTAB_H): Remove.
3968 2013-09-25 Tom Tromey <tromey@redhat.com>
3970 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
3972 2013-09-25 Tom Tromey <tromey@redhat.com>
3974 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
3976 2013-09-25 Tom Tromey <tromey@redhat.com>
3978 * config/i386/t-i386 (i386.o): Remove.
3979 (i386-c.o): Use COMPILE and POSTCOMPILE.
3981 2013-09-25 Tom Tromey <tromey@redhat.com>
3983 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
3985 2013-09-25 Tom Tromey <tromey@redhat.com>
3987 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
3988 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
3989 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
3990 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
3991 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
3992 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
3993 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
3994 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
3995 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
3996 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
3997 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
3998 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
3999 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
4000 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
4001 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
4002 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
4003 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
4004 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
4005 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
4006 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
4007 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
4008 (tree-ssa-pre.o, tree-ssa-sccvn.o)
4009 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
4010 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
4011 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
4012 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
4013 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
4014 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
4015 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
4016 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
4017 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
4018 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
4019 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
4020 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
4021 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
4022 (graphite-clast-to-gimple.o, graphite-dependences.o)
4023 (graphite-interchange.o, graphite-poly.o)
4024 (graphite-scop-detection.o, graphite-sese-to-poly.o)
4025 (graphite-optimize-isl.o, tree-vect-loop.o)
4026 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
4027 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
4028 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
4029 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
4030 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
4031 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
4032 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
4033 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
4034 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
4035 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
4036 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
4037 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
4038 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
4039 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
4040 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
4041 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
4042 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
4043 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
4044 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
4045 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
4046 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
4047 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
4048 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
4049 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
4050 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
4051 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
4052 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
4053 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
4054 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
4055 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
4056 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
4057 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
4058 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
4059 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
4060 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
4061 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
4062 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
4063 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
4064 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
4065 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
4066 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
4067 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
4068 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
4069 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
4070 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
4071 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
4072 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
4073 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
4074 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
4075 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
4076 (gcov-dump.o): Remove.
4077 (default-c.o): Use COMPILE and POSTCOMPILE.
4078 (CFLAGS-gcc.o): New variable.
4079 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
4081 2013-09-25 Tom Tromey <tromey@redhat.com>
4083 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
4084 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
4085 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
4086 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
4087 (c-family/c-pragma.o, c-family/c-pretty-print.o)
4088 (c-family/c-semantics.o, c-family/c-ada-spec.o)
4089 (c-family/array-notation-common.o, c-family/stub-objc.o)
4090 (c-family/c-ubsan.o): Remove.
4092 2013-09-25 Tom Tromey <tromey@redhat.com>
4094 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
4096 2013-09-25 Tom Tromey <tromey@redhat.com>
4098 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
4099 to add -DENABLE_SHARED_LIBGCC.
4100 (gcc.o): Don't use subshell.
4102 2013-09-25 Tom Tromey <tromey@redhat.com>
4104 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
4105 * configure.ac: Don't invoke AM_PROG_CC_C_O.
4106 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
4107 * configure, config.in: Rebuild.
4109 2013-09-25 Tom Tromey <tromey@redhat.com>
4111 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
4112 (COMPILE, POSTCOMPILE): New variables.
4113 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
4114 (DEPFILES): New variable.
4115 Include ".Po" files.
4116 * configure.ac: Add checks for dependency checking.
4117 * configure, aclocal.m4: Regenerate.
4119 2013-09-25 Tom Tromey <tromey@redhat.com>
4121 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
4122 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
4124 2013-09-25 Tom Tromey <tromey@redhat.com>
4126 * Makefile.in (generated_files): Add options.h,
4127 target-hooks-def.h, insn-opinit.h,
4128 common/common-target-hooks-def.h, pass-instances.def,
4129 c-family/c-target-hooks-def.h.
4131 2013-09-25 Jeff Law <law@redhat.com>
4133 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
4134 than foo[foo.length () - 1] to access last member in a vec.
4135 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
4137 2013-09-25 Richard Biener <rguenther@suse.de>
4140 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
4142 2013-09-25 Jan Hubicka <jh@suse.cz>
4144 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
4147 2013-09-25 Marek Polacek <polacek@redhat.com>
4150 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
4151 when determining the type name.
4153 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
4155 * config/sh/sh.md: Fix formatting.
4157 2013-09-24 Xinliang David Li <davidxl@google.com>
4159 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
4160 max peel iterations parameter.
4161 * param.def: New parameter.
4162 * doc/invoke.texi: Document New parameter.
4164 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
4166 * gimple-pretty-print.c: Various whitespace tweaks.
4167 * tree-core.h: Likewise.
4168 * tree-pretty-print.c: Likewise.
4169 * tree-ssa-alias.c: Likewise.
4170 * tree-ssa-copy.c: Likewise.
4171 * tree-ssanames.c: Likewise.
4172 * tree-ssanames.h: Likewise.
4173 * tree-vrp.c: Likewise.
4175 2013-09-24 Alan Modra <amodra@gmail.com>
4179 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
4180 for output operands that disallow regs. Don't use EXPAND_WRITE on
4183 2013-09-24 Richard Biener <rguenther@suse.de>
4186 * tree.c (reference_alias_ptr_type): Move ...
4187 * alias.c (reference_alias_ptr_type): ... here and implement
4188 in terms of the new reference_alias_ptr_type_1.
4189 (ref_all_alias_ptr_type_p): New helper.
4190 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
4191 use ref_all_alias_ptr_type_p.
4192 (get_deref_alias_set): Add flag_strict_aliasing check here.
4193 (reference_alias_ptr_type_1): New function, split out from ...
4194 (get_alias_set): ... here.
4195 (alias_ptr_types_compatible_p): New function.
4196 * alias.h (reference_alias_ptr_type): Declare.
4197 (alias_ptr_types_compatible_p): Likewise.
4198 * tree.h (reference_alias_ptr_type): Remove.
4199 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
4200 to compare MEM_REF alias types.
4202 2013-09-24 Richard Biener <rguenther@suse.de>
4204 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
4206 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
4208 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
4209 reload helper function arrays into a single array reg_addr.
4210 (reload_fpr_gpr): Likewise.
4211 (reload_gpr_vsx): Likewise.
4212 (reload_vsx_gpr): Likewise.
4213 (struct rs6000_reg_addr): Likewise.
4214 (reg_addr): Likewise.
4215 (rs6000_debug_reg_global): Change rs6000_vector_reload,
4216 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
4217 (rs6000_init_hard_regno_mode_ok): Likewise.
4218 (rs6000_secondary_reload_direct_move): Likewise.
4219 (rs6000_secondary_reload): Likewise.
4221 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
4222 constraints: wu, ww, and wy. Repurpose wv constraint added during
4223 power8 changes. Put wg constraint in alphabetical order.
4225 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
4226 for future work to add ISA 2.07 VSX single precision support.
4227 (-mvsx-scalar-double): Change default from -1 to 1, update
4228 documentation comment.
4229 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
4230 (-mupper-regs-df): New debug switch to control whether DF values
4231 can go in the traditional Altivec registers.
4232 (-mupper-regs-sf): New debug switch to control whether SF values
4233 can go in the traditional Altivec registers.
4235 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
4237 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
4238 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
4239 Add new constraints, wu/ww/wy. Repurpose wv constraint.
4240 (rs6000_debug_legitimate_address_p): Print if we are running
4241 before, during, or after reload.
4242 (rs6000_secondary_reload): Add a comment.
4243 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
4245 * config/rs6000/constraints.md (wa constraint): Sort w<x>
4246 constraints. Update documentation string.
4247 (wd constraint): Likewise.
4248 (wf constraint): Likewise.
4249 (wg constraint): Likewise.
4250 (wn constraint): Likewise.
4251 (ws constraint): Likewise.
4252 (wt constraint): Likewise.
4253 (wx constraint): Likewise.
4254 (wz constraint): Likewise.
4255 (wu constraint): New constraint for ISA 2.07 SFmode scalar
4257 (ww constraint): Likewise.
4258 (wy constraint): Likewise.
4259 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
4260 in the previous submissions.
4261 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
4263 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
4265 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
4266 (REG_BR_PROB): Say that the probability is stored in an int_list.
4267 * reg-notes.def: Update commentary to mention INT_LIST.
4268 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
4269 (INT_LIST): New rtx.
4270 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
4271 * rtlanal.c (int_reg_note_p): New function.
4272 (alloc_reg_note): Assert that the note does not have an int argument.
4273 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
4274 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
4275 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
4276 rather than an INSN_LIST. Handle INT_LIST.
4277 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
4278 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
4279 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
4280 Manipulate them as ints rather than rtxes.
4281 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
4282 * regmove.c (copy_src_to_dest): Likewise.
4283 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
4285 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
4286 into the cases that need it.
4287 * config/arm/arm.c (arm_unwind_emit): Likewise.
4289 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
4290 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
4291 * loop-doloop.c (add_test, doloop_modify): Likewise.
4292 * loop-unswitch.c (compare_and_jump_seq): Likewise.
4293 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
4294 * predict.c (combine_predictions_for_insn): Likewise.
4295 * print-rtl.c (print_rtx): Handle INT_LIST.
4296 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
4297 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
4298 * config/arm/arm.c (emit_unlikely_jump): Likewise.
4299 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
4300 (ix86_split_fp_branch, predict_jump): Likewise.
4301 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
4302 * config/sh/sh.c (expand_cbranchsi4): Likewise.
4303 * config/spu/spu.c (ea_load_store_inline): Likewise.
4305 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
4306 value of a REG_BR_PROB note.
4307 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
4308 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
4309 * emit-rtl.c (try_split): Likewise.
4310 * predict.c (br_prob_note_reliable_p): Likewise.
4311 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
4312 * reorg.c (mostly_true_jump): Likewise.
4313 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
4314 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
4315 * config/i386/i386.c (ix86_print_operand): Likewise.
4316 * config/ia64/ia64.c (ia64_print_operand): Likewise.
4317 * config/mmix/mmix.c (mmix_print_operand): Likewise.
4318 * config/rs6000/rs6000.c (output_cbranch): Likewise.
4319 * config/s390/s390.c (s390_expand_tbegin): Likewise.
4320 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
4321 * config/sparc/sparc.c (output_cbranch): Likewise.
4322 * config/spu/spu.c (get_branch_target): Likewise.
4323 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
4324 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
4326 2013-09-23 Jan Hubicka <jh@suse.cz>
4328 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
4330 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
4331 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
4333 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
4334 (gimple_fold_call): Dump inconsistent devirtualizations; add
4335 sanity check for type based devirtualizations.
4336 * ipa-prop.c: Include ipa-utils.h
4337 (ipa_intraprocedural_devirtualization): Add sanity check.
4338 (try_make_edge_direct_virtual_call): Likewise.
4340 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
4342 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
4343 assignment statements.
4345 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
4347 * gimple-pretty-print.c (dump_ssaname_info): New function.
4348 (dump_gimple_phi): Call it.
4349 (pp_gimple_stmt_1): Likewise.
4350 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
4351 (range_info_def): Declare.
4352 * tree-pretty-print.c (pp_double_int): New function.
4353 (dump_generic_node): Call it.
4354 * tree-pretty-print.h (pp_double_int): Declare.
4355 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
4356 * tree-ssanames.h (range_info_def): New structure.
4357 (value_range_type): Move definition here.
4358 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
4360 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
4362 (set_range_info): New function.
4363 (get_range_info): Likewise.
4364 (duplicate_ssa_name_range_info): Likewise.
4365 (duplicate_ssa_name_fn): Check pointer type and call
4366 duplicate_ssa_name_range_info.
4367 * tree-ssa-copy.c (fini_copy_prop): Likewise.
4368 * tree-vrp.c (value_range_type): Remove definition, now in
4370 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
4371 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
4372 (SSA_NAME_RANGE_INFO): New macro.
4374 2013-09-23 Richard Biener <rguenther@suse.de>
4376 PR tree-optimization/58464
4377 * tree-ssa-pre.c (phi_trans_lookup): Remove.
4378 (phi_trans_add): Change to add conditionally on being not present.
4379 (phi_translate_1): Remove recursion detection here.
4380 (phi_translate): Pre-seed the cache with NULL to catch
4381 recursion here in a more generic way.
4382 (bitmap_find_leader): Adjust comment.
4383 (get_representative_for): Dump value-numbers.
4384 (create_expression_by_pieces): Likewise.
4385 (insert_into_preds_of_block): Likewise.
4387 2013-09-23 Christian Bruel <christian.bruel@st.com>
4390 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
4391 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
4393 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
4396 2013-09-20 Renlin Li <renlin.li@arm.com>
4398 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
4400 (aarch64_expand_epilogue): Likewise.
4401 (aarch64_legitimize_reload_address): Likewise.
4403 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
4405 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
4406 NULL_TREE before pushing them onto the vector. Likewise for labels.
4408 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
4410 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
4411 comparison operators when -fno-trapping-math is in effect.
4412 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
4413 comparison operators in TFmode and assert that unsupported operators
4415 (ia64_print_operand): Likewise.
4417 2013-09-21 Jan Hubicka <jh@suse.cz>
4419 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
4420 (sse_typeless_stores): Enable for core
4421 (sse_load0_by_pxor): Likewise.
4422 (four_jump_limit): Disable for core.
4423 (pad_returns): Likewise.
4424 (avoid_vector_decode): Likewise.
4425 (fuse_cmp_and_branch): Enable for cores.
4426 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
4428 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4431 * config/pa/pa.c (pa_option_override): Disable auto increment and
4432 decrement instructions until reload is completed.
4434 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
4435 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
4436 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
4438 2013-09-20 DJ Delorie <dj@redhat.com>
4439 Nick Clifton <nickc@redhat.com>
4441 * config/rl78/rl78.c: Various whitespace and comment tweaks.
4442 (need_to_save): Save bank 0 on interrupts.
4443 (characterize_address): Strip far address wrappers.
4444 (rl78_as_legitimate_address): Likewise.
4445 (transcode_memory_rtx): Likewise.
4446 (rl78_peep_movhi_p): Disable this peephole after devirt.
4447 (rl78_propogate_register_origins): Forget all origins when a
4449 * config/rl78/rl78-virt.md: Various whitespace tweaks.
4450 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
4452 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
4453 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
4454 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
4456 * config/rl78/constraints.md: Various whitespace and paren tweaks.
4458 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4460 * config/pa/pa.md: In "scc" insn patterns, change output template to
4461 handle const0_rtx in reg_or_0_operand operands.
4463 2013-09-20 Martin Husemann <martin@NetBSD.org>
4466 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
4467 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
4469 2013-09-20 Richard Biener <rguenther@suse.de>
4472 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
4473 remembering SSA name version and block index.
4474 (new_scev_info_str): Adjust.
4475 (hash_scev_info): Likewise. Also hash the block index.
4476 (eq_scev_info): Adjust.
4477 (find_var_scev_info): Likewise.
4478 (struct instantiate_cache_entry): Remove.
4479 (struct instantiate_cache_type): Use a htab to map name, block
4481 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4482 (get_instantiated_value_entry): Likewise.
4483 (hash_idx_scev_info, eq_idx_scev_info): New functions.
4484 (instantiate_scev_name): Adjust.
4486 2013-09-20 Jeff Law <law@redhat.com>
4488 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
4490 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
4492 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
4493 Call aarch64_simd_expand_args to update op[argc].
4495 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4497 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
4500 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4502 * gengtype.c (file_rules): Added rule for *.cc files.
4503 (get_output_file_with_visibility): Give fatal message when no
4506 2013-09-20 Renlin Li <renlin.li@arm.com>
4508 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
4509 (aarch64_expand_epilogue): Likewise.
4510 (aarch64_legitimize_reload_address): Likewise.
4512 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
4515 * expr.c (expand_assignment): Remove misalignp code path.
4517 2013-09-20 Marek Polacek <polacek@redhat.com>
4520 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
4521 TYPE_PRECISION. Add asserts.
4523 2013-09-20 Richard Biener <rguenther@suse.de>
4525 PR tree-optimization/58453
4526 * tree-loop-distribution.c (distribute_loop): Apply the cost
4527 model for -ftree-loop-distribute-patterns, too.
4529 2013-09-20 Richard Biener <rguenther@suse.de>
4532 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
4533 make type comparison less strict.
4535 2013-09-20 Alan Modra <amodra@gmail.com>
4537 * configure: Regenerate.
4538 * aclocal.m4: Regenerate.
4540 2013-09-20 Marek Polacek <polacek@redhat.com>
4543 * doc/extend.texi: Document that attribute used is meant to be used
4544 on variables with static storage duration.
4546 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4548 PR tree-optimization/58472
4549 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
4550 simd_lane_access set inv_p = false.
4551 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
4552 the simduid magic VAR_DECL.
4554 2013-09-19 Jan Hubicka <jh@suse.cz>
4556 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
4558 2013-09-17 Jeff Law <law@redhat.com>
4560 * tree-ssa-dom.c (record_temporary_equivalences): New function
4561 split out of dom_opt_dom_walker::after_dom_children.
4562 (dom_opt_dom_walker::thread_across_edge): Move common code
4563 in here from dom_opt_dom_walker::after_dom_children.
4564 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
4566 2013-09-19 Jan Hubicka <jh@suse.cz>
4568 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
4569 (TARGET_GENERIC): Use PROCESOR_GENERIC
4570 (enum processor_type): Unify generic32 and 64.
4571 * i386.md (cpu): Likewise.
4572 * x86-tune.def (use_leave): Enable for generic32.
4573 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
4574 * athlon.md: Change generic64 to generic in all occurences.
4575 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
4576 (ix86_target_macros_internal): Likewise.
4577 * driver-i386.c (host_detect_local_cpu): Likewise.
4578 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
4580 (generic_memcpy, generic_memset, generic_cost): This one.
4581 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
4582 (m_GENERIC32, m_GENERIC64): Remove.
4583 (m_GENERIC): Turn into one flag.
4584 (processor_target): Unify generic tunnings.
4585 (ix86_option_override_internal): Replace generic32/64 by generic.
4586 (ix86_issue_rate): Likewise.
4587 (ix86_adjust_cost): Likewise.
4589 2013-09-19 Jan Hubicka <jh@suse.cz>
4591 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
4592 of speculative flag.
4594 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4596 * omp-low.c (expand_omp_sections): Always pass len - 1 to
4597 GOMP_sections_start, even if !exit_reachable.
4599 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4601 * lra-constraints.c (need_for_all_save_p): Use macro
4602 HARD_REGNO_CALL_PART_CLOBBERED.
4603 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
4604 set up pseudo conflict hard regs.
4606 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4609 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
4612 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4614 PR rtl-optimization/58438
4615 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
4616 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
4617 from previous subpasses.
4619 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
4621 * arm.c (arm_get_frame_offsets): Validate architecture supports
4622 LDRD/STRD before accepting the tuning preference.
4623 (arm_expand_prologue): Likewise.
4624 (arm_expand_epilogue): Likewise.
4626 2013-09-18 Richard Biener <rguenther@suse.de>
4628 PR tree-optimization/58417
4629 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
4630 have chrecs with symbols defined in the loop as operands.
4631 (chrec_fold_multiply): Likewise.
4632 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
4633 parameters before folding binary operations.
4634 (struct instantiate_cache_entry_hasher): Remove.
4635 (struct instantiate_cache_type): Use a pointer-map.
4636 (instantiate_cache_type::instantiate_cache_type): New function.
4637 (instantiate_cache_type::get): Likewise.
4638 (instantiate_cache_type::set): Likewise.
4639 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4640 (get_instantiated_value_entry): Likewise.
4641 (global_cache): New global.
4642 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
4643 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
4644 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
4645 (instantiate_scev_name): Adjust.
4646 (instantiate_scev): Construct global instead of local cache.
4647 (resolve_mixers): Likewise.
4649 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
4650 Paolo Carlini <paolo.carlini@oracle.com>
4653 * doc/implement-cxx.texi: Fix references to the C++ standards.
4655 2013-09-18 Jakub Jelinek <jakub@redhat.com>
4657 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
4658 * tree-vect-data-refs.c (vect_analyze_data_refs): For
4659 simd_lane_access drs, update also DR_ALIGNED_TO.
4661 2013-09-18 Marek Polacek <polacek@redhat.com>
4664 * doc/extend.texi: Document no_sanitize_undefined attribute.
4665 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
4666 no_sanitize_undefined attribute.
4668 2013-09-18 Nick Clifton <nickc@redhat.com>
4670 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
4671 (ASM_DECLARE_FUNCTION_NAME): Define.
4673 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
4675 * compare-elim.c (find_comparison_dom_walker): New class
4676 (find_comparisons_in_bb): Rename to
4677 find_comparison_dom_walker::before_dom_children
4678 (find_comparisons): Adjust
4679 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
4681 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
4682 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
4683 (init_walk_dominator_tree): Remove declaration.
4684 (fini_walk_dominator_tree): Remove declaration.
4685 * fwprop.c (single_def_use_dom_walker): New class
4686 (single_def_use_enter_block): Convert to
4687 single_def_use_dom_walker::before_dom_children.
4688 (single_def_use_leave_block): Convert to
4689 single_def_use_dom_walker::after_dom_children.
4690 (build_single_def_use_links): Adjust.
4691 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
4693 (find_candidates_in_block): Convert to
4694 find_candidates_dom_walker::before_dom_children.
4695 (execute_strength_reduction): Adjust.
4696 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
4697 (sese_dom_walker): New class.
4698 (sese_dom_walker::sese_dom_walker): New constructor.
4699 (sese_dom_walker::~sese_dom_walker): New destructor.
4700 (build_sese_conditions_before): Convert to
4701 sese_dom_walker::before_dom_children.
4702 (build_sese_conditions_after): Convert to
4703 sese_dom_walker::after_dom_children.
4704 (build_poly_scop): Adjust
4705 * tree-into-ssa.c (rewrite_dom_walker): New class
4706 (rewrite_enter_block): Convert to
4707 rewrite_dom_walker::before_dom_children.
4708 (rewrite_leave_block): Convert to
4709 rewrite_dom_walker::after_dom_children.
4710 (rewrite_update_dom_walker): New class.
4711 (rewrite_update_enter_block): Convert to
4712 rewrite_update_dom_walker::before_dom_children.
4713 (rewrite_update_leave_block): Convert to
4714 rewrite_update_dom_walker::after_dom_children.
4715 (rewrite_blocks, rewrite_into_ssa): Adjust.
4716 (mark_def_dom_walker): New class.
4717 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
4718 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
4719 (mark_def_sites_blocks): Convert to
4720 mark_def_dom_walker::before_dom_children.
4721 (mark_def_site_blocks): Remove.
4722 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
4723 (tree_ssa_dominator_optimize): Adjust.
4724 (dom_thread_across_edge): Convert to method
4725 dom_opt_dom_walker::thread_across_edge.
4726 (dom_opt_enter_block): Convert to member function
4727 dom_opt_dom_walker::before_dom_children.
4728 (dom_opt_leave_block): Convert to member function
4729 dom_opt_dom_walker::after_dom_children.
4730 * tree-ssa-dse.c (dse_dom_walker): New class.
4731 (dse_enter_block): Convert to member function
4732 dse_dom_walker::before_dom_children.
4733 (tree_ssa_dse): Adjust.
4734 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
4735 (determine_invariantness_stmt): Convert to method
4736 invariantness_dom_walker::before_dom_children.
4737 (determine_invariantness): Remove
4738 (move_computations_dom_walker): New class.
4739 (move_computations_stmt): Convert to method
4740 move_computations_dom_walker::before_dom_children.
4741 (move_computations, tree_ssa_lim): Adjust.
4742 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
4743 (nt_init_block): Make method
4744 notrappping_dom_walker::before_dom_children.
4745 (nt_fini_block): Make
4746 method nontrapping_dom_walker::after_dom_children.
4747 (get_non_trapping): Adjust.
4748 * tree-ssa-pre.c (eliminate_dom_walker): New class.
4749 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
4750 (eliminate_leave_block): Make method.
4751 eliminate_dom_walker::after_dom_children.
4753 * tree-ssa-strlen.c (strlen_dom_walker): New class.
4754 (strlen_enter_block): Make method
4755 strlen_dom_walker::before_dom_children.
4756 (strlen_leave_block): Make
4757 method strlen_dom_walker::after_dom_children.
4758 (tree_ssa_strlen): Adjust.
4759 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
4760 (tree_ssa_uncprop): Adjust.
4761 (uncprop_leave_block): Make method
4762 uncprop_dom_walker::after_dom_children.
4763 (uncprop_leave_block): Make method
4764 uncprop_dom_walker::before_dom_children.
4766 2013-09-18 Bin Cheng <bin.cheng@arm.com>
4768 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
4769 branch in same basic block. Check both src and dest of the move insn.
4771 2013-09-17 Nick Clifton <nickc@redhat.com>
4773 * config/rl78/rl78-real.md (bf): New pattern.
4775 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
4776 (rl78_print_operand): Do not put a # before a %B.
4777 * config/rl78/rl78.opt: Tweak doc strings.
4779 2013-09-17 DJ Delorie <dj@redhat.com>
4781 * config/rl78/constraints.md (Wcv): Allow up to $r31.
4782 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
4783 (rl78_option_override): Likewise, if -mallregs.
4784 (is_virtual_register): Likewise.
4785 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
4786 (REGNO_OK_FOR_BASE_P): Likewise.
4787 * config/rl78/rl78.opt (-mallregs): New.
4789 2013-09-17 Nick Clifton <nickc@redhat.com>
4791 * config/rl78/rl78.c (need_to_save): Change return type to bool.
4792 For interrupt functions: save all call clobbered registers if the
4793 interrupt handler is not a leaf function.
4794 (rl78_expand_prologue): Always recompute the frame information.
4795 For interrupt functions: only select bank 0 if one of the bank 0
4796 registers is going to be psuhed.
4798 2013-09-17 DJ Delorie <dj@redhat.com>
4800 * config/rl78/constraints.md: For each W* constraint, rename to C*
4801 and create a W* constraint that checks for an optional ES: prefix
4803 * config/rl78/rl78.md (UNS_ES_ADDR): New.
4804 (es_addr): New. Used to wrap far addresses.
4805 * config/rl78/rl78-protos.h (rl78_es_addr): New.
4806 (rl78_es_base): New.
4807 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
4808 wrapped far addresses.
4809 (rl78_print_operand_1): Unwrap far addresses before processing.
4810 (rl78_lo16): Wrap far addresses in unspecs.
4811 (rl78_es_addr): New.
4812 (rl78_es_base): New.
4813 (insn_ok_now): Check for not-yet-wrapped far addresses.
4814 (transcode_memory_rtx): Properly re-wrap far addresses.
4816 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
4818 * config/sparc/t-rtems: Add leon3 multilibs.
4820 2013-09-17 Cong Hou <congh@google.com>
4822 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
4823 when checking the dot production pattern. The type of rhs operand
4824 of multiply is now checked correctly.
4826 2013-09-17 Jeff Law <law@redhat.com>
4828 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
4829 edge implied equivalences into successor phis.
4830 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
4831 here from tree-ssa-threadedge.c.
4832 (mark_threaded_blocks): When threading through a joiner, if both
4833 successors of the joiner's clone reach the same block, verify the
4834 PHI arguments are equal. If not, cancel the jump threading request.
4835 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
4836 tree-ssa-threadupdate.c
4837 (thread_across_edge): Don't check PHI argument equality when
4838 threading through joiner block here.
4840 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
4842 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
4843 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
4844 (warn_uninit, warn_uninitialized_vars,
4845 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4846 Move to tree-ssa-uninit.c.
4847 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
4848 (has_undefined_value_p): New. Pass dependant parts of
4849 ssa_undefined_value_p.
4850 (uninit_undefined_value_p): Use has_undefined_value_p.
4851 (warn_uninit, warn_uninitialized_vars,
4852 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4853 Move from tree-ssa.c.
4854 * tree-ssa.h: Adjust prototypes.
4856 2013-09-17 Jan Hubicka <jh@suse.cz>
4859 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
4860 * ipa-inline.c (can_inline_edge_p): Do not downgrade
4861 FUNCTION_NOT_OPTIMIZED.
4862 * ipa-inline-analysis.c (compute_inline_parameters): Function
4863 not optimized is not inlinable unless it is alwaysinline.
4864 (inline_analyze_function): Force calls in not optimized
4865 function not inlinable.
4867 2013-09-17 Jan Hubicka <jh@suse.cz>
4870 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
4872 * ipa.c (function_and_variable_visibility): Likewise.
4873 * ipa-profile.c (ipa_profile): Likewise.
4875 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
4878 * cgraph.c (cgraph_function_body_availability): Check for ifunc
4879 attribute, and don't inline the resolver in this case.
4881 2013-09-17 Teresa Johnson <tejohnson@google.com>
4883 * coverage.c (get_coverage_counts): Add missing newline.
4885 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4887 PR tree-optimization/58088
4888 * fold-const.c (mask_with_trailing_zeros): New function.
4889 (fold_binary_loc): Make sure we don't recurse infinitely
4890 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
4891 Use mask_with_trailing_zeros where appropriate.
4893 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
4895 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
4896 of 'prev' var to get better distance estimation.
4898 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
4900 * tree-inline.h (struct copy_body_data): Add transform_parameter.
4901 * tree-inline.c (is_parameter_of): New predicate.
4902 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
4903 a parameter has been remapped.
4904 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
4905 (optimize_inline_calls): Initialize transform_parameter.
4906 (copy_gimple_seq_and_replace_locals): Likewise.
4907 (tree_function_versioning): Likewise.
4908 (maybe_inline_call_in_expr): Likewise.
4910 2013-09-17 Nick Clifton <nickc@redhat.com>
4912 * config/msp430/msp430-protos.h: Add prototypes for new functions.
4913 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
4915 (is_attr_func): New function.
4916 (msp430_is_interrupt_func): New function.
4917 (is_naked_func): New function.
4918 (is_reentrant_func): New function.
4919 (is_critical_func): New function.
4920 (msp430_start_function): Add annotations for function attributes.
4921 (msp430_attr): New function.
4922 (msp430_attribute_table): New.
4923 (msp430_function_section): New function.
4924 (TARGET_ASM_FUNCTION_SECTION): Define.
4925 (msp430_builtin): New enum.
4926 (msp430_init_builtins): New function.
4927 (msp430_builtin_devl): New function.
4928 (msp430_expand_builtin): New function.
4929 (TARGET_INIT_BUILTINS): Define.
4930 (TARGET_EXPAND_BUILTINS): Define.
4931 (TARGET_BUILTIN_DECL): Define.
4932 (msp430_expand_prologue): Add support for naked, interrupt,
4933 critical and reentrant functions.
4934 (msp430_expand_epilogue): Likewise.
4935 (msp430_print_operand): Handle 'O' character.
4936 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
4938 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
4939 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
4940 (pushm): Use a 'n' rather than an 'i' constraint.
4941 (msp_return): Add generation of the interrupt return instruction.
4942 (disable_interrupts): New pattern.
4943 (enable_interrupts): New pattern.
4944 (push_intr_state): New pattern.
4945 (pop_intr_state): New pattern.
4946 (bic_SR): New pattern.
4947 (bis_SR): New pattern.
4948 * doc/extend.texi: Document MSP430 function attributes and builtin
4951 2013-09-17 Richard Biener <rguenther@suse.de>
4953 PR tree-optimization/58432
4954 * tree-loop-distribution.c (tree_loop_distribution): Also
4955 scan PHIs for outside loop uses and seed a partition from them.
4957 2013-09-17 Bin Cheng <bin.cheng@arm.com>
4959 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
4960 (restructure_reference): Call backtrace_base_for_ref.
4962 2013-09-17 Alan Modra <amodra@gmail.com>
4965 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
4968 2013-09-16 DJ Delorie <dj@redhat.com>
4970 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
4971 vregs location for RL78/G10.
4972 (rl78_expand_prologue): Avoid SEL on G10.
4973 (rl78_expand_epilogue): Likewise.
4974 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
4975 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
4976 __RL78_G10__ when appropriate.
4977 (ASM_SPEC): Pass -mg10 along to the assembler.
4978 * config/rl78/rl78.md (sel_rb): Disable for G10.
4979 * config/rl78/rl78.opt: Add -mg10 option.
4980 * config/rl78/t-rl78: Add -mg10 multilib.
4982 2013-09-16 Xinliang David Li <davidxl@google.com>
4984 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
4985 * omp-low.c (omp_max_vf): Ditto.
4986 (expand_omp_simd): Ditto.
4987 * tree-vectorizer.c (vectorize_loops): Ditto.
4988 (gate_vect_slp): Ditto.
4989 (gate_increase_alignment): Ditto.
4990 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
4991 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
4992 (gate_tree_vectorize): Name change.
4993 (tree_vectorize): Ditto.
4994 (pass_vectorize::gate): Call new function.
4995 (pass_vectorize::execute): Ditto.
4996 * opts.c: O3 default setting change.
4997 (finish_options): Check new flag.
4998 * doc/invoke.texi: Document new flags.
4999 * common.opt: New flags.
5001 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
5003 * doc/tm.texi.in (Cond Exec Macros): Remove node.
5004 (Condition Code): Don't reference it.
5005 * doc/tm.texi: Regenerate.
5007 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
5010 * lra-constraints.c (undo_optional_reloads): Consider all optional
5011 reload even if it did not get a hard reg.
5013 2013-09-16 Teresa Johnson <tejohnson@google.com>
5015 * dumpfile.c (dump_loc): Remove newline emission.
5016 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
5017 emission to dump_printf_loc calls where missing.
5018 (vect_mark_for_runtime_alias_test): Ditto.
5019 (vect_analyze_data_ref_dependence): Ditto.
5020 (vect_analyze_data_ref_dependences): Ditto.
5021 (vect_slp_analyze_data_ref_dependence): Ditto.
5022 (vect_slp_analyze_data_ref_dependences): Ditto.
5023 (vect_compute_data_ref_alignment): Ditto.
5024 (vect_update_misalignment_for_peel): Ditto.
5025 (vect_verify_datarefs_alignment): Ditto.
5026 (vector_alignment_reachable_p): Ditto.
5027 (vect_get_data_access_cost): Ditto.
5028 (vect_enhance_data_refs_alignment): Ditto.
5029 (vect_find_same_alignment_drs): Ditto.
5030 (vect_analyze_data_refs_alignment): Ditto.
5031 (vect_analyze_group_access): Ditto.
5032 (vect_analyze_data_ref_access): Ditto.
5033 (vect_analyze_data_ref_accesses): Ditto.
5034 (vect_prune_runtime_alias_test_list): Ditto.
5035 (vect_analyze_data_refs): Ditto.
5036 (vect_create_addr_base_for_vector_ref): Ditto.
5037 (vect_create_data_ref_ptr): Ditto.
5038 (vect_grouped_store_supported): Ditto.
5039 (vect_grouped_load_supported): Ditto.
5040 * value-prof.c (check_counter): Ditto.
5041 (check_ic_target): Ditto.
5042 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
5043 (vect_recog_widen_mult_pattern): Ditto.
5044 (vect_recog_widen_sum_pattern): Ditto.
5045 (vect_recog_over_widening_pattern): Ditto.
5046 (vect_recog_widen_shift_pattern): Ditto.
5047 (vect_recog_rotate_pattern): Ditto.
5048 (vect_recog_vector_vector_shift_pattern): Ditto.
5049 (vect_recog_divmod_pattern): Ditto.
5050 (vect_recog_mixed_size_cond_pattern): Ditto.
5051 (vect_recog_bool_pattern): Ditto.
5052 (vect_pattern_recog_1): Ditto.
5053 (vect_pattern_recog): Ditto.
5054 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
5055 (vect_is_simple_iv_evolution): Ditto.
5056 (vect_analyze_scalar_cycles_1): Ditto.
5057 (vect_get_loop_niters): Ditto.
5058 (vect_analyze_loop_1): Ditto.
5059 (vect_analyze_loop_form): Ditto.
5060 (vect_analyze_loop_operations): Ditto.
5061 (vect_analyze_loop_2): Ditto.
5062 (vect_analyze_loop): Ditto.
5063 (report_vect_op): Ditto.
5064 (vect_is_slp_reduction): Ditto.
5065 (vect_is_simple_reduction_1): Ditto.
5066 (vect_get_known_peeling_cost): Ditto.
5067 (vect_estimate_min_profitable_iters): Ditto.
5068 (vect_model_reduction_cost): Ditto.
5069 (vect_model_induction_cost): Ditto.
5070 (get_initial_def_for_induction): Ditto.
5071 (vect_create_epilog_for_reduction): Ditto.
5072 (vectorizable_reduction): Ditto.
5073 (vectorizable_induction): Ditto.
5074 (vectorizable_live_operation): Ditto.
5075 (vect_loop_kill_debug_uses): Ditto.
5076 (vect_transform_loop): Ditto.
5077 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
5078 (vect_stmt_relevant_p): Ditto.
5079 (process_use): Ditto.
5080 (vect_mark_stmts_to_be_vectorized): Ditto.
5081 (vect_model_simple_cost): Ditto.
5082 (vect_model_promotion_demotion_cost): Ditto.
5083 (vect_model_store_cost): Ditto.
5084 (vect_get_store_cost): Ditto.
5085 (vect_model_load_cost): Ditto.
5086 (vect_get_load_cost): Ditto.
5087 (vect_init_vector_1): Ditto.
5088 (vect_get_vec_def_for_operand): Ditto.
5089 (vect_finish_stmt_generation): Ditto.
5090 (vectorizable_call): Ditto.
5091 (vectorizable_conversion): Ditto.
5092 (vectorizable_assignment): Ditto.
5093 (vectorizable_shift): Ditto.
5094 (vectorizable_operation): Ditto.
5095 (vectorizable_store): Ditto.
5096 (vectorizable_load): Ditto.
5097 (vectorizable_condition): Ditto.
5098 (vect_analyze_stmt): Ditto.
5099 (vect_transform_stmt): Ditto.
5100 (vect_is_simple_use): Ditto.
5101 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
5102 (vect_can_advance_ivs_p): Ditto.
5103 (vect_update_ivs_after_vectorizer): Ditto.
5104 (vect_do_peeling_for_loop_bound): Ditto.
5105 (vect_gen_niters_for_prolog_loop): Ditto.
5106 (vect_update_inits_of_drs): Ditto.
5107 (vect_create_cond_for_alias_checks): Ditto.
5108 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
5109 (vect_build_slp_tree_1): Ditto.
5110 (vect_supported_load_permutation_p): Ditto.
5111 (vect_analyze_slp_instance): Ditto.
5112 (vect_analyze_slp): Ditto.
5113 (vect_make_slp_decision): Ditto.
5114 (vect_detect_hybrid_slp): Ditto.
5115 (vect_bb_vectorization_profitable_p): Ditto.
5116 (vect_slp_analyze_bb_1): Ditto.
5117 (vect_update_slp_costs_according_to_vf): Ditto.
5118 (vect_get_mask_element): Ditto.
5119 (vect_transform_slp_perm_load): Ditto.
5120 (vect_schedule_slp_instance): Ditto.
5121 (vect_schedule_slp): Ditto.
5122 (vect_slp_transform_bb): Ditto.
5123 * profile.c (read_profile_edge_counts): Ditto.
5124 (compute_branch_probabilities): Ditto.
5125 * coverage.c (get_coverage_counts): Ditto.
5127 2013-09-16 Diego Novillo <dnovillo@google.com>
5129 * tree-core.h: Add missing comment lines from refactoring of tree.h.
5131 2013-09-16 Jan Hubicka <jh@suse.cz>
5133 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
5134 abstract functions; for static functions check the presence of body.
5136 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5138 * config/aarch64/aarch64-simd-builtins.def (fma): New.
5139 * config/aarch64/aarch64-simd.md
5140 (aarch64_mla_elt<mode>): New.
5141 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
5142 (aarch64_mls_elt<mode>): Likewise.
5143 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
5144 (aarch64_fma4_elt<mode>): Likewise.
5145 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
5146 (aarch64_fma4_elt_to_128v2df): Likewise.
5147 (aarch64_fma4_elt_to_64df): Likewise.
5148 (fnma<mode>4): Likewise.
5149 (aarch64_fnma4_elt<mode>): Likewise.
5150 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
5151 (aarch64_fnma4_elt_to_128v2df): Likewise.
5152 (aarch64_fnma4_elt_to_64df): Likewise.
5153 * config/aarch64/iterators.md (VDQSF): New.
5154 * config/aarch64/arm_neon.h
5155 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
5156 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
5158 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5160 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
5161 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
5162 (aarch64_mul3_elt_to_128df): Likewise.
5163 (aarch64_mul3_elt_to_64v2df): Likewise.
5164 * config/aarch64/iterators.md (VEL): Also handle DFmode.
5166 (VMUL_CHANGE_NLANES) Likewise.
5169 * config/aarch64/arm_neon.h
5170 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
5172 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5174 * config/aarch64/arm_neon.h
5175 (vcvtx_high_f32_f64): Fix parameters.
5177 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5178 Uros Bizjak <ubizjak@gmail.com>
5180 * config/alpha.c: Include tree-ssanames.h.
5182 2013-09-16 Richard Biener <rguenther@suse.de>
5184 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
5185 (dot_rdg_1): Handle control_dd.
5186 (create_edge_for_control_dependence): New function.
5187 (create_rdg_edges): Add control dependences if asked for.
5188 (build_rdg): Likewise.
5189 (generate_loops_for_partition): If there are not necessary
5190 control stmts remove all their dependencies.
5191 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
5192 (distribute_loop): Pass on control dependences.
5193 (tree_loop_distribution): Compute control dependences and remove
5194 restriction on number of loop nodes.
5196 2013-09-16 Jakub Jelinek <jakub@redhat.com>
5198 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
5201 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
5203 * cse.c (try_const_anchors): Punt on CC modes.
5205 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5207 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
5209 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
5212 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
5215 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5218 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
5220 (darwin_objc1_section): Likewise.
5221 (darwin_file_end): Emit Image Info section when required.
5223 2013-09-14 Jan Hubicka <jh@suse.cz>
5225 * tree-into-ssa.c (gate_into_ssa): New.
5226 (pass_data_build_ssa): Use it.
5227 * cgraph.h (expand_thunk): Update prototype.
5228 * cgraphunit.c (analyze_function): Expand thunks early.
5229 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
5230 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
5231 set lowered flag; do not add new function.
5232 (assemble_thunks_and_aliases): Update.
5233 * tree-ssa.c (gate_init_datastructures): New gate.
5234 (pass_data_init_datastructures): Use it.
5236 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5239 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
5240 xmm register set described in the psABI.
5242 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
5244 * dwarf2out.c (should_emit_struct_debug): Add check
5245 for type_decl variable is not NULL.
5247 2013-09-13 Jacek Caban <jacek@codeweavers.com>
5249 * config.gcc: Use new winnt-c.c target hooks
5250 * config/t-winnt: New file
5251 * config/winnt-c.c: New file
5252 * doc/tm.texi.in: Document new hook
5253 * doc/tm.texi: Regenerated
5255 2013-09-13 Jan Hubicka <jh@suse.cz>
5258 * ipa-inline.c (check_callers): New function.
5259 (check_caller_edge): Remove.
5260 (want_inline_function_to_all_callers_p): Also permit alises that are
5262 (inline_to_all_callers): Terminate the walk when devirtualization
5263 introduce new calls.
5265 2013-09-13 Jan Hubicka <jh@suse.cz>
5267 * ipa-inline-analysis.c (struct growth_data): Add node.
5268 (do_estimate_growth_1): Fix detection of recursion.
5270 2013-09-13 Jakub Jelinek <jakub@redhat.com>
5272 PR tree-optimization/58392
5273 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
5274 to avoid shadowing of outer loop variable. If
5275 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
5276 replace_by_duplicate_decl simduid of loops that have it set and
5277 set dest_cfun->has_simduid_loops and/or
5278 dest_cfun->has_force_vect_loops.
5279 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
5280 instead of maybe_lookup_decl.
5281 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
5282 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
5283 call. Copy over force_vect and copy and remap simduid. Set
5284 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
5285 (copy_cfg_body): Remove blocks_to_copy argument. Use
5286 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
5287 caller. Don't set cfun->has_simduid_loops and/or
5288 cfun->has_force_vect_loops here.
5289 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
5291 (expand_call_inline, tree_function_versioning): Adjust copy_body
5294 2013-09-13 Martin Jambor <mjambor@suse.cz>
5297 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
5298 the assert if the edge was a speculative one.
5300 2013-09-13 Richard Biener <rguenther@suse.de>
5302 * tree-data-ref.h (known_dependences_p): Move here ...
5303 * tree-loop-distribution.c (known_dependences_p): ... from here.
5304 (dump_rdg_component, debug_rdg_component): Remove.
5306 (generate_loops_for_partition): Use gimple_uid instead of
5307 relying on matching stmt visit order.
5308 (rdg_build_partitions): Take starting stmt vector.
5309 (ldist_gen): Merge into ...
5310 (distribute_loop): ... this function. Do not compute starting
5312 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
5314 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5316 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
5317 Set type attribute correctly. Set predicable_short_it attribute.
5318 (cmpsi_shiftsi): Remove %? from output template.
5320 2013-09-13 Richard Biener <rguenther@suse.de>
5322 * tree-loop-distribution.c (struct rdg_component,
5323 rdg_defs_used_in_other_loops_p, free_rdg_components,
5324 rdg_build_components): Remove.
5325 (stmts_from_loop): Do not record virtual PHIs.
5326 (generate_loops_for_partition): Skip virtual PHIs.
5327 (build_rdg_partition_for_component): Rename to ...
5328 (build_rdg_partition_for_vertex): ... this and adjust.
5329 (rdg_build_partitions): Take a vector of starting vertices
5330 instead of components. Remove unnecessary leftover handling.
5331 (ldist_gen): Do not build components or record other stores.
5332 (distribute_loop): Do not distribute loops containing stmts
5335 2013-09-13 Christian Bruel <christian.bruel@st.com>
5338 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
5340 2013-09-13 Kai Tietz <ktietz@redhat.com>
5342 * config.gcc: Separate cases for mingw and cygwin targets,
5343 and add 64-bit cygwin target case.
5345 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
5346 dll-export inline-functions.
5347 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
5350 2013-09-13 Jeff Law <law@redhat.com>
5354 2013-09-06 Jeff Law <law@redhat.com>
5356 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5357 edge implied equivalences into successor phis.
5359 2013-09-12 DJ Delorie <dj@redhat.com>
5361 * config/rl78/rl78-virt.md: Change from | to \; for asm line
5364 2013-09-12 Brooks Moses <bmoses@google.com>
5367 * Makefile.in: Do not install driver binaries in $(target)/bin.
5369 2013-09-12 DJ Delorie <dj@redhat.com>
5371 * config/rl78/rl78.opt (mrelax): New.
5372 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
5373 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
5375 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
5377 (rl78_expand_epilogue): Likewise.
5379 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
5382 * lra-eliminations.c (remove_reg_equal_offset_note): New.
5383 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
5384 pointer elimination with using remove_reg_equal_offset_note.
5386 2013-09-12 DJ Delorie <dj@redhat.com>
5388 * config/msp430/: New port.
5389 * config.gcc (msp430): Added.
5390 * doc/invoke.texi: Document MSP430 options.
5391 * doc/install.texi: Document msp430-elf
5392 * doc/md.texi: Document msp430-elf
5393 * doc/contrib.texi: Document msp430-elf
5395 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
5398 2013-09-12 Martin Jambor <mjambor@suse.cz>
5401 * ipa-prop.c (remove_described_reference): Give up if the edge in the
5402 reference descriptor is NULL.
5403 (ipa_edge_removal_hook): If owning a reference descriptor, set its
5406 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5408 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
5409 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
5410 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
5411 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
5412 make_temp_ssa_name): move to tree-ssanames.h
5413 * tree-ssa-alias.h: Move prototype.
5414 * tree-ssa.h: Include tree-ssanames.h.
5415 * tree-ssanames.c (FREE_SSANAMES): Move to here.
5416 * tree-ssanames.h: New. Move items from tree-flow*.h
5417 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
5419 2013-09-12 Richard Biener <rguenther@suse.de>
5421 PR tree-optimization/58404
5422 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
5423 propagate non-invariant addresses into dereferences wrapped
5424 in component references.
5426 2013-09-12 Richard Biener <rguenther@suse.de>
5428 PR tree-optimization/58402
5429 * passes.def: Move pass_late_warn_uninitialized later.
5431 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5433 * tree-ssa.h: New. Move content from tree-flow.h and
5435 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
5436 Move prototypes belonging to tree-ssa.c.
5437 * tree-flow-inline.h (redirect_edge_var_map_def,
5438 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
5440 * gimple.h: Adjust prototypes.
5441 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
5443 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
5444 * tree.h: Move prototype to tree-ssa.h.
5445 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
5446 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
5447 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
5448 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
5449 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
5450 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
5451 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
5452 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
5453 graphite-blocking.c, graphite-clast-to-gimple.c,
5454 graphite-dependences.c, graphite-interchange.c,
5455 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
5456 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
5457 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
5458 ipa-reference.c, ipa-split.c, ipa-utils.c,
5459 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
5460 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
5461 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
5462 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5463 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
5464 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
5465 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
5466 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
5467 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
5468 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
5469 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
5470 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
5471 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
5472 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
5473 value-prof.c, var-tracking.c,
5474 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
5476 2013-09-12 Richard Biener <rguenther@suse.de>
5478 PR tree-optimization/58396
5479 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
5480 (build_rdg): Take a loop-nest parameter, fix memleaks.
5481 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
5483 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
5485 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
5488 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
5490 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
5491 The FMA instruction names should have a 'v' prefix.
5493 2013-09-12 Richard Biener <rguenther@suse.de>
5495 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
5496 (dot_rdg): Use popen instead of system in optional code.
5497 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
5498 (already_processed_vertex_p): Adjust.
5499 (has_anti_or_output_dependence, predecessor_has_mem_write,
5500 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
5501 rdg_flag_uses): Remove.
5502 (rdg_flag_vertex): Simplify.
5503 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
5505 (build_rdg_partition_for_component): Process the first vertex
5506 of a component only.
5507 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
5509 2013-09-12 Alan Modra <amodra@gmail.com>
5511 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
5513 2013-09-11 DJ Delorie <dj@redhat.com>
5514 Nick Clifton <nickc@redhat.com>
5516 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
5517 (rl78_stack_based_mem): New.
5518 * config/rl78/constraints.md (Iv08): New.
5527 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
5529 (movsi): Change from expand to insn-and-split.
5530 (ashrsi3): Clobber AX.
5534 * config/rl78/rl78.md (CC_REG): Fix.
5535 (addsi3): Allow memory and immediate operands.
5536 (addsi3_internal): Split into...
5537 (addsi3_internal_virt): ...new, and ...
5538 (addsi3_internal_real): ...new.
5540 (subsi3_internal_virt): New.
5541 (subsi3_internal_real): New.
5542 (mulsi3): Add memory operand.
5543 (mulsi3_rl78): Likewise.
5544 (mulsi3_g13): Likewise.
5545 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
5546 (cbranchqi4_real): Add more constraint options.
5547 (cbranchhi4_real): Expand pattern.
5548 (cbranchhi4_real_signed): New.
5549 (cbranchhi4_real_inverted): New.
5550 (cbranchsi4_real_lt): New.
5551 (cbranchsi4_real_ge): New.
5552 (cbranchsi4_real_signed): New.
5553 (cbranchsi4_real): New.
5555 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
5557 (lshrsi3_virt): Likewise.
5558 (ashlsi3_virt): Likewise.
5559 (cbranchqi4_virt_signed): New.
5560 (cbranchhi4_virt_signed): New.
5561 (cbranchsi4_virt): New.
5562 * config/rl78/rl78.c: Whitespace fixes throughout.
5563 (move_elim_pass): New.
5564 (pass_data_rl78_move_elim): New.
5565 (pass_rl78_move_elim): New.
5566 (make_pass_rl78_move_elim): New.
5567 (rl78_devirt_info): Run devirt earlier.
5568 (rl78_move_elim_info): New.
5569 (rl78_asm_file_start): Register it.
5570 (rl78_split_movsi): New.
5571 (rl78_as_legitimate_address): Allow virtual base registers when
5573 (rl78_addr_space_convert): Remove spurious debug stuff.
5574 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
5575 (rl78_print_operand): More cases for not printing '#'.
5576 (rl78_expand_compare): Remove most of the logic.
5577 (content_memory): New.
5578 (clear_content_memory): New.
5579 (get_content_index): New.
5580 (get_content_name): New.
5581 (display_content_memory): New.
5582 (update_content): New.
5583 (record_content): New.
5584 (already_contains): New.
5585 (insn_ok_now): Re-recog insns with virtual registers.
5586 (add_postponed_content_update): New.
5587 (process_postponed_content_update): New.
5588 (gen_and_emit_move): New.
5589 (transcode_memory_rtx): Record new location content.
5590 Use gen_and_emit_move.
5591 (force_into_acc): New.
5592 (move_to_acc): Use gen_and_emit_move.
5593 (move_from_acc): Likewise.
5594 (move_acc_to_reg): Likewise.
5595 (move_to_x): Likewise.
5596 (move_to_hl): Likewise.
5597 (move_to_de): Likewise.
5598 (rl78_alloc_physical_registers_op1): Record location content.
5599 (has_constraint): New.
5600 (rl78_alloc_physical_registers_op2): Record location content.
5602 (rl78_alloc_physical_registers_ro1): Likewise.
5603 (rl78_alloc_physical_registers_cmp): Likewise.
5604 (rl78_alloc_physical_registers_umul): Likewise.
5605 (rl78_alloc_address_registers_macax): New.
5606 (rl78_alloc_physical_registers): Initialize and set location
5607 content memory as needed.
5608 (rl78_reorg): Make sure split2 is called.
5609 (rl78_rtx_costs): New.
5611 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
5613 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
5614 for (not (neg ...)) and (neg (not ...)) cases.
5616 2013-09-11 Richard Biener <rguenther@suse.de>
5619 * passes.def: Split critical edges before late uninit warning passes.
5620 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
5622 2013-09-11 Jakub Jelinek <jakub@redhat.com>
5624 PR tree-optimization/58385
5625 * fold-const.c (build_range_check): If both low and high are NULL,
5626 use omit_one_operand_loc to preserve exp side-effects.
5628 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5630 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
5632 2013-09-11 Richard Biener <rguenther@suse.de>
5634 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
5635 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
5636 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
5637 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
5638 stmts_from_loop, known_dependences_p, build_empty_rdg,
5639 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
5640 * tree-loop-distribution.c: ... here.
5641 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
5642 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
5643 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
5644 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
5645 * tree-loop-distribution.c: ... here.
5646 * tree-loop-distribution.c: Include gimple-pretty-print.h.
5647 (struct partition_s): Add loops member.
5648 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
5649 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
5650 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
5652 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5653 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5654 Sergey Lega <sergey.s.lega@intel.com>
5655 Anna Tikhonova <anna.tikhonova@intel.com>
5656 Ilya Tocar <ilya.tocar@intel.com>
5657 Andrey Turetskiy <andrey.turetskiy@intel.com>
5658 Ilya Verbin <ilya.verbin@intel.com>
5659 Kirill Yukhin <kirill.yukhin@intel.com>
5660 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5662 * config/i386/constraints.md (k): New.
5664 * config/i386/i386.c (const regclass_map): Add new mask registers.
5665 (dbx_register_map): Ditto.
5666 (dbx64_register_map): Ditto.
5667 (svr4_dbx_register_map): Ditto.
5668 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
5670 (ix86_preferred_reload_class): Disable constants for mask registers.
5671 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
5672 (ix86_hard_regno_mode_ok): Support new mask registers.
5673 (x86_order_regs_for_local_alloc): Ditto.
5674 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
5675 (FIXED_REGISTERS): Add new mask registers.
5676 (CALL_USED_REGISTERS): Ditto.
5677 (REG_ALLOC_ORDER): Ditto.
5678 (VALID_MASK_REG_MODE): New.
5679 (FIRST_MASK_REG): Ditto.
5680 (LAST_MASK_REG): Ditto.
5681 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
5682 (MAYBE_MASK_CLASS_P): New.
5683 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
5684 (REG_CLASS_CONTENTS): Ditto.
5685 (MASK_REGNO_P): New.
5686 (ANY_MASK_REG_P): Ditto.
5687 (HI_REGISTER_NAMES): Add new mask registers.
5688 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
5689 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
5691 (attribute "type"): Add mskmov, msklog.
5692 (attribute "length_immediate"): Support them.
5693 (attribute "memory"): Ditto.
5694 (attribute "prefix_0f"): Ditto.
5695 (*movhi_internal): Support new mask registers.
5696 (*movqi_internal): Ditto.
5697 (define_split): Split out clobber pattern is a logic
5698 insn on mask registers.
5699 (*k<logic><mode>): New.
5700 (*andhi_1): Extend to support mask regs.
5701 (*andqi_1): Extend to support mask regs.
5703 (define_split): Split and-not to and and not if operands
5705 (*<code><mode>_1): Separate HI mode to new pattern...
5706 (*<code>hi_1): This.
5707 (*<code>qi_1): Extend to support mask regs.
5709 (kortestzhi): Ditto.
5710 (kortestchi): Ditto.
5712 (*one_cmpl<mode>2_1): Remove HImode and handle it...
5713 (*one_cmplhi2_1): ...Here, now with mask registers support.
5714 (*one_cmplqi2_1): Support new mask registers.
5715 (HI/QImode arithmetics splitter): Don't split if mask registers
5717 (HI/QImode not splitter): Ditto.
5718 * config/i386/predicated.md (mask_reg_operand): New.
5719 (general_reg_operand): Ditto.
5721 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5723 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
5724 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
5726 2013-09-10 Jeff Law <law@redhat.com>
5728 PR tree-optimization/58380
5729 * tree-ssa-threadupdate.c (thread_block): Recognize another case
5730 of threading through a buried loop header.
5732 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
5733 return value for single successor case.
5735 2013-09-10 Jan Hubicka <jh@suse.cz>
5737 * ipa-devirt.c (ipa_devirt): Enable with LTO.
5739 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
5742 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
5743 support conditional execution.
5744 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
5746 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
5748 * lra.c (lra): Clear lra_optional_reload_pseudos before every
5750 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
5751 Check destination too to check move insn.
5752 (undo_optional_reloads): Add check that the original peudo did not
5753 changed its allocation and the optional reload was inherited on last
5754 inheritance pass. Break loop after deciding to keep optional reload.
5755 (lra_undo_inheritance): Add check that inherited pseudo still in
5758 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
5760 * config/aarch64/aarch64.md (generic_sched): New.
5761 * config/aarch64/aarch64-generic.md (load): Make conditional
5762 on generic_sched attribute.
5763 (nonload): Likewise.
5765 2013-09-10 Jan Hubicka <jh@suse.cz>
5767 * lto-cgraph.c: Include ipa-utils.h.
5768 (compute_ltrans_boundary): Also add possible targets into the boundary.
5770 2013-09-10 Jan Hubicka <jh@suse.cz>
5772 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
5773 VAR_DECL of vtable rather than full expression.
5775 2013-09-10 Jan Hubicka <jh@suse.cz>
5776 Paolo Carlini <paolo.carlini@oracle.com>
5778 * cgraphunit.c (analyze_functions): Save input_location, set it
5779 to UNKNOWN_LOCATION and restore it at the end.
5781 2013-09-10 Martin Jambor <mjambor@suse.cz>
5783 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
5784 represented by a thunk.
5786 2013-09-10 Jeff Law <law@redhat.com>
5788 PR tree-optimization/58343
5789 * tree-ssa-threadupdate.c (thread_block): Identify and disable
5790 jump threading requests through loop headers buried in the middle
5791 of a jump threading path.
5793 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
5794 in return value/type.
5796 2013-09-10 Jakub Jelinek <jakub@redhat.com>
5798 PR rtl-optimization/58365
5799 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
5800 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
5803 2013-09-10 Richard Biener <rguenther@suse.de>
5805 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
5806 * tree-data-ref.c (create_rdg_vertices): Collect all data
5807 references, signal failure to the caller, use data-ref API.
5808 (build_rdg): Compute data references only once. Maintain lifetime
5809 of data references and data dependences from within RDG.
5810 (free_rdg): Free dependence relations.
5811 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
5812 inventing extra dependences.
5813 (distribute_loop): Update for RDG API changes.
5815 2013-09-10 Kai Tietz <ktietz@redhat.com>
5817 * doc/invoke.texi (fms-extensions): Document changed
5818 behavior for ms-abi targets.
5819 * config/i386/i386.c (ix86_option_override_internal):
5820 Set default value of option -fms-extension for ms-abi targets.
5822 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
5824 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
5826 2013-09-10 Alan Modra <amodra@gmail.com>
5829 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
5831 2013-09-10 Alan Modra <amodra@gmail.com>
5833 * config/rs6000/predicates.md (add_cint_operand): New.
5834 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
5835 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
5836 using add_cint_operand.
5837 (largetoc_high_plus_aix): Likewise.
5839 2013-09-09 Jakub Jelinek <jakub@redhat.com>
5841 PR tree-optimization/58364
5842 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
5843 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
5844 the current range can't be an unconditional true or false.
5846 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
5848 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
5850 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
5852 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
5854 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
5857 * doc/invoke.texi (-Wdelete-incomplete): Document it.
5859 2013-09-09 Ian Bolton <ian.bolton@arm.com>
5861 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
5862 NO_REGS for immediate that can't be moved directly into FP_REGS.
5864 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5866 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
5867 comparison with negated operand.
5868 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
5871 2013-09-09 Richard Biener <rguenther@suse.de>
5874 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
5875 of a subloop record all its block as affecting loop-closed SSA form.
5877 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5879 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
5880 of an rtx/bitpos pair.
5881 (store_fixed_bit_field): Update accordingly.
5883 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5885 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
5887 * builtins.c (expand_errno_check): Likewise.
5888 * dwarf2cfi.c (init_return_column_size): Likewise.
5889 * except.c (sjlj_mark_call_sites): Likewise.
5890 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
5891 * lra-constraints.c (emit_inc): Likewise.
5892 * ree.c (combine_set_extension): Likewise.
5893 * regmove.c (fixup_match_2): Likewise.
5894 * reload1.c (inc_for_reload): Likewise.
5896 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5898 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
5899 (force_to_mode, simplify_shift_const_1, simplify_comparison):
5900 Use gen_int_mode with the mode of the associated simplify_* call.
5901 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
5902 * expmed.c (expand_shift_1): Likewise.
5903 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5904 * loop-iv.c (iv_number_of_iterations): Likewise.
5905 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5906 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
5908 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5910 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
5911 of the associated expand_* call.
5912 (asan_emit_stack_protection): Likewise.
5913 * builtins.c (round_trampoline_addr): Likewise.
5914 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
5915 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
5916 (emit_store_flag): Likewise.
5917 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
5918 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
5920 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5921 * ifcvt.c (noce_try_store_flag_constants): Likewise.
5922 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5923 * modulo-sched.c (generate_prolog_epilog): Likewise.
5924 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
5925 (expand_ctz, expand_ffs, expand_unop): Likewise.
5927 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5929 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
5930 of the associated gen_rtx_* call.
5931 * caller-save.c (init_caller_save): Likewise.
5932 * combine.c (find_split_point, make_extraction): Likewise.
5933 (make_compound_operation): Likewise.
5934 * dwarf2out.c (mem_loc_descriptor): Likewise.
5935 * explow.c (plus_constant, probe_stack_range): Likewise.
5936 * expmed.c (expand_mult_const): Likewise.
5937 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
5938 * reload1.c (init_reload): Likewise.
5939 * valtrack.c (cleanup_auto_inc_dec): Likewise.
5940 * var-tracking.c (adjust_mems): Likewise.
5941 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
5942 rather than gen_rtx_fmt_ee.
5944 2013-09-09 Jan Hubicka <jh@suse.cz>
5947 * value-prof.c (gimple_ic): Copy also abnormal edges.
5949 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5951 * asan.c (asan_shadow_cst): Use gen_int_mode.
5953 2013-09-08 Jan Hubicka <jh@suse.cz>
5955 * ipa-profile.c: Add toplevel comment.
5956 (ipa_propagate_frequency_1): Be more conservative when profile is read.
5957 (contains_hot_call_p): New function.
5958 (ipa_propagate_frequency): Set frequencies based on counts when
5960 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
5961 profile; do not tamper with profile after inlining if it is read.
5963 2013-09-08 Jan Hubicka <jh@suse.cz>
5965 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
5968 2013-09-08 Jan Hubicka <jh@suse.cz>
5970 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
5973 2013-09-08 Jeff Law <law@redhat.com>
5976 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
5979 2013-09-08 Andi Kleen <ak@linux.intel.com>
5981 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
5983 2013-09-08 Jan Hubicka <jh@suse.cz>
5985 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
5987 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
5989 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
5990 for non-debug insns.
5991 * lra.c (new_insn_reg): Take the containing insn as a parameter.
5992 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
5993 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
5996 2013-09-08 Jan Hubicka <jh@suse.cz>
5998 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
5999 targets and devirtualize to BUILT_IN_UNREACHABLE.
6000 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
6001 * ipa.c (walk_polymorphic_call_targets): New function.
6002 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
6003 functions; use the new timevar.
6004 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
6005 was removed from the program.
6006 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
6007 not consider it in the walk when its vtable is dead.
6008 (possible_polymorphic_call_targets_1): Pass anonymous flag to
6010 (devirt_variable_node_removal_hook): New function.
6011 (possible_polymorphic_call_targets): Also register
6012 devirt_variable_node_removal_hook.
6013 (ipa_devirt): Do not do non-speculative devirtualization.
6014 (gate_ipa_devirt): One execute if devirtualizing speculatively.
6016 2013-09-08 Jan Hubicka <jh@suse.cz>
6018 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
6019 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
6020 varpool_remove_variable_insertion_hook): Declare.
6021 * varpool.c (varpool_node_hook_list): New structure.
6022 (first_varpool_node_removal_hook,
6023 first_varpool_variable_insertion_hook): New variables.
6024 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
6025 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
6026 varpool_remove_variable_insertion_hook,
6027 varpool_call_variable_insertion_hooks): New functions.
6028 (varpool_remove_node): Use it.
6030 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
6033 * diagnostic.c (diagnostic_build_prefix): When s.file is
6034 "<built-in>" don't output line and column numbers.
6036 2013-09-06 Jan Hubicka <jh@suse.cz>
6038 * cgraphunit.c (expand_thunk): Get body before touching arguments.
6039 * lto-streamer-out.c: Stream thunks, too.
6040 * lto-streamer-in.c (input_function): Pop cfun here
6041 (lto_read_body): Instead of here.
6043 2013-09-06 Caroline Tice <cmtice@google.com>
6045 * doc/install.texi: Add documentation for the --enable-vtable-verify
6046 and the --disable-libvtv configure options.
6048 2013-09-06 Jeff Law <law@redhat.com>
6050 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
6051 edge implied equivalences into successor phis.
6053 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
6055 * resource.c (mark_referenced_resources): Handle COND_EXEC.
6057 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
6059 * resource.c (mark_target_live_regs): Compute resources taking
6060 into account if a call is predicated or not.
6062 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6064 * toplev.c (output_stack_usage): Be prepared for suffixes created by
6065 the compiler in the function names.
6067 2013-09-06 Jan Hubicka <jh@suse.cz>
6070 * ipa-inline.c (has_caller_p): New function.
6071 (want_inline_function_to_all_callers_p): Use it.
6072 (sum_callers, inline_to_all_callers): Break out from ...
6073 (ipa_inline): ... here.
6075 2013-09-06 Jan Hubicka <jh@suse.cz>
6077 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
6078 only when AVX is enabled.
6080 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6082 * config/aarch64/aarch64.md
6083 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
6084 is fpsimd_<load/store>2.
6085 (load_pair<mode>): Likewise.
6086 (store_pair<mode>): Likewise.
6088 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6090 * config/arm/types.md (type): Add "mrs" type.
6091 * config/aarch64/aarch64.md
6092 (aarch64_load_tp_hard): Make type "mrs".
6094 (load_tp_hard): Make type "mrs".
6095 * config/arm/cortex-a15.md: Update with new attributes.
6096 * config/arm/cortex-a5.md: Update with new attributes.
6097 * config/arm/cortex-a53.md: Update with new attributes.
6098 * config/arm/cortex-a7.md: Update with new attributes.
6099 * config/arm/cortex-a8.md: Update with new attributes.
6100 * config/arm/cortex-a9.md: Update with new attributes.
6101 * config/arm/cortex-m4.md: Update with new attributes.
6102 * config/arm/cortex-r4.md: Update with new attributes.
6103 * config/arm/fa526.md: Update with new attributes.
6104 * config/arm/fa606te.md: Update with new attributes.
6105 * config/arm/fa626te.md: Update with new attributes.
6106 * config/arm/fa726te.md: Update with new attributes.
6108 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6110 * config/aarch64/aarch64.md
6111 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
6112 (*movtf_aarch64): Likewise.
6114 (thumb1_movdi_insn): Use "multiple" for type where more than one
6115 instruction is used for a move.
6116 (*arm32_movhf): Likewise.
6117 (*thumb_movdf_insn): Likewise.
6119 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6121 * config/arm/types.md (type): Rename fcpys to fmov.
6123 (*arm_movsi_vfp): Rename type fcpys as fmov.
6124 (*thumb2_movsi_vfp): Likewise
6125 (*movhf_vfp_neon): Likewise
6126 (*movhf_vfp): Likewise
6127 (*movsf_vfp): Likewise
6128 (*thumb2_movsf_vfp): Likewise
6129 (*movsfcc_vfp): Likewise
6130 (*thumb2_movsfcc_vfp): Likewise
6131 * config/aarch64/aarch64-simd.md
6132 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
6133 * config/aarch64/aarch64.md
6134 (*movsi_aarch64): Replace type mov_reg with fmovs.
6135 (*movdi_aarch64): Likewise
6136 (*movsf_aarch64): Likewise
6137 (*movdf_aarch64): Likewise
6139 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
6140 * config/arm/iwmmxt.md
6141 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
6142 * config/arm/arm1020e.md: Update with new attributes.
6143 * config/arm/cortex-a15-neon.md: Update with new attributes.
6144 * config/arm/cortex-a5.md: Update with new attributes.
6145 * config/arm/cortex-a53.md: Update with new attributes.
6146 * config/arm/cortex-a7.md: Update with new attributes.
6147 * config/arm/cortex-a8-neon.md: Update with new attributes.
6148 * config/arm/cortex-a9.md: Update with new attributes.
6149 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6150 * config/arm/cortex-r4f.md: Update with new attributes.
6151 * config/arm/marvell-pj4.md: Update with new attributes.
6152 * config/arm/vfp11.md: Update with new attributes.
6154 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6156 * config/aarch64/aarch64.md
6157 (*madd<mode>): Fix type attribute.
6158 (*maddsi_uxtw): Likewise.
6159 (*msub<mode>): Likewise.
6160 (*msubsi_uxtw): Likewise.
6161 (<su_optab>maddsidi4): Likewise.
6162 (<su_optab>msubsidi4): Likewise.
6164 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6166 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
6167 * config/arm/arm.md (core_cycles): Remove fdiv.
6168 * config/arm/vfp.md:
6169 (*sqrtsf2_vfp): Update for attribute changes.
6170 (*sqrtdf2_vfp): Likewise.
6171 * config/aarch64/aarch64.md:
6172 (sqrt<mode>2): Update for attribute changes.
6173 * config/arm/arm1020e.md: Update with new attributes.
6174 * config/arm/cortex-a15-neon.md: Update with new attributes.
6175 * config/arm/cortex-a5.md: Update with new attributes.
6176 * config/arm/cortex-a53.md: Update with new attributes.
6177 * config/arm/cortex-a7.md: Update with new attributes.
6178 * config/arm/cortex-a8-neon.md: Update with new attributes.
6179 * config/arm/cortex-a9.md: Update with new attributes.
6180 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6181 * config/arm/cortex-r4f.md: Update with new attributes.
6182 * config/arm/marvell-pj4.md: Update with new attributes.
6183 * config/arm/vfp11.md: Update with new attributes.
6185 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6187 * config/arm/types.md
6188 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
6189 * config/aarch64/aarch64.md
6190 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
6192 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
6193 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
6194 (float<GPI:mode><GPF:mode>2): Likewise.
6196 (*truncsisf2_vfp): Update with new attributes.
6197 (*truncsidf2_vfp): Likewise.
6198 (fixuns_truncsfsi2): Likewise.
6199 (fixuns_truncdfsi2): Likewise.
6200 (*floatsisf2_vfp): Likewise.
6201 (*floatsidf2_vfp): Likewise.
6202 (floatunssisf2): Likewise.
6203 (floatunssidf2): Likewise.
6204 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
6205 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
6206 * config/arm/arm1020e.md: Update with new attributes.
6207 * config/arm/cortex-a15-neon.md: Update with new attributes.
6208 * config/arm/cortex-a5.md: Update with new attributes.
6209 * config/arm/cortex-a53.md: Update with new attributes.
6210 * config/arm/cortex-a7.md: Update with new attributes.
6211 * config/arm/cortex-a8-neon.md: Update with new attributes.
6212 * config/arm/cortex-a9.md: Update with new attributes.
6213 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6214 * config/arm/cortex-r4f.md: Update with new attributes.
6215 * config/arm/marvell-pj4.md: Update with new attributes.
6216 * config/arm/vfp11.md: Update with new attributes.
6218 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6220 * config/aarch64/arm_neon.h
6221 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
6222 (vqtbx<1,2,3,4><q>_s8): Likewise.
6224 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6226 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
6227 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
6228 (add<mode>3): Add type attribute.
6229 (add<mode>3): Likewise.
6230 (usadd<mode>3): Likewise.
6231 (ssadd<mode>3): Likewise.
6232 (sub<mode>3): Likewise.
6233 (sub<mode>3): Likewise.
6234 (ussub<mode>3): Likewise.
6235 (sssub<mode>3): Likewise.
6236 (ssmulsa3): Likewise.
6237 (usmulusa3): Likewise.
6238 (arm_usatsihi): Likewise.
6240 (*movdi_vfp): Add types for all instructions.
6241 (*movdi_vfp_cortexa8): Likewise.
6242 (*movhf_vfp_neon): Likewise.
6243 (*movhf_vfp): Likewise.
6244 (*movdf_vfp): Likewise.
6245 (*thumb2_movdf_vfp): Likewise.
6246 (*thumb2_movdfcc_vfp): Likewise.
6247 * config/arm/arm.md: Add type attribute to all insn patterns.
6248 (*thumb1_adddi3): Add type attribute.
6249 (*arm_adddi3): Likewise.
6250 (*adddi_sesidi_di): Likewise.
6251 (*adddi_zesidi_di): Likewise.
6252 (*thumb1_addsi3): Likewise.
6253 (addsi3_compare0): Likewise.
6254 (*addsi3_compare0_scratch): Likewise.
6255 (*compare_negsi_si): Likewise.
6256 (cmpsi2_addneg): Likewise.
6257 (*addsi3_carryin_<optab>): Likewise.
6258 (*addsi3_carryin_alt2_<optab>): Likewise.
6259 (*addsi3_carryin_clobercc_<optab>): Likewise.
6260 (*subsi3_carryin): Likewise.
6261 (*subsi3_carryin_const): Likewise.
6262 (*subsi3_carryin_compare): Likewise.
6263 (*subsi3_carryin_compare_const): Likewise.
6264 (*arm_subdi3): Likewise.
6265 (*thumb_subdi3): Likewise.
6266 (*subdi_di_zesidi): Likewise.
6267 (*subdi_di_sesidi): Likewise.
6268 (*subdi_zesidi_di): Likewise.
6269 (*subdi_sesidi_di): Likewise.
6270 (*subdi_zesidi_ze): Likewise.
6271 (thumb1_subsi3_insn): Likewise.
6272 (*arm_subsi3_insn): Likewise.
6273 (*anddi3_insn): Likewise.
6274 (*anddi_zesidi_di): Likewise.
6275 (*anddi_sesdi_di): Likewise.
6276 (*ne_zeroextracts): Likewise.
6277 (*ne_zeroextracts): Likewise.
6278 (*ite_ne_zeroextr): Likewise.
6279 (*ite_ne_zeroextr): Likewise.
6280 (*anddi_notdi_di): Likewise.
6281 (*anddi_notzesidi): Likewise.
6282 (*anddi_notsesidi): Likewise.
6283 (andsi_notsi_si): Likewise.
6284 (thumb1_bicsi3): Likewise.
6285 (*iordi3_insn): Likewise.
6286 (*iordi_zesidi_di): Likewise.
6287 (*iordi_sesidi_di): Likewise.
6288 (*thumb1_iorsi3_insn): Likewise.
6289 (*xordi3_insn): Likewise.
6290 (*xordi_zesidi_di): Likewise.
6291 (*xordi_sesidi_di): Likewise.
6292 (*arm_xorsi3): Likewise.
6293 (*andsi_iorsi3_no): Likewise.
6294 (*smax_0): Likewise.
6295 (*smax_m1): Likewise.
6296 (*arm_smax_insn): Likewise.
6297 (*smin_0): Likewise.
6298 (*arm_smin_insn): Likewise.
6299 (*arm_umaxsi3): Likewise.
6300 (*arm_uminsi3): Likewise.
6301 (*minmax_arithsi): Likewise.
6302 (*minmax_arithsi_): Likewise.
6303 (*satsi_<SAT:code>): Likewise.
6304 (arm_ashldi3_1bit): Likewise.
6305 (arm_ashrdi3_1bit): Likewise.
6306 (arm_lshrdi3_1bit): Likewise.
6307 (*arm_negdi2): Likewise.
6308 (*thumb1_negdi2): Likewise.
6309 (*arm_negsi2): Likewise.
6310 (*thumb1_negsi2): Likewise.
6311 (*negdi_extendsid): Likewise.
6312 (*negdi_zero_extend): Likewise.
6313 (*arm_abssi2): Likewise.
6314 (*thumb1_abssi2): Likewise.
6315 (*arm_neg_abssi2): Likewise.
6316 (*thumb1_neg_abss): Likewise.
6317 (one_cmpldi2): Likewise.
6318 (extend<mode>di2): Likewise.
6319 (*compareqi_eq0): Likewise.
6320 (*arm_extendhisi2addsi): Likewise.
6321 (*arm_movdi): Likewise.
6322 (*thumb1_movdi_insn): Likewise.
6323 (*arm_movt): Likewise.
6324 (*thumb1_movsi_insn): Likewise.
6325 (pic_add_dot_plus_four): Likewise.
6326 (pic_add_dot_plus_eight): Likewise.
6327 (tls_load_dot_plus_eight): Likewise.
6328 (*thumb1_movhi_insn): Likewise.
6329 (*thumb1_movsf_insn): Likewise.
6330 (*movdf_soft_insn): Likewise.
6331 (*thumb_movdf_insn): Likewise.
6332 (cbranchsi4_insn): Likewise.
6333 (cbranchsi4_scratch): Likewise.
6334 (*negated_cbranchsi4): Likewise.
6335 (*tbit_cbranch): Likewise.
6336 (*tlobits_cbranch): Likewise.
6337 (*tstsi3_cbranch): Likewise.
6338 (*cbranchne_decr1): Likewise.
6339 (*addsi3_cbranch): Likewise.
6340 (*addsi3_cbranch_scratch): Likewise.
6341 (*arm_cmpdi_insn): Likewise.
6342 (*arm_cmpdi_unsig): Likewise.
6343 (*arm_cmpdi_zero): Likewise.
6344 (*thumb_cmpdi_zero): Likewise.
6345 (*deleted_compare): Likewise.
6346 (*mov_scc): Likewise.
6347 (*mov_negscc): Likewise.
6348 (*mov_notscc): Likewise.
6349 (*cstoresi_eq0_thumb1_insn): Likewise.
6350 (cstoresi_nltu_thumb1): Likewise.
6351 (cstoresi_ltu_thu): Likewise.
6352 (thumb1_addsi3_addgeu): Likewise.
6353 (*arm_jump): Likewise.
6354 (*thumb_jump): Likewise.
6355 (*check_arch2): Likewise.
6356 (arm_casesi_internal): Likewise.
6357 (thumb1_casesi_dispatch): Likewise.
6358 (*arm_indirect_jump): Likewise.
6359 (*thumb1_indirect_jump): Likewise.
6361 (*and_scc): Likewise.
6362 (*ior_scc): Likewise.
6363 (*compare_scc): Likewise.
6364 (*cond_move): Likewise.
6365 (*cond_arith): Likewise.
6366 (*cond_sub): Likewise.
6367 (*cmp_ite0): Likewise.
6368 (*cmp_ite1): Likewise.
6369 (*cmp_and): Likewise.
6370 (*cmp_ior): Likewise.
6371 (*ior_scc_scc): Likewise.
6372 (*ior_scc_scc_cmp): Likewise.
6373 (*and_scc_scc): Likewise.
6374 (*and_scc_scc_cmp): Likewise.
6375 (*and_scc_scc_nod): Likewise.
6376 (*negscc): Likewise.
6377 (movcond_addsi): Likewise.
6378 (movcond): Likewise.
6379 (*ifcompare_plus_move): Likewise.
6380 (*if_plus_move): Likewise.
6381 (*ifcompare_move_plus): Likewise.
6382 (*if_move_plus): Likewise.
6383 (*ifcompare_arith_arith): Likewise.
6384 (*if_arith_arith): Likewise.
6385 (*ifcompare_arith_move): Likewise.
6386 (*if_arith_move): Likewise.
6387 (*ifcompare_move_arith): Likewise.
6388 (*if_move_arith): Likewise.
6389 (*ifcompare_move_not): Likewise.
6390 (*if_move_not): Likewise.
6391 (*ifcompare_not_move): Likewise.
6392 (*if_not_move): Likewise.
6393 (*ifcompare_shift_move): Likewise.
6394 (*if_shift_move): Likewise.
6395 (*ifcompare_move_shift): Likewise.
6396 (*if_move_shift): Likewise.
6397 (*ifcompare_shift_shift): Likewise.
6398 (*ifcompare_not_arith): Likewise.
6399 (*ifcompare_arith_not): Likewise.
6400 (*if_arith_not): Likewise.
6401 (*ifcompare_neg_move): Likewise.
6402 (*if_neg_move): Likewise.
6403 (*ifcompare_move_neg): Likewise.
6404 (*if_move_neg): Likewise.
6405 (prologue_thumb1_interwork): Likewise.
6406 (*cond_move_not): Likewise.
6407 (*sign_extract_onebit): Likewise.
6408 (*not_signextract_onebit): Likewise.
6409 (stack_tie): Likewise.
6410 (align_4): Likewise.
6411 (align_8): Likewise.
6412 (consttable_end): Likewise.
6413 (consttable_1): Likewise.
6414 (consttable_2): Likewise.
6415 (consttable_4): Likewise.
6416 (consttable_8): Likewise.
6417 (consttable_16): Likewise.
6418 (*thumb1_tablejump): Likewise.
6419 (prefetch): Likewise.
6420 (force_register_use): Likewise.
6421 (thumb_eh_return): Likewise.
6422 (load_tp_hard): Likewise.
6423 (load_tp_soft): Likewise.
6424 (tlscall): Likewise.
6425 (*arm_movtas_ze): Likewise.
6426 (*arm_rev): Likewise.
6427 (*arm_revsh): Likewise.
6428 (*arm_rev16): Likewise.
6429 * config/arm/thumb2.md
6430 (*thumb2_smaxsi3): Likewise.
6431 (*thumb2_sminsi3): Likewise.
6432 (*thumb32_umaxsi3): Likewise.
6433 (*thumb2_uminsi3): Likewise.
6434 (*thumb2_negdi2): Likewise.
6435 (*thumb2_abssi2): Likewise.
6436 (*thumb2_neg_abss): Likewise.
6437 (*thumb2_movsi_insn): Likewise.
6438 (tls_load_dot_plus_four): Likewise.
6439 (*thumb2_movhi_insn): Likewise.
6440 (*thumb2_mov_scc): Likewise.
6441 (*thumb2_mov_negs): Likewise.
6442 (*thumb2_mov_negs): Likewise.
6443 (*thumb2_mov_nots): Likewise.
6444 (*thumb2_mov_nots): Likewise.
6445 (*thumb2_movsicc_): Likewise.
6446 (*thumb2_movsfcc_soft_insn): Likewise.
6447 (*thumb2_indirect_jump): Likewise.
6448 (*thumb2_and_scc): Likewise.
6449 (*thumb2_ior_scc): Likewise.
6450 (*thumb2_ior_scc_strict_it): Likewise.
6451 (*thumb2_cond_move): Likewise.
6452 (*thumb2_cond_arith): Likewise.
6453 (*thumb2_cond_ari): Likewise.
6454 (*thumb2_cond_sub): Likewise.
6455 (*thumb2_negscc): Likewise.
6456 (*thumb2_movcond): Likewise.
6457 (thumb2_casesi_internal): Likewise.
6458 (thumb2_casesi_internal_pic): Likewise.
6459 (*thumb2_alusi3_short): Likewise.
6460 (*thumb2_mov<mode>_shortim): Likewise.
6461 (*thumb2_addsi_short): Likewise.
6462 (*thumb2_subsi_short): Likewise.
6463 (thumb2_addsi3_compare0): Likewise.
6464 (*thumb2_cbz): Likewise.
6465 (*thumb2_cbnz): Likewise.
6466 (*thumb2_one_cmplsi2_short): Likewise.
6467 (*thumb2_negsi2_short): Likewise.
6468 (*orsi_notsi_si): Likewise.
6469 * config/arm/arm1020e.md: Update with new attributes.
6470 * config/arm/arm1026ejs.md: Update with new attributes.
6471 * config/arm/arm1136jfs.md: Update with new attributes.
6472 * config/arm/arm926ejs.md: Update with new attributes.
6473 * config/arm/cortex-a15.md: Update with new attributes.
6474 * config/arm/cortex-a5.md: Update with new attributes.
6475 * config/arm/cortex-a53.md: Update with new attributes.
6476 * config/arm/cortex-a7.md: Update with new attributes.
6477 * config/arm/cortex-a8.md: Update with new attributes.
6478 * config/arm/cortex-a9.md: Update with new attributes.
6479 * config/arm/cortex-m4.md: Update with new attributes.
6480 * config/arm/cortex-r4.md: Update with new attributes.
6481 * config/arm/fa526.md: Update with new attributes.
6482 * config/arm/fa606te.md: Update with new attributes.
6483 * config/arm/fa626te.md: Update with new attributes.
6484 * config/arm/fa726te.md: Update with new attributes.
6486 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6488 * config/aarch64/aarch64-simd.md
6489 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
6490 <vwx> iterator to ensure correct register choice.
6491 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
6492 (aarch64_sqdmull_n<mode>): Likewise.
6493 (aarch64_sqdmull2_n<mode>_internal): Likewise.
6494 * config/aarch64/arm_neon.h
6495 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
6496 (vml<as><q>_n_<su>16): Likewise.
6497 (vml<as>l_high_lane<q>_<su>16): Likewise.
6498 (vml<as>l_high_n_<su>16): Likewise.
6499 (vml<as>l_lane<q>_<su>16): Likewise.
6500 (vml<as>l_n_<su>16): Likewise.
6501 (vmul<q>_lane<q>_<su>16): Likewise.
6502 (vmul<q>_n_<su>16): Likewise.
6503 (vmull_lane<q>_<su>16): Likewise.
6504 (vmull_n_<su>16): Likewise.
6505 (vmull_high_lane<q>_<su>16): Likewise.
6506 (vmull_high_n_<su>16): Likewise.
6507 (vqrdmulh<q>_n_s16): Likewise.
6509 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
6511 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
6512 have the correct lane parameter.
6514 2013-09-06 Richard Biener <rguenther@suse.de>
6516 * cfganal.c (control_dependences::~control_dependences):
6517 Properly free all of the vector.
6519 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
6522 * config/i386/i386.c (ix86_conditional_register_usage):
6523 Proper initialize extended SSE registers.
6525 2013-09-06 Jan Hubicka <jh@suse.cz>
6527 PR tree-optimization/58311
6528 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
6530 2013-09-06 Jan Hubicka <jh@suse.cz>
6532 * Makefile.in (tree-sra.o): Update dependencies.
6533 * tree-sra.c: Include ipa-utils.h
6534 (scan_function): Use recursive_call_p.
6535 (has_caller_p): New function.
6536 (cgraph_for_node_and_aliases): Count also callers of aliases.
6538 2013-09-06 Jan Hubicka <jh@suse.cz>
6541 * cgraph.h (symtab_semantically_equivalent_p): Declare.
6542 * tree-tailcall.c: Include ipa-utils.h.
6543 (find_tail_calls): Use it.
6544 * ipa-pure-const.c (check_call): Likewise.
6545 * ipa-utils.c (recursive_call_p): New function.
6546 * ipa-utils.h (recursive_call_p): Dclare.
6547 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
6548 (symtab_semantically_equivalent_p): New function.
6549 * Makefile.in (tree-tailcall.o): Update dependencies.
6551 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6553 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
6556 2013-09-06 Richard Biener <rguenther@suse.de>
6558 * lto-streamer.h (lto_global_var_decls): Remove.
6559 * Makefile.in (OBJS): Remove lto-symtab.o.
6560 (lto-symtab.o): Remove.
6561 (GTFILES): Remove lto-symtab.c
6562 * lto-symtab.c: Move to lto/
6564 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6566 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
6567 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
6568 (UNSPEC_FPINT_RINT): New constant definitions.
6569 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
6570 definition with 2 attributes.
6571 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
6572 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
6575 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6577 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
6578 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
6579 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
6580 attribute to "z196_alone".
6581 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
6584 2013-09-06 Richard Biener <rguenther@suse.de>
6586 * basic-block.h (class control_dependences): New.
6587 * tree-ssa-dce.c (control_dependence_map): Remove.
6589 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
6590 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
6591 find_pdom, find_control_dependence, find_all_control_dependences):
6593 (mark_control_dependent_edges_necessary,
6594 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
6595 tree_dce_done, perform_tree_ssa_dce): Adjust.
6596 * cfganal.c (set_control_dependence_map_bit,
6597 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
6598 find_all_control_dependences): Move from tree-ssa-dce.c and
6599 implement as methods of control_dependences class.
6600 (control_dependences::control_dependences): New.
6601 (control_dependences::~control_dependences): Likewise.
6602 (control_dependences::get_edges_dependent_on): Likewise.
6603 (control_dependences::get_edge): Likewise.
6605 2013-09-04 Jan Hubicka <jh@suse.cz>
6607 * tree.c (types_same_for_odr): Drop overactive check.
6608 * ipa-devirt.c (hash_type_name): Likewise.
6610 2013-09-04 Jan Hubicka <jh@suse.cz>
6612 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
6613 (analyze_functions): ... here.
6615 2013-09-04 Jan Hubicka <jh@suse.cz>
6618 * cgraphunit.c (analyze_functions): Clear AUX fields
6619 after processing; initialize assembler name has.
6621 2013-09-05 Jeff Law <law@redhat.com>
6623 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
6624 from thread_around_empty_block. Record threading path into PATH.
6625 Recurse if threading through the initial block is successful.
6626 (thread_across_edge): Corresponding changes to slightly simplify.
6628 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6630 * config/aarch64/aarch64.md
6631 (type): Remove frecpe, frecps, frecpx.
6632 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
6633 fix to be a TARGET_SIMD instruction.
6634 (aarch64_frecps): Remove.
6635 * config/aarch64/aarch64-simd.md
6636 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
6637 (aarch64_frecps<mode>): Handle all float/vector of float modes.
6639 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6640 Sofiane Naci <sofiane.naci@arm.com>
6642 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
6643 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
6644 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
6645 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
6646 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
6647 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
6648 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
6649 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
6650 into "clz, "rbit". Rename "shift" to "shift_imm".
6651 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
6652 changes. Update for attribute changes all occurrences of arlo_* and
6654 * config/arm/arm-fixed.md: Update for attribute changes
6655 all occurrences of arlo_* types.
6656 * config/arm/thumb2.md: Update for attribute changes all occurrences
6658 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
6659 (cortexa7_older_only): Likewise.
6660 (cortexa7_younger): Likewise.
6661 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
6662 (1020alu_shift_op): Likewise.
6663 (1020alu_shift_reg_op): Likewise.
6664 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
6665 (alu_shift_op): Likewise.
6666 (alu_shift_reg_op): Likewise.
6667 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
6668 (11_alu_shift_op): Likewise.
6669 (11_alu_shift_reg_op): Likewise.
6670 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
6671 (9_alu_shift_reg_op): Likewise.
6672 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
6674 (cortex_a15_alu_shift): Likewise.
6675 (cortex_a15_alu_shift_reg): Likewise.
6676 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
6678 (cortex_a5_alu_shift): Likewise.
6679 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
6681 (cortex_a53_alu_shift): Likewise.
6682 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
6684 (cortex_a7_alu_reg): Likewise.
6685 (cortex_a7_alu_shift): Likewise.
6686 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
6688 (cortex_a8_alu_shift): Likewise.
6689 (cortex_a8_alu_shift_reg): Likewise.
6690 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
6691 (cortex_a9_dp_shift): Likewise.
6692 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
6694 * config/arm/cortex-r4.md
6695 (cortex_r4_alu): Update for attribute changes.
6696 (cortex_r4_mov): Likewise.
6697 (cortex_r4_alu_shift_reg): Likewise.
6698 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
6699 (526_alu_shift_op): Likewise.
6700 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
6701 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
6702 (626te_alu_shift_op): Likewise.
6703 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
6704 (726te_alu_shift_op): Likewise.
6705 (726te_alu_shift_reg_op): Likewise.
6706 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
6707 (mp626_alu_shift_op): Likewise.
6708 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
6709 (pj4_alu_conds): Likewise.
6710 (pj4_shift): Likewise.
6711 (pj4_shift_conds): Likewise.
6712 (pj4_alu_shift): Likewise.
6713 (pj4_alu_shift_conds): Likewise.
6714 * config/aarch64/aarch64.md: Update for attribute change
6715 all occurrences of arlo_* and shift* types.
6717 2013-09-05 Mike Stump <mikestump@comcast.net>
6719 * tree.h: Move documentation for tree_function_decl to tree-core.h
6720 with the declaration.
6722 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
6725 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
6726 looking for widest mode.
6728 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
6730 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
6732 2013-09-05 Richard Biener <rguenther@suse.de>
6734 PR tree-optimization/58137
6735 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
6736 Do not create vectors of pointers.
6737 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
6738 types for the components of the vector initializer.
6739 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
6740 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
6742 2013-09-05 Martin Jambor <mjambor@suse.cz>
6744 * ipa-prop.c (remove_described_reference): Accept missing references,
6745 return false if that hppens, otherwise return true.
6746 (cgraph_node_for_jfunc): New function.
6747 (try_decrement_rdesc_refcount): Likewise.
6748 (try_make_edge_direct_simple_call): Use them.
6749 (ipa_edge_removal_hook): Remove references from rdescs.
6750 (ipa_edge_duplication_hook): Clone rdescs and their references
6751 when the new edge has the same caller as the old one.
6752 * cgraph.c (cgraph_resolve_speculation): Remove speculative
6753 reference before removing any edges.
6755 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
6757 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
6759 * thumb2.md (thumb2_storewb_parisi): New pattern.
6761 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
6763 * config/aarch64/aarch64-option-extensions.def: Add
6764 AARCH64_OPT_EXTENSION of 'crc'.
6765 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
6766 (AARCH64_ISA_CRC): Ditto.
6767 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
6768 description of the CRC extension.
6770 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
6772 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
6773 * config/rs6000/linux.h: Ditto.
6774 * alpha/linux.h: Ditto.
6775 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
6776 no_c99_libc_has_function.
6777 * config/c6x/uclinux-elf.h: Ditto.
6778 * config/lm32/uclinux-elf.h: Ditto.
6779 * config/m68k/uclinux.h: Ditto.
6780 * config/moxie/uclinux.h: Ditto.
6781 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
6782 (crisv32-*-linux*, cris-*-linux*): Ditto.
6783 * config/bfin/bfin.c: Include "tm_p.h".
6785 2013-09-05 Richard Biener <rguenther@suse.de>
6787 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
6788 check for a definition without a basic-block.
6790 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6791 Sofiane Naci <sofiane.naci@arm.com>
6793 * config/aarch64/aarch64.md
6794 (*movti_aarch64): Rename r_2_f and f_2_r.
6795 (*movsf_aarch64): Likewise.
6796 (*movdf_aarch64): Likewise.
6797 (*movtf_aarch64): Likewise.
6798 (aarch64_movdi_<mode>low): Likewise.
6799 (aarch64_movdi_<mode>high): Likewise.
6800 (aarch64_mov<mode>high_di): Likewise.
6801 (aarch64_mov<mode>low_di): Likewise.
6802 (aarch64_movtilow_tilow): Likewise.
6803 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
6804 values to config/arm/types.md
6805 (attribute "conds"): Update for attribute change.
6806 (anddi3_insn): Likewise.
6807 (iordi3_insn): Likewise.
6808 (xordi3_insn): Likewise.
6809 (one_cmpldi2): Likewise.
6810 * config/arm/types.md (type): Add Neon types.
6811 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
6812 use "type" attribute.
6813 (movmisalign<mode>_neon_store): Likewise.
6814 (movmisalign<mode>_neon_load): Likewise.
6815 (vec_set<mode>_internal): Likewise.
6816 (vec_setv2di_internal): Likewise.
6817 (vec_extract<mode>): Likewise.
6818 (vec_extractv2di): Likewise.
6819 (add<mode>3_neon): Likewise.
6820 (adddi3_neon): Likewise.
6821 (sub<mode>3_neon): Likewise.
6822 (subdi3_neon): Likewise.
6823 (mul<mode>3_neon): Likewise.
6824 (mul<mode>3add<mode>_neon): Likewise.
6825 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
6826 (fma<VCVTF:mode>4)): Likewise.
6827 (fma<VCVTF:mode>4_intrinsic): Likewise.
6828 (fmsub<VCVTF:mode>4)): Likewise.
6829 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
6830 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
6831 (ior<mode>3): Likewise.
6832 (and<mode>3): Likewise.
6833 (anddi3_neon): Likewise.
6834 (orn<mode>3_neon): Likewise.
6835 (orndi3_neon): Likewise.
6836 (bic<mode>3_neon): Likewise.
6837 (bicdi3_neon): Likewise.
6838 (xor<mode>3): Likewise.
6839 (one_cmpl<mode>2): Likewise.
6840 (abs<mode>2): Likewise.
6841 (neg<mode>2): Likewise.
6842 (umin<mode>3_neon): Likewise.
6843 (umax<mode>3_neon): Likewise.
6844 (smin<mode>3_neon): Likewise.
6845 (smax<mode>3_neon): Likewise.
6846 (vashl<mode>3): Likewise.
6847 (vashr<mode>3_imm): Likewise.
6848 (vlshr<mode>3_imm): Likewise.
6849 (ashl<mode>3_signed): Likewise.
6850 (ashl<mode>3_unsigned): Likewise.
6851 (neon_load_count): Likewise.
6852 (ashldi3_neon_noclobber): Likewise.
6853 (signed_shift_di3_neon): Likewise.
6854 (unsigned_shift_di3_neon): Likewise.
6855 (ashrdi3_neon_imm_noclobber): Likewise.
6856 (lshrdi3_neon_imm_noclobber): Likewise.
6857 (widen_ssum<mode>3): Likewise.
6858 (widen_usum<mode>3): Likewise.
6859 (quad_halves_<code>v4si): Likewise.
6860 (quad_halves_<code>v4sf): Likewise.
6861 (quad_halves_<code>v8hi): Likewise.
6862 (quad_halves_<code>v16qi): Likewise.
6863 (reduc_splus_v2di): Likewise.
6864 (neon_vpadd_internal<mode>): Likewise.
6865 (neon_vpsmin<mode>): Likewise.
6866 (neon_vpsmax<mode>): Likewise.
6867 (neon_vpumin<mode>): Likewise.
6868 (neon_vpumax<mode>): Likewise.
6869 (ss_add<mode>_neon): Likewise.
6870 (us_add<mode>_neon): Likewise.
6871 (ss_sub<mode>_neon): Likewise.
6872 (us_sub<mode>_neon): Likewise.
6873 (neon_vadd<mode>_unspec): Likewise.
6874 (neon_vaddl<mode>): Likewise.
6875 (neon_vaddw<mode>): Likewise.
6876 (neon_vhadd<mode>): Likewise.
6877 (neon_vqadd<mode>): Likewise.
6878 (neon_vaddhn<mode>): Likewise.
6879 (neon_vmul<mode>): Likewise.
6880 (neon_vmla<mode>): Likewise.
6881 (neon_vmlal<mode>): Likewise.
6882 (neon_vmls<mode>): Likewise.
6883 (neon_vmlsl<mode>): Likewise.
6884 (neon_vqdmulh<mode>): Likewise.
6885 (neon_vqdmlal<mode>): Likewise.
6886 (neon_vqdmlsl<mode>): Likewise.
6887 (neon_vmull<mode>): Likewise.
6888 (neon_vqdmull<mode>): Likewise.
6889 (neon_vsub<mode>_unspec): Likewise.
6890 (neon_vsubl<mode>): Likewise.
6891 (neon_vsubw<mode>): Likewise.
6892 (neon_vqsub<mode>): Likewise.
6893 (neon_vhsub<mode>): Likewise.
6894 (neon_vsubhn<mode>): Likewise.
6895 (neon_vceq<mode>): Likewise.
6896 (neon_vcge<mode>): Likewise.
6897 (neon_vcgeu<mode>): Likewise.
6898 (neon_vcgt<mode>): Likewise.
6899 (neon_vcgtu<mode>): Likewise.
6900 (neon_vcle<mode>): Likewise.
6901 (neon_vclt<mode>): Likewise.
6902 (neon_vcage<mode>): Likewise.
6903 (neon_vcagt<mode>): Likewise.
6904 (neon_vtst<mode>): Likewise.
6905 (neon_vabd<mode>): Likewise.
6906 (neon_vabdl<mode>): Likewise.
6907 (neon_vaba<mode>): Likewise.
6908 (neon_vabal<mode>): Likewise.
6909 (neon_vmax<mode>): Likewise.
6910 (neon_vmin<mode>): Likewise.
6911 (neon_vpaddl<mode>): Likewise.
6912 (neon_vpadal<mode>): Likewise.
6913 (neon_vpmax<mode>): Likewise.
6914 (neon_vpmin<mode>): Likewise.
6915 (neon_vrecps<mode>): Likewise.
6916 (neon_vrsqrts<mode>): Likewise.
6917 (neon_vqabs<mode>): Likewise.
6918 (neon_vqneg<mode>): Likewise.
6919 (neon_vcls<mode>): Likewise.
6920 (clz<mode>2): Likewise.
6921 (popcount<mode>2): Likewise.
6922 (neon_vrecpe): Likewise.
6923 (neon_vrsqrte): Likewise.
6924 (neon_vget_lane<mode>_sext_internal): Likewise.
6925 (neon_vget_lane<mode>_zext_internal): Likewise.
6926 (neon_vdup_n<mode>): Likewise.
6927 (neon_vdup_nv2di): Likewise.
6928 (neon_vdpu_lane<mode>_internal): Likewise.
6929 (neon_vswp<mode>): Likewise.
6930 (float<mode><V_cvtto>2): Likewise.
6931 (floatuns<mode><V_cvtto>2): Likewise.
6932 (fix_trunc<mode><V_cvtto>)2): Likewise
6933 (fixuns_trunc<mode><V_cvtto)2): Likewise.
6934 (neon_vcvt<mode>): Likewise.
6935 (neon_vcvtv4sfv4hf): Likewise.
6936 (neon_vcvtv4hfv4sf): Likewise.
6937 (neon_vcvt_n<mode>): Likewise.
6938 (neon_vmovn<mode>): Likewise.
6939 (neon_vqmovn<mode>): Likewise.
6940 (neon_vqmovun<mode>): Likewise.
6941 (neon_vmovl<mode>): Likewise.
6942 (neon_vmul_lane<mode>): Likewise.
6943 (neon_vmull_lane<mode>): Likewise.
6944 (neon_vqdmull_lane<mode>): Likewise.
6945 (neon_vqdmulh_lane<mode>): Likewise.
6946 (neon_vmla_lane<mode>): Likewise.
6947 (neon_vmlal_lane<mode>): Likewise.
6948 (neon_vqdmlal_lane<mode>): Likewise.
6949 (neon_vmls_lane<mode>): Likewise.
6950 (neon_vmlsl_lane<mode>): Likewise.
6951 (neon_vqdmlsl_lane<mode>): Likewise.
6952 (neon_vext<mode>): Likewise.
6953 (neon_vrev64<mode>): Likewise.
6954 (neon_vrev32<mode>): Likewise.
6955 (neon_vrev16<mode>): Likewise.
6956 (neon_vbsl<mode>_internal): Likewise.
6957 (neon_vshl<mode>): Likewise.
6958 (neon_vqshl<mode>): Likewise.
6959 (neon_vshr_n<mode>): Likewise.
6960 (neon_vshrn_n<mode>): Likewise.
6961 (neon_vqshrn_n<mode>): Likewise.
6962 (neon_vqshrun_n<mode>): Likewise.
6963 (neon_vshl_n<mode>): Likewise.
6964 (neon_vqshl_n<mode>): Likewise.
6965 (neon_vqshlu_n<mode>): Likewise.
6966 (neon_vshll_n<mode>): Likewise.
6967 (neon_vsra_n<mode>): Likewise.
6968 (neon_vsri_n<mode>): Likewise.
6969 (neon_vsli_n<mode>): Likewise.
6970 (neon_vtbl1v8qi): Likewise.
6971 (neon_vtbl2v8qi): Likewise.
6972 (neon_vtbl3v8qi): Likewise.
6973 (neon_vtbl4v8qi): Likewise.
6974 (neon_vtbx1v8qi): Likewise.
6975 (neon_vtbx2v8qi): Likewise.
6976 (neon_vtbx3v8qi): Likewise.
6977 (neon_vtbx4v8qi): Likewise.
6978 (neon_vtrn<mode>_internal): Likewise.
6979 (neon_vzip<mode>_internal): Likewise.
6980 (neon_vuzp<mode>_internal): Likewise.
6981 (neon_vld1<mode>): Likewise.
6982 (neon_vld1_lane<mode>): Likewise.
6983 (neon_vld1_dup<mode>): Likewise.
6984 (neon_vld1_dupv2di): Likewise.
6985 (neon_vst1<mode>): Likewise.
6986 (neon_vst1_lane<mode>): Likewise.
6987 (neon_vld2<mode>): Likewise.
6988 (neon_vld2_lane<mode>): Likewise.
6989 (neon_vld2_dup<mode>): Likewise.
6990 (neon_vst2<mode>): Likewise.
6991 (neon_vst2_lane<mode>): Likewise.
6992 (neon_vld3<mode>): Likewise.
6993 (neon_vld3qa<mode>): Likewise.
6994 (neon_vld3qb<mode>): Likewise.
6995 (neon_vld3_lane<mode>): Likewise.
6996 (neon_vld3_dup<mode>): Likewise.
6997 (neon_vst3<mode>): Likewise.
6998 (neon_vst3qa<mode>): Likewise.
6999 (neon_vst3qb<mode>): Likewise.
7000 (neon_vst3_lane<mode>): Likewise.
7001 (neon_vld4<mode>): Likewise.
7002 (neon_vld4qa<mode>): Likewise.
7003 (neon_vld4qb<mode>): Likewise.
7004 (neon_vld4_lane<mode>): Likewise.
7005 (neon_vld4_dup<mode>): Likewise.
7006 (neon_vst4<mode>): Likewise.
7007 (neon_vst4qa<mode>): Likewise.
7008 (neon_vst4qb<mode>): Likewise.
7009 (neon_vst4_lane<mode>): Likewise.
7010 (neon_vec_unpack<US>_lo_<mode>): Likewise.
7011 (neon_vec_unpack<US>_hi_<mode>): Likewise.
7012 (neon_vec_<US>mult_lo_<mode>): Likewise.
7013 (neon_vec_<US>mult_hi_<mode>): Likewise.
7014 (neon_vec_<US>shiftl_<mode>): Likewise.
7015 (neon_unpack<US>_<mode>): Likewise.
7016 (neon_vec_<US>mult_<mode>): Likewise.
7017 (vec_pack_trunc_<mode>): Likewise.
7018 (neon_vec_pack_trunk_<mode>): Likewise.
7019 (neon_vabd<mode>_2): Likewise.
7020 (neon_vabd<mode>_3): Likewise.
7021 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
7022 (thumb2_movsi_vfp): Likewise.
7023 (movdi_vfp): Likewise.
7024 (movdi_vfp_cortexa8): Likewise.
7025 (movhf_vfp_neon): Likewise.
7026 (movhf_vfp): Likewiwse.
7027 (movsf_vfp): Likewiwse.
7028 (thumb2_movsf_vfp): Likewiwse.
7029 (movdf_vfp): Likewise.
7030 (thumb2_movdf_vfp): Likewise.
7031 (movsfcc_vfp): Likewise.
7032 (thumb2_movsfcc_vfp): Likewise.
7033 (movdfcc_vfp): Likewise.
7034 (thumb2_movdfcc_vfp): Likewise.
7035 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
7036 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
7037 (v10_v2c): Likewise.
7038 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
7040 (cortex_a15_neon_int_2): Likewise.
7041 (cortex_a15_neon_int_3): Likewise.
7042 (cortex_a15_neon_int_4): Likewise.
7043 (cortex_a15_neon_int_5): Likewise.
7044 (cortex_a15_neon_vqneg_vqabs): Likewise.
7045 (cortex_a15_neon_vmov): Likewise.
7046 (cortex_a15_neon_vaba): Likewise.
7047 (cortex_a15_neon_vaba_qqq): Likewise.
7048 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7049 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7050 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
7051 scalar_64_32_long_scalar): Likewise.
7052 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7053 (cortex_a15_neon_mla_qqq_8_16): Likewise.
7054 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7055 lotype_qdd_64_32_long): Likewise.
7056 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7057 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7058 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
7059 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7060 (cortex_a15_neon_shift_1): Likewise.
7061 (cortex_a15_neon_shift_2): Likewise.
7062 (cortex_a15_neon_shift_3): Likewise.
7063 (cortex_a15_neon_vshl_ddd): Likewise.
7064 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7065 (cortex_a15_neon_vsra_vrsra): Likewise.
7066 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
7067 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
7068 (cortex_a15_neon_fp_vmul_ddd): Likewise.
7069 (cortex_a15_neon_fp_vmul_qqd): Likewise.
7070 (cortex_a15_neon_fp_vmla_ddd): Likewise.
7071 (cortex_a15_neon_fp_vmla_qqq): Likewise.
7072 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
7073 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
7074 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7075 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7076 (cortex_a15_neon_bp_simple): Likewise.
7077 (cortex_a15_neon_bp_2cycle): Likewise.
7078 (cortex_a15_neon_bp_3cycle): Likewise.
7079 (cortex_a15_neon_vld1_1_2_regs): Likewise.
7080 (cortex_a15_neon_vld1_3_4_regs): Likewise.
7081 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7082 (cortex_a15_neon_vld2_4_regs): Likewise.
7083 (cortex_a15_neon_vld3_vld4): Likewise.
7084 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7085 (cortex_a15_neon_vst1_3_4_regs): Likewise.
7086 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
7087 (cortex_a15_neon_vst3_vst4): Likewise.
7088 (cortex_a15_neon_vld1_vld2_lane): Likewise.
7089 (cortex_a15_neon_vld3_vld4_lane" 10
7090 (cortex_a15_neon_vst1_vst2_lane): Likewise.
7091 (cortex_a15_neon_vst3_vst4_lane): Likewise.
7092 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
7093 (cortex_a15_neon_ldm_2): Likewise.0
7094 (cortex_a15_neon_stm_2): Likewise.
7095 (cortex_a15_neon_mcr): Likewise.
7096 (cortex_a15_neon_mcr_2_mcrr): Likewise.
7097 (cortex_a15_neon_mrc): Likewise.
7098 (cortex_a15_neon_mrrc): Likewise.
7099 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
7101 (cortex_a15_alu_shift): Likewise.
7102 (cortex_a15_alu_shift_reg): Likewise.
7103 (cortex_a15_mult32): Likewise.
7104 (cortex_a15_mult64): Likewise.
7105 (cortex_a15_block): Likewise.
7106 (cortex_a15_branch): Likewise.
7107 (cortex_a15_load1): Likewise.
7108 (cortex_a15_load3): Likewise.
7109 (cortex_a15_store1): Likewise.
7110 (cortex_a15_store3): Likewise.
7111 (cortex_a15_call): Likewise.
7112 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
7113 (cortex_a5_f2r): Likewise.
7114 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
7116 (cortex_a53_f2r): Likewise.
7117 * config/arm/cortex-a7.md
7118 (cortex_a7_branch): Update for attribute change.
7119 (cortex_a7_call): Likewise.
7120 (cortex_a7_alu_imm): Likewise.
7121 (cortex_a7_alu_reg): Likewise.
7122 (cortex_a7_alu_shift): Likewise.
7123 (cortex_a7_mul): Likewise.
7124 (cortex_a7_load1): Likewise.
7125 (cortex_a7_store1): Likewise.
7126 (cortex_a7_load2): Likewise.
7127 (cortex_a7_store2): Likewise.
7128 (cortex_a7_load3): Likewise.
7129 (cortex_a7_store3): Likewise.
7130 (cortex_a7_load4): Likewise.
7131 (cortex_a7_store4): Likewise.
7132 (cortex_a7_fpalu): Likewise.
7133 (cortex_a7_fconst): Likewise.
7134 (cortex_a7_fpmuls): Likewise.
7135 (cortex_a7_neon_mul): Likewise.
7136 (cortex_a7_fpmacs): Likewise.
7137 (cortex_a7_neon_mla: Likewise.
7138 (cortex_a7_fpmuld: Likewise.
7139 (cortex_a7_fpmacd: Likewise.
7140 (cortex_a7_fpfmad: Likewise.
7141 (cortex_a7_fdivs: Likewise.
7142 (cortex_a7_fdivd: Likewise.
7143 (cortex_a7_r2f: Likewise.
7144 (cortex_a7_f2r: Likewise.
7145 (cortex_a7_f_flags: Likewise.
7146 (cortex_a7_f_loads: Likewise.
7147 (cortex_a7_f_loadd: Likewise.
7148 (cortex_a7_f_stores: Likewise.
7149 (cortex_a7_f_stored: Likewise.
7150 (cortex_a7_neon): Likewise.
7151 * config/arm/cortex-a8-neon.md
7152 (cortex_a8_neon_mrc): Update for attribute change.
7153 (cortex_a8_neon_mrrc): Likewise.
7154 (cortex_a8_neon_int_1): Likewise.
7155 (cortex_a8_neon_int_2): Likewise.
7156 (cortex_a8_neon_int_3): Likewise.
7157 (cortex_a8_neon_int_4): Likewise.
7158 (cortex_a8_neon_int_5): Likewise.
7159 (cortex_a8_neon_vqneg_vqabs): Likewise.
7160 (cortex_a8_neon_vmov): Likewise.
7161 (cortex_a8_neon_vaba): Likewise.
7162 (cortex_a8_neon_vaba_qqq): Likewise.
7163 (cortex_a8_neon_vsma): Likewise.
7164 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7165 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7166 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7167 long_scalar): Likewise.
7168 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7169 (cortex_a8_neon_mla_qqq_8_16): Likewise.
7170 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7171 long_scalar_qdd_64_32_long): Likewise.
7172 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7173 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7174 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
7175 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7176 (cortex_a8_neon_shift_1): Likewise.
7177 (cortex_a8_neon_shift_2): Likewise.
7178 (cortex_a8_neon_shift_3): Likewise.
7179 (cortex_a8_neon_vshl_ddd): Likewise.
7180 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7181 (cortex_a8_neon_vsra_vrsra): Likewise.
7182 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
7183 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
7184 (cortex_a8_neon_fp_vsum): Likewise.
7185 (cortex_a8_neon_fp_vmul_ddd): Likewise.
7186 (cortex_a8_neon_fp_vmul_qqd): Likewise.
7187 (cortex_a8_neon_fp_vmla_ddd): Likewise.
7188 (cortex_a8_neon_fp_vmla_qqq): Likewise.
7189 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
7190 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
7191 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7192 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7193 (cortex_a8_neon_bp_simple): Likewise.
7194 (cortex_a8_neon_bp_2cycle): Likewise.
7195 (cortex_a8_neon_bp_3cycle): Likewise.
7196 (cortex_a8_neon_ldr): Likewise.
7197 (cortex_a8_neon_str): Likewise.
7198 (cortex_a8_neon_vld1_1_2_regs): Likewise.
7199 (cortex_a8_neon_vld1_3_4_regs): Likewise.
7200 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7201 (cortex_a8_neon_vld2_4_regs): Likewise.
7202 (cortex_a8_neon_vld3_vld4): Likewise.
7203 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7204 (cortex_a8_neon_vst1_3_4_regs): Likewise.
7205 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
7206 (cortex_a8_neon_vst3_vst4): Likewise.
7207 (cortex_a8_neon_vld1_vld2_lane): Likewise.
7208 (cortex_a8_neon_vld3_vld4_lane): Likewise.
7209 (cortex_a8_neon_vst1_vst2_lane): Likewise.
7210 (cortex_a8_neon_vst3_vst4_lane): Likewise.
7211 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
7212 (cortex_a8_neon_mcr): Likewise.
7213 (cortex_a8_neon_mcr_2_mcrr): Likewise.
7214 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
7215 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
7217 (ca9_neon_mrrc): Likewise.
7218 (cortex_a9_neon_int_1): Likewise.
7219 (cortex_a9_neon_int_2): Likewise.
7220 (cortex_a9_neon_int_3): Likewise.
7221 (cortex_a9_neon_int_4): Likewise.
7222 (cortex_a9_neon_int_5): Likewise.
7223 (cortex_a9_neon_vqneg_vqabs): Likewise.
7224 (cortex_a9_neon_vmov): Likewise.
7225 (cortex_a9_neon_vaba): Likewise.
7226 (cortex_a9_neon_vaba_qqq): Likewise.
7227 (cortex_a9_neon_vsma): Likewise.
7228 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7229 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7230 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7231 long_scalar): Likewise.
7232 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7233 (cortex_a9_neon_mla_qqq_8_16): Likewise.
7234 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7235 long_scalar_qdd_64_32_long): Likewise.
7236 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7237 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7238 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
7239 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7240 (cortex_a9_neon_shift_1): Likewise.
7241 (cortex_a9_neon_shift_2): Likewise.
7242 (cortex_a9_neon_shift_3): Likewise.
7243 (cortex_a9_neon_vshl_ddd): Likewise.
7244 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7245 (cortex_a9_neon_vsra_vrsra): Likewise.
7246 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
7247 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
7248 (cortex_a9_neon_fp_vsum): Likewise.
7249 (cortex_a9_neon_fp_vmul_ddd): Likewise.
7250 (cortex_a9_neon_fp_vmul_qqd): Likewise.
7251 (cortex_a9_neon_fp_vmla_ddd): Likewise.
7252 (cortex_a9_neon_fp_vmla_qqq): Likewise.
7253 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
7254 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
7255 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7256 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7257 (cortex_a9_neon_bp_simple): Likewise.
7258 (cortex_a9_neon_bp_2cycle): Likewise.
7259 (cortex_a9_neon_bp_3cycle): Likewise.
7260 (cortex_a9_neon_ldr): Likewise.
7261 (cortex_a9_neon_str): Likewise.
7262 (cortex_a9_neon_vld1_1_2_regs): Likewise.
7263 (cortex_a9_neon_vld1_3_4_regs): Likewise.
7264 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7265 (cortex_a9_neon_vld2_4_regs): Likewise.
7266 (cortex_a9_neon_vld3_vld4): Likewise.
7267 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7268 (cortex_a9_neon_vst1_3_4_regs): Likewise.
7269 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
7270 (cortex_a9_neon_vst3_vst4): Likewise.
7271 (cortex_a9_neon_vld1_vld2_lane): Likewise.
7272 (cortex_a9_neon_vld3_vld4_lane): Likewise.
7273 (cortex_a9_neon_vst1_vst2_lane): Likewise.
7274 (cortex_a9_neon_vst3_vst4_lane): Likewise.
7275 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
7276 (cortex_a9_neon_mcr): Likewise.
7277 (cortex_a9_neon_mcr_2_mcrr): Likewise.
7278 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
7279 (cortex_a9_fps): Likewise.
7280 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
7282 (cortex_m4_fmuls): Likewise.
7283 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
7285 (cortex_r4_mrc): Likewise.
7286 * config/arm/iterators.md: Update comment referring to neon_type.
7287 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
7288 (iwmmxt_movsi_insn): Likewise.
7289 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
7291 (pj4_core_to_vfp): Likewise.
7292 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
7294 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
7295 (vfp_fstore): Likewise.
7296 * doc/md.texi: Change references to neon_type to refer to type.
7298 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7300 * tree.h (DECL_BUILT_IN): Fix typo in comment.
7302 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
7304 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
7307 2013-09-04 Easwaran Raman <eraman@google.com>
7310 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
7311 (build_and_add_sum): Use it.
7312 (appears_later_in_bb): Simplify code.
7314 2013-09-04 Teresa Johnson <tejohnson@google.com>
7316 * dumpfile.c (dump_finish): Don't close stderr/stdout.
7318 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
7320 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
7322 2013-09-04 Jan Hubicka <jh@suse.cz>
7324 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
7325 * ipa-devirt.c: Include diganostic.h
7326 (odr_type_d): Add types and types_set.
7327 (hash_type_name): Work for types with vtables during LTO.
7328 (odr_hasher::remove): Fix comment; destroy types_set.
7329 (add_type_duplicate): New function,
7330 (get_odr_type): Use it.
7331 (dump_type_inheritance_graph): Dump type duplicates.
7332 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
7334 * tree.c (types_same_for_odr): Give exact answers on types with
7337 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7339 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
7340 explaining their differences.
7342 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
7344 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
7346 2013-09-03 Jeff Law <law@redhat.com>
7348 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
7349 when not threading through a joiner block. Pass joiner/no joiner
7350 state to register_jump_thread.
7351 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
7352 state from argument rather than implying on path length.
7353 Dump the entire jump thread path into debugging dump.
7354 * tree-flow.h (register_jump_thread): Update prototype.
7356 2013-08-29 Xinliang David Li <davidxl@google.com>
7358 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7359 Remove a trivial gcc_assert.
7361 2013-08-29 Xinliang David Li <davidxl@google.com>
7363 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
7364 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
7365 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7366 Delay base decl alignment adjustment.
7367 * tree-vectorizer.c (vect_destroy_datarefs): New function.
7368 * tree-vectorizer.h: New data structure.
7369 (set_dr_misalignment): New function.
7370 (dr_misalignment): Ditto.
7371 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
7372 (vectorizable_load): Ditto.
7373 (ensure_base_align): New function.
7374 (vectorize_loops): Add dbg_cnt support.
7375 (execute_vect_slp): Ditto.
7376 * dbgcnt.def: New debug counter.
7377 * Makefile: New dependency.
7379 2013-09-03 Meador Inge <meadori@codesourcery.com>
7383 2013-08-30 Meador Inge <meadori@codesourcery.com>
7385 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7387 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
7389 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
7390 function descriptor.
7392 2013-09-03 Richard Biener <rguenther@suse.de>
7394 * tree-affine.c (add_elt_to_tree): Fix association issue,
7395 avoid useless converts and make sure to always return a
7396 properly typed result.
7398 2013-09-03 Richard Biener <rguenther@suse.de>
7401 * fold-const.c (negate_expr_p): Fix division case.
7402 (negate_expr): Likewise.
7404 2013-09-03 Richard Biener <rguenther@suse.de>
7407 * tree-streamer-out.c: Include tm.h.
7408 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
7410 2013-09-03 Jan Hubicka <jh@suse.cz>
7412 * tree-profile.c (tree_profiling): Cleanup CFG when done.
7414 2013-09-03 Alan Modra <amodra@gmail.com>
7416 * config.gcc (powerpc*-*-linux*): Add support for little-endian
7417 multilibs to big-endian target and vice versa.
7418 * config/rs6000/t-linux64: Use := assignment on all vars.
7419 (MULTILIB_EXTRA_OPTS): Remove fPIC.
7420 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
7421 * config/rs6000/t-linux64le: New file.
7422 * config/rs6000/t-linux64bele: New file.
7423 * config/rs6000/t-linux64lebe: New file.
7425 2013-09-02 Jan Hubicka <jh@suse.cz>
7427 * ipa-inline-transform.c (inline_transform): Do not
7428 optimize_inline_calls when not optimizing.
7430 2013-09-02 Jan Hubicka <jh@suse.cz>
7432 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
7433 duplicated nodes for assembler names.
7434 * symtab.c (symtab_unregister_node): Do not attempt to unlink
7435 hard registers from assembler name hash.
7437 2013-09-02 Jan Hubicka <jh@suse.cz>
7439 * ipa-split.c (execute_split_functions): Split externally visible
7440 functions called once.
7442 2013-09-02 Martin Jambor <mjambor@suse.cz>
7445 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
7446 linked list. When finding the correct duplicate, also consider also
7447 the caller in additon to its inlined_to node.
7449 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
7451 * config/aarch64/aarch64-simd-builtins.def
7452 (dup_lane_scalar): Remove.
7453 * config/aarch64/aarch64-simd.md
7454 (aarch64_simd_dup): Add 'w->w' alternative.
7455 (aarch64_dup_lane<mode>): Allow for VALL.
7456 (aarch64_dup_lane_scalar<mode>): Remove.
7457 (aarch64_dup_lane_<vswap_width_name><mode>): New.
7458 (aarch64_get_lane_signed<mode>): Add w->w altenative.
7459 (aarch64_get_lane_unsigned<mode>): Likewise.
7460 (aarch64_get_lane<mode>): Likewise.
7461 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
7462 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
7463 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
7464 (VCON): Change container of V2SF.
7465 (vswap_width_name): Likewise.
7466 * config/aarch64/arm_neon.h
7467 (__aarch64_vdup_lane_any): New.
7468 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7469 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
7470 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7472 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
7475 * expr.c (emit_move_complex): Do not move complex FP values as parts if
7476 the source or the destination is a single hard register.
7478 2013-09-02 Richard Biener <rguenther@suse.de>
7481 * tree-scalar-evolution.c (instantiate_scev_name): Allow
7484 2013-09-02 Richard Biener <rguenther@suse.de>
7486 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
7487 arithmetic to sizetype.
7489 2013-09-02 Bin Cheng <bin.cheng@arm.com>
7491 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
7492 Find auto-increment use both before and after candidate.
7494 2013-09-02 Marek Polacek <polacek@redhat.com>
7496 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
7498 2013-09-01 Jan Hubicka <jh@suse.cz>
7500 * Makefile.in: Add ipa-profile.o
7501 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
7502 * cgraph.c (struct cgraph_propagate_frequency_data,
7503 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
7504 ipa-profile.c; replace cgraph_ by ipa_ prefix.
7505 * cgraph.h (cgraph_propagate_frequency): Remove.
7506 * ipa-inline-analysis.c: Include ipa-utils.h;
7507 drop duplicated cfgloop.h.
7508 (inline_update_callee_summaries): Update.
7509 * ipa-profile.c: New file.
7510 * ipa-utils.h (ipa_propagate_frequency): Declare.
7511 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
7512 data-streamer.h, value-prof.h.
7513 (symtab_remove_unreachable_nodes): Update profile.
7514 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
7515 account_time_size, cmp_counts, dump_histogram,
7516 ipa_profile_generate_summary, ipa_profile_write_summary,
7517 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
7518 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
7519 Move to ipa-profile.c.
7521 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
7523 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
7525 2013-09-01 Jan Hubicka <jh@suse.cz>
7527 * common.opt (fdevirtualize-speculatively): New function.
7528 * invoke.texi (fdevirtualize-speculatively): Document.
7529 * ipa-devirt.c: Include ipa-inline.h
7530 (likely_target_p): New function.
7531 (ipa_devirt): New function.
7532 (gate_ipa_devirt): New function.
7533 (pass_data_ipa_devirt): New static var.
7534 (pass_ipa_devirt): Likewise.
7535 (make_pass_ipa_devirt): New function.
7536 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
7537 (common_handle_option): Disable devirtualization when
7538 value range profiling is available.
7539 * passes.def (pass_ipa_devirt): Add.
7540 * timever.def (TV_IPA_DEVIRT): New timevar.
7541 * tree-pass.h (make_pass_ipa_devirt):
7543 2013-09-01 Iain Sandoe <iain@codesourcery.com>
7545 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
7546 include sanitize(undefined).
7548 2013-08-31 Diego Novillo <dnovillo@google.com>
7550 * Makefile.in (TREE_CORE_H): Define.
7552 (GTFILES): Add tree-core.h.
7553 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
7555 * tree-core.h: New file.
7556 Move all data structures, enum, typedefs, global
7557 declarations and constants from ...
7560 2013-08-31 Jan Hubicka <jh@suse.cz>
7562 * bulitins.c (expand_builtin): Do not early exit for gcov
7563 instrumented functions.
7565 2013-08-31 Marek Polacek <polacek@redhat.com>
7567 * ubsan.c: Include tm_p.h.
7569 2013-08-31 Jan Hubicka <jh@suse.cz>
7571 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
7574 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
7575 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
7576 * tree-cfg.c (verify_gimple_label): ... here.
7577 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
7578 (ipa_merge_profiles): New function.
7579 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
7580 (lto_input_function_body): Likewise.
7581 * ipa-utils.h (ipa_merge_profiles): Declare.
7582 * lto-streamer.h (lto_input_function_body): Update prototype.
7583 (emit_label_in_global_context_p): Remove.
7584 * lto-symtab.c: Include ipa-utils.h
7585 (lto_cgraph_replace_node): Use ipa_merge_profiles.
7587 2013-08-31 Jan Hubicka <jh@suse.cz>
7589 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
7591 2013-08-31 Jan Hubicka <jh@suse.cz>
7593 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
7595 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
7597 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
7598 "cmp" RTX before signed_comparison_operator check to account
7601 2013-08-30 Jan Hubicka <jh@suse.cz>
7603 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
7604 (detect_type_change_1): Rename to ...
7605 (detect_type_change): ... this one; early return on non-polymorphic
7607 (detect_type_change_ssa): Add comp_type parameter; update
7608 use of detect_type_change.
7609 (compute_complex_assign_jump_func): Add param_type parameter;
7610 update use of detect_type_change_ssa.
7611 (compute_complex_ancestor_jump_func): Likewise.
7612 (ipa_get_callee_param_type): New function.
7613 (ipa_compute_jump_functions_for_edge): Compute parameter type;
7614 update calls to the jump function computation functions.
7616 2013-08-30 Teresa Johnson <tejohnson@google.com>
7617 Steven Bosscher <steven@gcc.gnu.org>
7619 * cfgrtl.c (fixup_new_cold_bb): New routine.
7620 (commit_edge_insertions): Invoke fixup_partitions.
7621 (find_partition_fixes): New routine.
7622 (fixup_partitions): Ditto.
7623 (verify_hot_cold_block_grouping): Update comments.
7624 (rtl_verify_edges): Invoke find_partition_fixes.
7625 (rtl_verify_bb_pointers): Update comments.
7626 (rtl_verify_bb_layout): Ditto.
7627 * basic-block.h (probably_never_executed_edge_p): Declare.
7628 (fixup_partitions): Ditto.
7629 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
7630 * bb-reorder.c (sanitize_hot_paths): New function.
7631 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
7633 * predict.c (probably_never_executed_edge_p): New routine.
7634 * cfg.c (check_bb_profile): Add partition insanity warnings.
7636 2013-08-30 Meador Inge <meadori@codesourcery.com>
7638 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7640 2013-08-30 Marek Polacek <polacek@redhat.com>
7642 * Makefile.in (ubsan.o): Add.
7643 (c-family/c-ubsan.o): Add.
7644 (builtins.o): Add ubsan.h dependency.
7645 * ubsan.h: New file.
7646 * ubsan.c: New file.
7647 * common.opt: Add -fsanitize=undefined option.
7648 (flag_sanitize): Add variable.
7649 (fsanitize=): Add option. Add Driver.
7650 (fsanitize=thread): Remove option.
7651 (fsanitize=address): Likewise.
7652 (static-libubsan): New option.
7653 * doc/invoke.texi: Document the new flag and -static-libubsan.
7654 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
7655 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
7656 * builtin-attrs.def (ATTR_COLD): Define.
7657 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7658 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
7659 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
7660 * flag-types.h (sanitize_code): New enum.
7661 * opts.c (common_handle_option): Parse command line arguments
7662 of -fsanitize=. Add -fsanitize=unreachable option.
7663 * varasm.c (get_variable_section): Adjust.
7664 (assemble_noswitch_variable): Likewise.
7665 (assemble_variable): Likewise.
7666 (output_constant_def_contents): Likewise.
7667 (categorize_decl_for_section): Likewise.
7668 (place_block_symbol): Likewise.
7669 (output_object_block): Likewise.
7670 * builtins.def: Likewise.
7671 * toplev.c (compile_file): Likewise.
7672 (process_options): Likewise.
7673 * cppbuiltin.c: Likewise.
7674 * tsan.c (tsan_pass): Likewise.
7675 (tsan_gate): Likewise.
7676 (tsan_gate_O0): Likewise.
7677 * cfgexpand.c (partition_stack_vars): Likewise.
7678 (expand_stack_vars): Likewise.
7679 (defer_stack_allocation): Likewise.
7680 (expand_used_vars): Likewise.
7681 * cfgcleanup.c (old_insns_match_p): Likewise.
7682 * asan.c (asan_finish_file): Likewise.
7683 (asan_instrument): Likewise.
7684 (gate_asan): Likewise.
7685 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
7686 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7687 (asan_global_struct): Use pointer_sized_int_node instead
7688 calling build_nonstandard_integer_type.
7689 (initialize_sanitizer_builtins): Likewise.
7690 (asan_finish_file): Likewise.
7691 * gcc.c: Document %{%:function(args):X}.
7692 (static_spec_functions): Add sanitize.
7693 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
7694 store funcval != NULL there.
7695 (do_spec_1): Adjust handle_spec_function caller.
7696 (handle_braces): Allow %:function(args) as condition.
7697 (sanitize_spec_function): New function.
7698 (ADD_STATIC_LIBUBSAN_LIBS): Define.
7699 (LIBUBSAN_SPEC): Likewise.
7700 (LIBUBSAN_EARLY_SPEC): Likewise.
7701 (SANITIZER_SPEC): Handle libubsan.
7702 (SANITIZER_EARLY_SPEC): Likewise.
7703 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
7704 instead of fsanitize=address.
7705 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
7706 instead of fsanitize=address*.
7707 * builtins.c: Include ubsan.h.
7708 (fold_builtin_0): Instrument __builtin_unreachable.
7709 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
7710 instead of flag_asan.
7711 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
7712 (pointer_sized_int_node): Define.
7713 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
7715 2013-08-30 Mike Stump <mikestump@comcast.net>
7717 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
7720 2013-08-29 Jan Hubicka <jh@suse.cz>
7722 * cgraph.c (cgraph_function_body_availability): Handle weakref
7724 * passes.def: Remove pass_fixup_cfg.
7725 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
7726 track when we need to remove functions.
7727 (gate_ipa_inline): Execute inlining always; add comment why.
7728 (pass_data_ipa_inline): Remove TODO_remove_functions.
7729 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
7730 do not produce summaries.
7731 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
7732 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
7733 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
7734 constant pool and vtable.
7736 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
7738 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
7739 New arm_neon.h's internal macros to specify 64-bit constants.
7740 Avoid using stdint.h's macros.
7742 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
7744 * recog.c (verify_changes): Verify that changes[i].old is non-zero
7745 before applying REG_P.
7747 2013-08-30 Jakub Jelinek <jakub@redhat.com>
7749 PR tree-optimization/58277
7750 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
7751 after seeing too many stmts with vdef in between dombb and current
7752 bb, invalidate everything.
7754 2013-08-30 Richard Biener <rguenther@suse.de>
7756 * fold-const.c (fold_single_bit_test): Fix overflow test.
7758 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
7760 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
7761 and which can live in a register, always retrieve the value on entry.
7762 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
7763 passed by invisible reference specially.
7764 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
7765 (vt_add_function_parameter): Correctly deal with a parameter passed by
7766 invisible reference.
7768 2013-08-30 Jan Hubicka <jh@suse.cz>
7770 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
7772 2013-08-30 Richard Biener <rguenther@suse.de>
7774 PR tree-optimization/58228
7775 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
7776 allow invariant loads in nested loop vectorization.
7778 2013-08-30 Richard Biener <rguenther@suse.de>
7780 PR tree-optimization/58223
7781 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
7782 (has_anti_or_output_dependence): ... this and adjust to also
7783 look for output dependences.
7784 (mark_nodes_having_upstream_mem_writes): Adjust.
7785 (rdg_flag_uses): Likewise.
7787 2013-08-30 Richard Biener <rguenther@suse.de>
7789 PR tree-optimization/58010
7790 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
7791 assert that we have a loop-closed PHI.
7793 2013-08-29 Jan Hubicka <jh@suse.cz>
7795 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
7796 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
7797 * lto-section-in.c (lto_free_function_in_decl_state): New function.
7798 (lto_free_function_in_decl_state_for_node): New function.
7800 2013-08-29 Xinliang David Li <davidxl@google.com>
7802 * loop-unroll.c (report_unroll_peel): Minor message change.
7803 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
7804 Emit alignment peeling message with default -fopt-info.
7805 (vect_loop_versioning): Emit loop version info message.
7806 * tree-vectorizer.c (vectorize_loops): Minor message change.
7807 (execute_vect_slp): Ditto.
7809 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
7811 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
7812 of the clone from the DECL_NAME of the original function.
7814 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
7816 * passes.c (register_pass): Add overload.
7817 * tree-pass.h (register_pass): Forward declare it. Add comment.
7819 2013-08-29 Jan Hubicka <jh@suse.cz>
7821 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
7822 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
7823 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
7824 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7825 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
7826 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
7827 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7828 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
7830 2013-08-29 Teresa Johnson <tejohnson@google.com>
7832 * dumpfile.c (dump_loc): Output column number.
7833 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
7834 * doc/invoke.texi: Document optall -fopt-info flag.
7835 * profile.c (read_profile_edge_counts): Use new dump framework.
7836 (compute_branch_probabilities): Ditto.
7837 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
7838 when pass not in any opt group.
7839 * pass_manager.h (pass_manager::get_pass_profile): New method.
7840 * value-prof.c (check_counter): Use new dump framework.
7841 (check_ic_target): Ditto.
7842 * coverage.c (get_coverage_counts): Ditto.
7843 (coverage_init): Setup new dump framework.
7845 2013-08-29 Richard Biener <rguenther@suse.de>
7847 PR tree-optimization/58246
7848 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
7849 handle the dominance check inside a basic-block.
7851 2013-08-29 Richard Biener <rguenther@suse.de>
7854 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
7855 of default defs that appear in abnormal PHI nodes.
7857 2013-08-29 Richard Biener <rguenther@suse.de>
7859 PR tree-optimization/57685
7860 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
7861 single-use operands to avoid exponential complexity.
7863 2013-08-28 Dehao Chen <dehao@google.com>
7865 * ipa-inline.c (edge_badness): Fix integer underflow.
7867 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
7869 * gtm-builtins.def (_ITM_free): Declare leaf.
7871 2013-08-28 Jakub Jelinek <jakub@redhat.com>
7874 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
7875 (*tls_local_dynamic_base_64_largepic): Likewise.
7876 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
7877 Remove predicate from call operand.
7878 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
7879 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
7881 2013-08-28 Jeff Law <law@redhat.com>
7883 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
7884 checks for the number of predecessors and successors allowed.
7885 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
7886 which require copying a joiner block if there is a request which
7887 is a subpath that requires no joiner block copying.
7889 2013-08-28 Jan Hubicka <jh@suse.cz>
7891 * lto-streamer-out.c (DFS_write_tree_body): Drop
7892 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
7893 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
7894 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
7895 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
7896 Do not read DECL_ERROR_ISSUED.
7897 (unpack_ts_decl_with_vis_value_fields): Do not read
7899 (lto_input_ts_binfo_tree_pointers): Do not read
7900 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
7901 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
7902 write DECL_ERROR_ISSUED..
7903 (pack_ts_decl_with_vis_value_fields): Do not write
7905 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
7906 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
7907 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
7908 * tree.h (tree_decl_common): Update comment.
7909 (DECL_ERROR_ISSUED): Remove.
7911 2013-08-28 Jakub Jelinek <jakub@redhat.com>
7914 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
7916 2013-08-28 Jan Hubicka <jh@suse.cz>
7918 * builtins.def (free): Declare leaf.
7920 2013-08-27 David Malcolm <dmalcolm@redhat.com>
7923 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
7924 * configure: Regenerate.
7926 2013-08-27 Martin Jambor <mjambor@suse.cz>
7928 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
7929 (ipa_ancestor_jf_data): Likewise.
7930 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
7931 (ipa_get_jf_pass_through_type_preserved): New function.
7932 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
7933 (ipa_get_jf_ancestor_type_preserved): New function.
7934 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
7935 (ipa_get_jf_ancestor_result): Likewise.
7936 (propagate_vals_accross_pass_through): Use
7937 ipa_get_jf_pass_through_result to do all the value mappings.
7938 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
7939 type_preserved flag.
7940 (ipa_set_jf_cst_copy): New function.
7941 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
7942 (ipa_set_jf_arith_pass_through): Likewise.
7943 (ipa_set_ancestor_jf): Likewise.
7944 (compute_complex_assign_jump_func): Set type_preserved instead of
7946 (ipa_compute_jump_functions_for_edge): Likewise.
7947 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
7948 (update_jump_functions_after_inlining): Update type_preserved.
7949 Explicitely create jump functions when combining one with pass_through.
7950 (ipa_write_jump_function): Stream the type_preserved flags.
7951 (ipa_read_jump_function): Likewise.
7953 2013-08-27 Jakub Jelinek <jakub@redhat.com>
7954 Aldy Hernandez <aldyh@redhat.com>
7956 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
7957 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
7958 * function.h (struct function): Add has_force_vect_loops and
7960 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
7961 * gimple.c (gimple_build_omp_critical): Add KIND argument and
7963 * gimple.def: Update CLAUSES comments.
7964 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
7965 (gimple_build_omp_for): Add argument to prototype.
7966 (gimple_omp_for_kind): New.
7967 (gimple_omp_for_set_kind): New.
7968 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
7969 GOVD_DATA_SHARE_CLASS.
7970 (enum omp_region_type): Add ORT_SIMD.
7971 (gimple_add_tmp_var): Handle ORT_SIMD.
7972 (gimplify_var_or_parm_decl): Same.
7973 (is_gimple_stmt): Same.
7974 (omp_firstprivatize_variable): Same.
7975 (omp_add_variable): Only use splay_tree_insert if lookup failed.
7976 (omp_notice_variable): Handle ORT_SIMD.
7977 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
7978 (omp_check_private): Handle ORT_SIMD.
7979 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7981 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
7982 Handle OMP_CLAUSE_LASTPRIVATE.
7983 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7985 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
7986 (gimplify_expr): Handle OMP_SIMD.
7987 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
7988 (expand_GOMP_SIMD_VF): New.
7989 (expand_GOMP_SIMD_LAST_LANE): New.
7990 * internal-fn.def (GOMP_SIMD_LANE): New.
7991 (GOMP_SIMD_VF): New.
7992 (GOMP_SIMD_LAST_LANE): New.
7993 * omp-low.c: Include target.h.
7994 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
7996 (check_omp_nesting_restrictions): Same.
7998 (lower_rec_simd_input_clauses): New.
7999 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
8001 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
8002 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
8003 (expand_omp_build_assign): New.
8004 (expand_omp_for_init_counts): New.
8005 (expand_omp_for_init_vars): New.
8006 (extract_omp_for_update_vars): New.
8007 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
8008 and rewrite accordingly.
8009 (expand_omp_simd): New.
8010 (expand_omp_for): Use expand_omp_simd.
8011 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
8012 (lower_omp_for): Do not lower the body.
8013 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
8015 * tree-flow.h (find_omp_clause): Remove prototype.
8016 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
8017 forcing vectorization of the loop, or if flag_tree_vectorize.
8018 (gate_tree_if_conversion): Similarly.
8019 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
8020 gimple_build_omp_for.
8021 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
8022 * tree-parloops (create_parallel_loop): Pass kind argument to
8023 gimple_build_omp_for.
8024 * tree-pretty-print.c (dump_omp_clause): Add cases for
8025 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
8026 OMP_CLAUSE__SIMDUID_.
8027 (dump_generic_node): Handle OMP_SIMD.
8028 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
8029 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
8030 unroll OMP_SIMD loops here.
8031 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
8032 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
8034 (vect_analyze_data_refs): Handle simd loops.
8035 * tree-vect-loop.c (vectorizable_live_operation): Handle
8037 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
8038 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
8039 (vectorizable_load): Same.
8040 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
8041 (struct simduid_to_vf): New.
8042 (simduid_to_vf::hash): New.
8043 (simduid_to-vf::equal): New.
8044 (struct simd_array_to_simduid): New.
8045 (simd_array_to_simduid::hash): New.
8046 (simd_array_to_simduid::equal): New.
8047 (adjust_simduid_builtins): New.
8048 (struct note_simd_array_uses_struct): New.
8049 (note_simd_array_uses_cb): New.
8050 (note_simd_array_uses): New.
8051 (vectorize_loops): Handle simd hints and adjust simd builtins
8053 * tree-vectorizer.h (struct _stmt_vec_info): Add
8054 simd_lane_access_p field.
8055 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
8056 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
8057 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
8058 (omp_clause_code_name): Same.
8059 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
8060 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
8061 * tree.def (OMP_SIMD): New entry.
8062 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
8063 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
8064 (OMP_CLAUSE_DECL): Adjust range for new clauses.
8065 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
8066 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
8067 (OMP_CLAUSE_LINEAR_STEP): New.
8068 (OMP_CLAUSE_SAFELEN_EXPR): New.
8069 (OMP_CLAUSE__SIMDUID__DECL): New.
8070 (find_omp_clause): New prototype.
8072 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
8074 * config/i386/driver-i386.c (host_detect_local_cpu): Update
8075 Haswell processor detection.
8077 2013-08-27 Christian Widmer <shadow@umbrox.de>
8080 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
8081 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
8082 AVX2 capable processors.
8084 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
8086 * config/aarch64/arm_neon.h: Replace all inline asm implementations
8087 of vget_low_* with implementations in terms of other intrinsics.
8089 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
8092 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
8093 values to -1, 0 and 1.
8095 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
8097 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
8098 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
8099 (<optab><mode>3_insn): Remove.
8100 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
8101 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
8102 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
8103 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
8104 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
8105 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
8106 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
8107 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
8108 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
8109 (ror<mode>3_insn): Likewise.
8110 * config/aarch64/predicates.md (aarch64_simd_register): New.
8112 2013-08-27 Richard Biener <rguenther@suse.de>
8114 PR tree-optimization/57521
8115 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
8116 one edge is non-critical.
8117 (find_phi_replacement_condition): Make sure to use a non-critical
8118 edge. Cleanup and remove old bug workarounds.
8119 (bb_postdominates_preds): Remove.
8120 (if_convertible_loop_p_1): Do not compute post-dominators.
8121 (combine_blocks): Do not free post-dominators.
8122 (main_tree_if_conversion): Likewise.
8123 (pass_data_if_conversion): Add TODO_verify_ssa.
8125 2013-08-27 DJ Delorie <dj@redhat.com>
8127 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
8129 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
8131 * function.c (assign_parm_find_data_types): Set passed_mode and
8132 nominal_mode to the TYPE_MODE of nominal_type for the built
8133 pointer type in case of the struct-pass-by-reference.
8135 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8137 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
8138 (UINT16_TYPE): Change default to "unsigned int".
8140 * config/avr/avr.opt (mfract-convert-truncate): New option.
8141 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
8142 is set, round negative fractional integers according to n1169
8143 when converting to integer types.
8145 2013-08-26 Jan Hubicka <jh@suse.cz>
8147 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
8148 methods can not be called indirectly when their address is not taken.
8150 2013-08-26 Jan Hubicka <jh@suse.cz>
8152 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
8155 2013-08-26 Jan Hubicka <jh@suse.cz>
8157 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
8160 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8162 * reload.c (find_valid_class): Allow classes that do not include
8163 FIRST_PSEUDO_REGISTER - 1.
8165 2013-08-26 Jan Hubicka <jh@suse.cz>
8167 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
8168 fix edge count/frequency when speculation failed; fix type check
8169 for the direct call.
8171 2013-08-26 Jan Hubicka <jh@suse.cz>
8173 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
8175 2013-08-26 Jan Hubicka <jh@suse.cz>
8177 * ipa-inline-transform.c (inline_transform): Be ready for basic block
8178 to be changed by edge redirection.
8180 2013-08-26 Jan Hubicka <jh@suse.cz>
8182 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
8183 formating; add sanity check.
8184 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
8185 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
8186 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
8187 (resolve_noninline_speculation): Update callee keys, too.
8189 2013-08-26 Jan Hubicka <jh@suse.cz>
8191 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
8192 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
8194 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8196 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
8199 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
8201 * config/i386/i386.c (ix86_debug_options): Remove prototype.
8202 (x86_64_elf_select_section): Ditto.
8203 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
8205 (ix86_pass_by_reference): Ditto.
8206 (output_set_got): Ditto.
8207 (ix86_unary_operator_ok): Ditto.
8208 (ix86_expand_builtin): Ditto.
8210 2013-08-23 Jan Hubicka <jh@suse.cz>
8212 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
8214 2013-08-23 Jan Hubicka <jh@suse.cz>
8216 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
8217 (tree_decl_with_vis): Add FINAL field.
8219 2013-08-23 Jeff Law <law@redhat.com>
8221 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
8222 the debugging dump when the expression is fully redundant.
8224 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8226 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
8227 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
8228 * pretty-print.c (pp_formatted_text_data): Likewise.
8229 (pp_write_text_to_stream): Likewise.
8230 (pp_write_text_as_dot_label_to_stream): Likewise.
8231 (pp_append_r): Likewise.
8232 (pp_format): Likewise.
8233 (pp_flush): Likewise.
8234 (pp_clear_output_area): Likewise.
8235 (pp_append_text): Likewise.
8236 (pp_formatted_text): Likewise.
8237 (pp_remaining_character_count_for_line): Likewise.
8238 (pp_newline): Likewise.
8239 (pp_character): Likewise.
8240 (output_buffer::~output_buffer): Define.
8241 (pretty_printer::~pretty_printer): Destruct output buffer.
8242 * pretty-print.h (output_buffer::~output_buffer): Declare.
8243 (pretty_printer::~pretty_printer): Declare virtual.
8245 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
8248 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
8249 HOST_WIDE_INT_M1U): New macros.
8250 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
8251 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
8252 unsigned -1 for lshift.
8253 * cse.c (cse_insn): Likewise.
8254 * double-int.c (rshift_double, lshift_double): Likewise.
8255 * builtins.c (fold_builtin_bitop): Likewise.
8256 * combine.c (force_to_mode): Likewise.
8257 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
8258 * simplify-rtx.c (simplify_const_unary_operation,
8259 simplify_const_binary_operation): Likewise.
8260 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
8261 check_va_list_escapes): Likewise.
8262 * rtlanal.c (nonzero_bits1): Likewise.
8263 * expmed.c (expand_smod_pow2): Likewise.
8264 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
8266 2013-08-23 Jan Hubicka <jh@suse.cz>
8268 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
8269 as having address taken.
8271 2013-08-23 Jan Hubicka <jh@suse.cz>
8273 * ipa-utils.h (method_class_type): Declare.
8274 * ipa-devirt.c (method_class_type): Export.
8276 * cgraphunit.c (analyze_functions): Do basic devirtualization;
8277 do not walk base classes of anonymous types.
8279 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
8281 PR rtl-optimization/58220
8283 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
8284 handle SEQUENCE insns properly.
8286 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8288 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
8290 (pp_newline_and_indent): Likewise.
8291 (pp_separate_with): Likewise.
8292 * pretty-print.c (pp_newline_and_flush): Define.
8293 (pp_newline_and_indent): Likewise.
8294 (pp_separate_with): Likewise.
8296 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8299 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
8300 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
8302 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
8304 * config/i386/predicates.md (ext_sse_reg_operand): New.
8305 * config/i386/i386.md (*movti_internal): Use
8306 predicate to determine if EVEX is needed.
8307 (*movsi_internal): Ditto.
8308 (*movdf_internal): Ditto.
8309 (*movsf_internal): Ditto.
8310 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
8312 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8314 PR tree-optimization/58209
8315 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
8316 (find_tail_calls): Give up for pointer result types if m is non-NULL.
8317 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
8318 emit POINTER_PLUS_EXPR.
8319 (create_tailcall_accumulator): For pointer result type accumulate in
8322 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
8324 * configure.ac: Add backslashes missing from the last change.
8325 * configure: Regenerate.
8327 2013-08-22 Jan Hubicka <jh@susue.cz>
8329 * ipa.c (function_and_variable_visibility): First remember function
8330 was global and then make it local.
8332 2013-08-22 Julian Brown <julian@codesourcery.com>
8334 * configure.ac: Add aarch64 to list of arches which use "nop" in
8336 * configure: Regenerate.
8338 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8340 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
8341 gnu_libc_has_function.
8342 * config/s390/tpf.h: Likewise.
8344 2013-08-22 Jan Hubicka <jh@susue.cz>
8346 * timevar.c (validate_phases): Add cast.
8348 2013-08-22 Jan Hubicka <jh@susue.cz>
8350 * timevar.c (validate_phases): Use size_t for memory.
8351 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
8353 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
8355 * pretty-print.h (output_buffer::output_buffer): Declare.
8356 (pretty_printer::pretty_printer): Likewise.
8357 (pp_construct): Remove.
8358 * pretty-print.c (output_buffer::output_buffer): Define.
8359 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
8360 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
8361 (print_gimple_expr): Likewise.
8362 (print_gimple_seq): Likewise.
8363 (gimple_dump_bb): Likewise.
8364 * sched-vis.c (dump_value_slim): Likewise.
8365 (dump_insn_slim): Likewise.
8366 (dump_rtl_slim): Likewise.
8367 (str_pattern_slim): Likewise.
8368 * tree-mudflap.c (mf_varname_tree): Likewise.
8369 * graph.c (print_graph_cfg): Likewise.
8370 (start_graph_dump): Likewise.
8371 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
8373 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
8374 pretty printer initialization.
8375 * coretypes.h (diagnostic_context): Remove superflous type alias
8377 (pretty_printer): Likewise. Declare directly as a class.
8378 (pretty_print_info): Remove declaration as class.
8379 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
8380 and pp_clear_output_area.
8381 (asan_add_global): Likewise.
8383 2013-08-22 Jan Hubicka <jh@suse.cz>
8385 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
8386 * ipa-utils.h (update_type_inheritance_graph): Declare.
8387 (possible_polymorphic_call_target_p): Declare.
8388 (possible_polymorphic_call_target_p): New.
8389 * ipa-devirt.c: Update toplevel comments.
8390 (cached_polymorphic_call_targets): Move up.
8391 (odr_type_d): Move ID down.
8392 (polymorphic_type_binfo_p): Update comment.
8393 (odr_hasher::remove): Likewise;
8394 (get_odr_type): Set anonymous_namespace.
8395 (dump_odr_type): Dump it.
8396 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
8397 (maybe_record_node): Record node in cached_polymorphic_call_targets.
8398 (record_binfo): Add comment.
8399 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
8401 (devirt_node_removal_hook): Do not iCE when cache is freed.
8402 (possible_polymorphic_call_target_p): New predicate.
8403 (update_type_inheritance_graph): New function.
8405 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
8406 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8407 Sergey Lega <sergey.s.lega@intel.com>
8408 Anna Tikhonova <anna.tikhonova@intel.com>
8409 Ilya Tocar <ilya.tocar@intel.com>
8410 Andrey Turetskiy <andrey.turetskiy@intel.com>
8411 Ilya Verbin <ilya.verbin@intel.com>
8412 Kirill Yukhin <kirill.yukhin@intel.com>
8413 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8415 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
8416 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
8417 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
8418 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
8419 (OPTION_MASK_ISA_AVX2_UNSET): Update.
8420 (OPTION_MASK_ISA_AVX512F_UNSET): New.
8421 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
8422 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
8423 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
8424 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
8425 OPT_mavx512pf, OPT_mavx512er cases.
8426 * config/i386/constraints.md (v): New constraint.
8427 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
8428 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
8429 (bit_AVX512CD): New.
8430 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8431 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
8432 * config/i386/i386-c.c (ix86_target_macros_internal):
8433 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
8435 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
8436 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
8437 * config/i386/i386.c (regclass_map, dbx_register_map)
8438 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
8439 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
8440 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
8442 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
8443 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
8444 -mavx512pf options. Fix formatting.
8445 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
8446 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
8447 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
8448 -mavx512cd, -mavx512pf options.
8449 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
8450 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
8451 (ix86_preferred_output_reload_class): Replace SSE_REGS with
8453 (ix86_hard_regno_mode_ok): Support 512-bit registers.
8454 (ix86_set_reg_reg_cost): Ditto.
8455 (x86_order_regs_for_local_alloc): Ditto.
8456 (MAX_VECT_LEN): Extend to 64-byte.
8457 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
8458 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
8459 (TARGET_AVX512ER, TARGET_AVX512CD): New.
8460 (BIGGEST_ALIGNMENT): Extend to 512-bits.
8461 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
8462 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
8463 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
8464 (SSE_REG_MODE_P): Support new modes.
8465 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
8466 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
8467 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
8468 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
8469 (REG_CLASS_CONTENTS): Add new registers.
8470 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
8471 (EXT_REX_SSE_REGNO_P): New.
8472 (HI_REGISTER_NAMES): Add new registers.
8473 * config/i386/i386.md: Define constants for new registers.
8474 (mode): Add new 512-bit modes.
8475 (prefix): Support evex prefix.
8476 (isa): Support avx512f, noavx512f, fma_avx512f.
8477 (ssemodesuffix): Add new 512-bit modes.
8479 (*movxi_internal_avx512f): Ditto.
8480 (*movdi_internal): Replace constraint "x" with the new constraint "v".
8482 (*movsi_internal): Likewise.
8483 (*movdf_internal): Likewise.
8484 (*movsf_internal): Likewise.
8485 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
8486 (<code><mode>3): Likewise.
8487 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
8489 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
8490 with the new constraint "v".
8491 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
8493 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
8494 with the new constraint "v".
8495 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
8496 (<sse2>_storedqu<avxsizesuffix>): Likewise.
8497 (*<plusminus_insn><mode>3): Likewise.
8498 (<sse>_vm<plusminus_insn><mode>3): Likewise.
8499 (*mul<mode>3): Likewise.
8500 (<sse>_vmmul<mode>3): Likewise.
8501 (<sse>_div<mode>3): Likewise.
8502 (<sse>_vmdiv<mode>3): Likewise.
8503 (<sse>_sqrt<mode>2): Likewise.
8504 (<sse>_vmsqrt<mode>2): Likewise.
8505 (*<code><mode>3_finite): Likewise.
8506 (*<code><mode>3) <smaxmin>: Likewise.
8507 (<sse>_vm<code><mode>3): Likewise.
8508 (*<code><mode>3) <any_logic>: Likewise.
8509 (*fma_fmadd_<mode>): Likewise.
8510 (*fma_fmsub_<mode>): Likewise.
8511 (*fma_fnmadd_<mode>): Likewise.
8512 (*fma_fnmsub_<mode>): Likewise.
8513 (*fma_fmaddsub_<mode>): Likewise.
8514 (*fma_fmsubadd_<mode>): Likewise.
8515 (*fmai_fmadd_<mode>): Likewise.
8516 (*fmai_fmsub_<mode>): Likewise.
8517 (*fmai_fnmadd_<mode>): Likewise.
8518 (*fmai_fnmsub_<mode>): Likewise.
8519 (sse_cvtsi2ss): Likewise.
8520 (sse_cvtsi2ssq): Likewise.
8521 (sse_cvtss2si): Likewise.
8522 (sse_cvtss2si_2): Likewise.
8523 (sse_cvtss2siq): Likewise.
8524 (sse_cvtss2siq_2): Likewise.
8525 (sse_cvttss2si): Likewise.
8526 (sse_cvtss2siq_2): Likewise.
8527 (float<sseintvecmodelower><mode>2): Likewise.
8528 (sse2_cvtsd2si_2): Likewise.
8529 (sse2_cvtsd2siq_2): Likewise.
8530 (*<plusminus_insn><mode>3): Likewise.
8531 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
8532 (*<sse4_1_avx2>_mul<mode>3): Likewise.
8533 (ashr<mode>3): Likewise.
8534 (<shift_insn><mode>3): Likewise.
8535 (avx2_<code><mode>3): Likewise.
8536 (*avx2_<code><mode>3): Likewise.
8537 (*andnot<mode>3): Likewise.
8538 (*<code><mode>3) <any_logic>: Likewise.
8539 (abs<mode>2): Likewise.
8540 (avx2_permvar<mode>): Likewise.
8541 (avx2_perm<mode>_1): Likewise.
8542 (*avx_vpermilp<mode>): Likewise.
8543 (avx_vpermilvar<mode>3): Likewise.
8544 (avx2_ashrv<mode>): Likewise.
8545 (avx2_<shift_insn>v<mode>): Likewise.
8546 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
8548 * doc/rtl.texi: Document XImode.
8550 2013-08-21 Jeff Law <law@redhat.com>
8552 * tree-flow.h (register_jump_thread): Pass vector of edges
8553 instead of each important edge.
8554 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
8555 thread path into a vector and pass that to register_jump_thread.
8556 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
8557 passed in edge vector to the current 3-edge form.
8560 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8562 * dce.c (fini_dce): Call df_analyze again just in case
8563 delete_unmarked_insns removed anything.
8565 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
8567 * reload.h (struct reg_equivs): Rename to ..
8568 (struct reg_equivs_s): .. this.
8570 2013-08-20 Martin Liska <marxin.liska@gmail.com>
8572 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
8574 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8576 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
8578 2013-08-21 Jeff Law <law@redhat.com>
8580 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
8581 simplify assignments too. If the RHS collapses to a singleton
8582 range, then return the value for the range.
8584 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
8586 * config/i386/sse.md (V16): Rename to...
8588 (mov<mode>): Update iterator name.
8589 (*mov<mode>_internal): Ditto.
8590 (push<mode>1): Ditto.
8591 (movmisalign<mode>): Ditto.
8593 2013-08-20 Jan Hubicka <jh@suse.cz>
8596 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
8597 entry for direct edges.
8598 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
8600 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8602 Revert my last two changes, r201865 and r201864:
8605 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8607 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8608 instances can own GC refs.
8610 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8611 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8612 (gcc::context::gt_pch_nx): Likewise.
8613 (gcc::context::gt_pch_nx): Likewise.
8614 * ggc.h (gt_ggc_mx <T>): New.
8615 (gt_pch_nx_with_op <T>): New.
8616 (gt_pch_nx <T>): New.
8617 * passes.c (opt_pass::gt_ggc_mx): New.
8618 (opt_pass::gt_pch_nx): New.
8619 (opt_pass::gt_pch_nx_with_op): New.
8620 (pass_manager::gt_ggc_mx): New.
8621 (pass_manager::gt_pch_nx): New.
8622 (pass_manager::gt_pch_nx_with_op): New.
8623 (pass_manager::operator new): Use
8624 ggc_internal_cleared_alloc_stat rather than xcalloc.
8625 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8626 (pass_manager::gt_ggc_mx): New.
8627 (pass_manager::gt_pch_nx): New.
8628 (pass_manager::gt_pch_nx_with_op): New.
8629 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8630 (opt_pass::operator new): New.
8631 (opt_pass::gt_ggc_mx): New.
8632 (opt_pass::gt_pch_nx): New.
8633 (opt_pass::gt_pch_nx_with_op): New.
8636 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8638 * Makefile.in (GTFILES): Add context.h.
8639 * context.c (gcc::context::operator new): New.
8640 (gcc::context::gt_ggc_mx): New.
8641 (gcc::context::gt_pch_nx): New.
8642 (gcc::context::gt_pch_nx): New.
8643 * context.h (gcc::context): Add GTY((user)) marking.
8644 (gcc::context::operator new): New.
8645 (gcc::context::gt_ggc_mx): New.
8646 (gcc::context::gt_pch_nx): New.
8647 (gcc::context::gt_pch_nx): New.
8648 (g): Add GTY marking.
8649 (gt_ggc_mx (gcc::context *)): New.
8650 (gt_pch_nx (gcc::context *)): New.
8651 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8652 void *cookie)): New.
8653 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8655 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8657 * dce.c (fini_dce): Call df_analyze again just in case
8658 delete_unmarked_insns removed anything.
8660 2013-08-20 Teresa Johnson <tejohnson@google.com>
8662 PR rtl-optimizations/57451
8663 * final.c (reemit_insn_block_notes): Prevent lexical blocks
8664 from crossing split section boundaries.
8666 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8668 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
8669 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
8670 with MULTLIB_DEFAULTS.
8672 2013-08-20 Nick Clifton <nickc@redhat.com>
8674 * target.def (narrow_volatile_bitfield): Note that the default
8675 value is false, not !TARGET_STRICT_ALIGN.
8676 * doc/tm.texi: Regenerate.
8678 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
8680 Fix LIB_SPEC for systems without libpthread.
8682 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
8683 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
8685 * config/i386/linux-common.h: Likewise.
8686 * config/mips/linux-common.h: Likewise.
8688 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
8690 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
8693 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8695 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8696 instances can own GC refs.
8698 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8699 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8700 (gcc::context::gt_pch_nx): Likewise.
8701 (gcc::context::gt_pch_nx): Likewise.
8702 * ggc.h (gt_ggc_mx <T>): New.
8703 (gt_pch_nx_with_op <T>): New.
8704 (gt_pch_nx <T>): New.
8705 * passes.c (opt_pass::gt_ggc_mx): New.
8706 (opt_pass::gt_pch_nx): New.
8707 (opt_pass::gt_pch_nx_with_op): New.
8708 (pass_manager::gt_ggc_mx): New.
8709 (pass_manager::gt_pch_nx): New.
8710 (pass_manager::gt_pch_nx_with_op): New.
8711 (pass_manager::operator new): Use
8712 ggc_internal_cleared_alloc_stat rather than xcalloc.
8713 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8714 (pass_manager::gt_ggc_mx): New.
8715 (pass_manager::gt_pch_nx): New.
8716 (pass_manager::gt_pch_nx_with_op): New.
8717 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8718 (opt_pass::operator new): New.
8719 (opt_pass::gt_ggc_mx): New.
8720 (opt_pass::gt_pch_nx): New.
8721 (opt_pass::gt_pch_nx_with_op): New.
8723 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8725 * Makefile.in (GTFILES): Add context.h.
8726 * context.c (gcc::context::operator new): New.
8727 (gcc::context::gt_ggc_mx): New.
8728 (gcc::context::gt_pch_nx): New.
8729 (gcc::context::gt_pch_nx): New.
8730 * context.h (gcc::context): Add GTY((user)) marking.
8731 (gcc::context::operator new): New.
8732 (gcc::context::gt_ggc_mx): New.
8733 (gcc::context::gt_pch_nx): New.
8734 (gcc::context::gt_pch_nx): New.
8735 (g): Add GTY marking.
8736 (gt_ggc_mx (gcc::context *)): New.
8737 (gt_pch_nx (gcc::context *)): New.
8738 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8739 void *cookie)): New.
8740 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8742 2013-08-20 Alan Modra <amodra@gmail.com>
8745 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
8746 (rs6000_emit_epilogue): Likewise.
8748 2013-08-19 Dehao Chen <dehao@google.com>
8750 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
8752 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
8753 Jakub Jelinek <jakub@redhat.com>
8755 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
8756 (BUILT_IN_FABSD64): Likewise.
8757 (BUILT_IN_FABSD128): Likewise.
8758 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
8759 (fold_builtin_1): Likewise.
8760 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
8761 destination and source operands.
8762 (*abstd2_fpr): Likewise.
8763 (*nabstd2_fpr): Likewise.
8765 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
8767 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
8770 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
8772 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
8775 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
8777 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
8778 * builtins.c (default_libc_has_function): New.
8779 (gnu_libc_has_function): Ditto.
8780 (no_c99_libc_has_function): Ditto.
8781 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
8782 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
8783 (fold_builtin_sincos): Likewise.
8784 (fold_builtin_cexp): Likewise.
8785 * builtins.def (DEF_C94_BUILTIN): Likewise.
8786 (DEF_C99_BUILTIN): Likewise.
8787 (DEF_C99_C90RES_BUILTIN): Likewise.
8788 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
8789 definitions to using this define.
8790 * config/darwin-protos.h (darwin_libc_has_function): New.
8791 * config/darwin.c (darwin_libc_has_function): Ditto.
8792 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
8793 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
8794 * config/darwin.h: Ditto.
8795 * config/elfos.h: Ditto.
8796 * config/freebsd.h: Ditto.
8797 * config/i386/cygming.h: Ditto.
8798 * config/i386/djgpp.h: Ditto.
8799 * config/i386/i386-interix.h: Ditto.
8800 * config/microblaze/microblaze.h: Ditto.
8801 * config/mmix/mmix.h: Ditto.
8802 * config/gnu-user.h: Ditto.
8803 * config/ia64/hpux.h: Ditto.
8804 * config/pa/pa-hpux.h: Ditto.
8805 * config/pdp11/pdp11.h: Ditto.
8806 * config/picochip/picochip.h: Ditto.
8807 * config/linux.h: Ditto.
8808 * config/netbsd.h: Ditto.
8809 * config/openbsd.h: Ditto.
8810 * config/rs6000/aix43.h: Ditto.
8811 * config/rs6000/aix51.h: Ditto.
8812 * config/rs6000/aix52.h: Ditto.
8813 * config/rs6000/aix53.h: Ditto.
8814 * config/rs6000/aix61.h: Ditto.
8815 * config/rs6000/darwin.h: Ditto.
8816 * config/rs6000/linux.h: Ditto.
8817 * config/rs6000/linux64.h: Ditto.
8818 * config/s390/tpf.h: Ditto.
8819 * config/sol2-10.h: Ditto.
8820 * config/sol2.h: Ditto.
8821 * config/vms/vms.h: Ditto.
8822 * config/vxworks.h: Ditto.
8823 * config/linux-android.c (linux_android_libc_has_function):
8824 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
8825 * config/linux-protos.h (linux_android_libc_has_function):
8827 * config/i386/i386.c (ix86_libc_has_function): New.
8828 * config/i386/i386-protos.h
8829 (ix86_libc_has_function): New declaration.
8830 * config/i386/i386.md
8831 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
8832 ("isinf<mode>2): Likewise.
8833 * convert.c (convert_to_integer): Using new target hook
8834 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
8835 TARGET_C99_FUNCTIONS.
8836 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
8837 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
8838 * coretypes.h (function_class): New enum for different
8839 classes of functions.
8840 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
8841 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
8842 (TARGET_HAS_SINCOS): Likewise.
8843 (TARGET_LIBC_HAS_FUNCTION): New.
8844 * doc/tm.texi: Regenerated.
8845 * targhooks.h (default_libc_has_function): New declaration.
8846 (no_c99_libc_has_function): Ditto.
8847 (gnu_libc_has_function): Ditto.
8848 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
8849 and TARGET_HAS_SINCOS.
8851 2013-08-18 Jan Hubicka <jh@suse.cz>
8853 * Makeifle-in (ipa-devirt.o): New.
8854 (GTFILES): Add ipa-utils.h and ipa-devirt.c
8855 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
8856 (analyze_functions): Look into possible targets of polymorphic call.
8857 * dumpfile.c (dump_files): Add type-inheritance dump.
8858 * dumpfile.h (TDI_inheritance): New.
8859 * ipa-devirt.c: New file.
8860 * ipa-utils.h (odr_type_d): Forward declare.
8861 (odr_type): New type.
8862 (build_type_inheritance_graph): Declare.
8863 (possible_polymorphic_call_targets): Declare and introduce inline
8864 variant when only edge is pased.
8865 (dump_possible_polymorphic_call_targets): Likewise.
8866 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
8867 * tree.c (type_in_anonymous_namespace_p): Break out from ...
8868 (types_same_for_odr): ... here.
8869 * tree.h (type_in_anonymous_namespace_p): Declare.
8871 2013-08-18 Jakub Jelinek <jakub@redhat.com>
8873 PR tree-optimization/58006
8874 * tree-parloops.c (take_address_of): Don't ICE if get_name
8876 (eliminate_local_variables_stmt): Remove clobber stmts.
8878 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
8880 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
8881 error message is issued for an alias to undefined symbol.
8883 2013-08-18 Jan Hubicka <jh@suse.cz>
8885 * cgraph.c (cgraph_create_indirect_edge): Discover
8886 polymorphic calls and record basic info into indirect_info.
8887 * gimple-fold.c (gimple_fold_call): When doing BINFO based
8888 devirtualization, ignore objc function calls.
8889 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
8890 call with no parm index info.
8891 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
8892 * tree.c (virtual_method_call_p): New function.
8893 * tree.h (virtual_method_call_p): Declare.
8895 2013-08-16 Jan Hubicka <jh@suse.cz>
8898 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
8900 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
8902 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
8905 2013-08-16 David Malcolm <dmalcolm@redhat.com>
8907 * gengtype.c (type_for_name): Add special-case support for
8908 locating types within the "gcc::" namespace.
8909 (open_base_files): Emit a "using namespace gcc" directive.
8911 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
8914 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
8915 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
8917 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
8918 array instead of each individual operand as a separate argument.
8919 (emit_fusion_gpr_load): Likewise.
8920 (expand_fusion_gpr_load): Add new function declaration.
8922 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
8923 signature to have the operands passed as an array, instead of as
8924 separate arguments. Allow ZERO_EXTEND to be in the memory
8925 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
8926 depend on the register live/dead flags when peepholes are run.
8927 (expand_fusion_gpr_load): New function to be called from the
8928 peephole2 pass, to change the register that addis sets to be the
8930 (emit_fusion_gpr_load): Change the calling signature to have the
8931 operands passed as an array, instead of as separate arguments.
8932 Allow ZERO_EXTEND to be in the memory address, and also
8933 SIGN_EXTEND if -mpower8-fusion-sign.
8935 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
8937 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
8938 adjust the register addis loads up in the peephole2 pass. Do not
8939 depend on the register live/dead state when the peephole pass is done.
8941 2013-08-16 David Malcolm <dmalcolm@redhat.com>
8943 * gengtype.c (create_user_defined_type): Ensure that the kind
8944 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
8945 declaration is seen before the GTY((user)) marking.
8947 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
8950 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
8952 2013-08-16 Jan Hubicka <jh@suse.cz>
8954 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
8955 arugment expected_type.
8956 (gimple_fold_call): Use it.
8957 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
8958 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
8959 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
8960 (try_make_edge_direct_virtual_call): Likewise.
8961 * tree.c (obj_type_ref_class): New.
8962 * tree.h (obj_type_ref_class): Use it.
8964 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
8966 * sched-vis.c (rtl_slim_pp_initialized): Remove.
8967 (rtl_slim_pp): Likewise.
8968 (init_rtl_slim_pretty_print): Likewise.
8969 (dump_value_slim): Don't call it. Use local pretty printer.
8970 (dump_insn_slim): Likewise.
8971 (dump_rtl_slim): Likewise.
8972 (str_pattern_slim): Likewise.
8973 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
8976 2013-08-16 Jakub Jelinek <jakub@redhat.com>
8978 PR tree-optimization/58164
8979 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
8980 walk gimple_goto_dest of GIMPLE_GOTO.
8982 PR tree-optimization/58165
8983 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
8984 bi_call must be the last stmt in a bb, don't split_block, instead
8985 use fallthru edge from it and give up if there is none.
8986 Release conds vector when returning early.
8988 2013-08-14 Xinliang David Li <davidxl@google.com>
8990 * config/i386/i386.c (ix86_option_override_internal):
8991 Remove unused variable and field.
8993 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8996 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
8997 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
8998 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
8999 and Linux, correct BLKmode alignment when 128-bit alignment is
9000 required and compatibility flag is not set.
9001 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
9002 for zero-size arguments when compatibility flag is not set.
9004 2013-08-14 Jakub Jelinek <jakub@redhat.com>
9006 PR tree-optimization/58145
9007 * tree-sra.c (build_ref_for_offset): If prev_base has
9008 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
9010 2013-08-14 Xinliang David Li <davidxl@google.com>
9012 * config/i386/i386.c (ix86_option_override_internal):
9013 Fix uninitialized variable error.
9015 2013-08-14 Xinliang David Li <davidxl@google.com>
9017 * config/i386/i386.opt: Define two new options.
9018 * config/i386/x86-tune.def: Add arch selector field in macros.
9019 * config/i386/i386.h: Adjust macro definition.
9020 * config/i386/i386.c (ix86_option_override_internal):
9022 (parse_mtune_ctrl_str): New function.
9023 (set_ix86_tune_features): New function.
9024 (ix86_function_specific_restore): Call the new helper function.
9026 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
9028 PR rtl-optimization/57662
9029 * sel-sched.c (code_motion_process_successors): When the current insn
9030 is removed after the recursive traversal, break from the loop.
9031 Add comments and debug printouts.
9033 2013-08-14 Jakub Jelinek <jakub@redhat.com>
9034 Alexandre Oliva <aoliva@redhat.com>
9037 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
9038 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
9039 there also UNSPEC_PLTOFF.
9041 2013-08-14 Marek Polacek <polacek@redhat.com>
9043 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
9044 NUM_CONDITIONS bit positions.
9046 2013-08-13 Cary Coutant <ccoutant@google.com>
9048 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
9049 (attr_checksum): Hash vector contents instead of pointer.
9050 (attr_checksum_ordered): Likewise.
9052 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
9054 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
9055 when Pmode != word_mode. Add length_address attribute.
9056 (sse3_monitor_<mode>): Merge from sse3_monitor and
9057 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
9058 Pmode != word_mode. Update insn length attribute.
9059 * config/i386/i386.c (ix86_option_override_internal): Update
9060 ix86_gen_monitor selection for merged sse3_monitor insn.
9062 2013-08-13 Julian Brown <julian@codesourcery.com>
9064 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
9065 perform invalid legitimization on greater-than-word-size modes for
9068 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
9070 * ira.c (setup_class_translate_array): Use aclass instead of cl
9071 for classes not fully covered by allocno classes.
9073 2013-08-13 Jakub Jelinek <jakub@redhat.com>
9075 PR tree-optimization/57661
9076 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
9077 * tree-inline.c (tree_function_versioning): Initialize it.
9078 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
9079 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
9080 that is not being copied.
9083 * asan.c (instrument_strlen_call): Fix typo in comment.
9084 Use char * type even for the lhs of POINTER_PLUS_EXPR.
9086 2013-08-13 Steve Ellcey <sellcey@mips.com>
9088 * config/mips/mips.md (prefetch): Use lw instead of ld on
9089 loongson in 32bit mode.
9091 2013-08-13 Nick Clifton <nickc@redhat.com>
9093 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
9095 2013-08-13 Jan Hubicka <jh@suse.cz>
9097 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
9098 introduced edge; fix typo in sanity check.
9099 (cgraph_resolve_speculation): Export; improve diagnostic.
9100 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
9101 speculation at type mismatch.
9102 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
9103 (cgraph_resolve_speculation): Declare.
9104 (symtab_can_be_discarded): New function.
9105 * value-prof.c (gimple_ic_transform): Remove actual transform code.
9106 * ipa-inline-transform.c (speculation_removed): New global var.
9107 (clone_inlined_nodes): See if speculation can be removed.
9108 (inline_call): If speculations was removed, we growths may not match.
9109 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
9110 (speculation_useful_p): New function.
9111 (resolve_noninline_speculation): New function.
9112 (inline_small_functions): Resolve useless speculations.
9113 * ipa-inline.h (speculation_useful_p): Declare
9114 * ipa.c (can_replace_by_local_alias): Simplify.
9115 (ipa_profile): Produce speculative calls in non-lto, too;
9116 add simple cost model; produce local aliases.
9118 2013-08-13 David Malcolm <dmalcolm@redhat.com>
9120 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
9123 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
9125 * config/i386/i386.c (ix86_function_versions): Use error + inform.
9127 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
9129 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
9130 iterator instead of X87MODEF.
9132 2013-08-12 Perez Read <netfirewall@gmail.com>
9135 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
9136 operand 0 for intel asm alternative.
9137 (*movabs<mode>_2): Ditto for operand 1.
9139 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
9141 * config/aarch64/arm_none.h
9142 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
9144 2013-08-12 Nick Clifton <nickc@redhat.com>
9146 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
9148 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
9150 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
9151 expand for QI/HImode operand to produce more effictive code for
9152 unsigned char(short) --> float(double) conversion.
9154 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
9156 * doc/invoke.texi: Mention that -ftls-model does not force the final
9159 2013-08-12 Marek Polacek <polacek@redhat.com>
9160 Marc Glisse <marc.glisse@inria.fr>
9162 PR tree-optimization/57980
9163 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
9164 when creating -1 constant.
9166 2013-08-10 Jan Hubicka <jh@suse.cz>
9168 Workaround binutils PR14342.
9169 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
9170 (gimple_init_edge_profiler): Likewise.
9171 (gimple_gen_ic_func_profiler): Likewise.
9173 2013-08-09 Jan Hubicka <jh@suse.cz>
9175 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
9177 2013-08-09 Xinliang David Li <davidxl@google.com>
9179 * config/i386/stringop.def: New file.
9180 * config/i386/stringop.opt: New file.
9181 * config/i386/i386-opts.h: Include stringopt.def.
9182 * config/i386/i386.opt: Include stringopt.opt.
9183 * config/i386/i386.c (ix86_option_override_internal):
9184 Override default size based stringop inline strategies with options.
9185 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
9188 2013-08-09 Jan Hubicka <jh@suse.cz>
9190 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
9192 2013-08-09 Jan Hubicka <jh@suse.cz>
9194 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
9196 (dump_cgraph_node): Dump profile-id.
9197 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
9198 and common_target_probability.
9199 * lto-cgraph.c (lto_output_edge): Stream common targets.
9200 (lto_output_node): Stream profile ids.
9201 (input_node): Stream profile ids.
9202 (input_edge): Stream common targets.
9203 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
9204 * ipa.c: Include value-prof.h
9205 (ipa_profile_generate_summary): Turn indirect call statement histograms
9206 into common targets.
9207 (ipa_profile): Turn common targets into speculative edges.
9209 2013-08-09 Jan Hubicka <jh@suse.cz>
9211 * cgraph.h (cgraph_node): Add profile_id.
9212 * value-prof.c (cgraph_node_map): Turn into pointer_map.
9213 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
9214 (del_node_map): Update.
9215 (find_func_by_funcdef_no): Replace by ...
9216 (find_func_by_profile_id): ... this one.
9217 (gimple_ic_transform): Do not remove useful histograms when
9218 speculation is not done; dump info when indirect call removal
9220 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
9221 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
9222 (__gcov_indirect_call_profiler_v2): .. this one.
9223 * profile.h (init_node_map): Update.
9224 * coverage.c (coverage_compute_profile_id): New function.
9225 * coverage.h (coverage_compute_profile_id): Declare.
9226 * tree-profile.c (init_ic_make_global_vars): Make
9227 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
9228 (gimple_init_edge_profiler): Update prototype of
9229 __gcov_indirect_call_profiler.
9230 (gimple_gen_ic_func_profiler): Simplify.
9231 (tree_profiling): Use init_node_map
9233 2013-08-09 Jan Hubicka <jh@suse.cz>
9235 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
9236 non-speculative refs.
9237 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
9238 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
9239 (cgraph_set_call_stmt): Likewise.
9240 (cgraph_create_edge_1): Fix release checking compilatoin;
9242 (cgraph_free_edge): Free indirect info.
9243 (cgraph_turn_edge_to_speculative): New function.
9244 (cgraph_speculative_call_info): New function.
9245 (cgraph_make_edge_direct): Return direct edge; handle speculation.
9246 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
9247 (dump_cgraph_node): Dump speculation.
9248 (verify_edge_count_and_frequency): Accept speculative edges.
9249 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
9250 (verify_cgraph_node): Handle speculation.
9251 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
9252 (cgraph_set_call_stmt): Update prototype.
9253 (cgraph_make_edge_direct): Update prototype.
9254 (cgraph_speculative_call_info): Declare.
9255 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
9256 to change; update call of ipa_find_references.
9257 * ipa-ref.c (ipa_record_reference): Fix return value; clear
9258 lto_stmt_uid and speculative flags.
9259 (ipa_dump_references): Dump speculation.
9260 (ipa_clone_references): Clone speculative flag.
9261 (ipa_clone_referring): Likewise.
9262 (ipa_clone_ref): New function.
9263 (ipa_find_reference): Look into lto_stmt_uids
9264 (ipa_clear_stmts_in_references): Do not clear speculative calls.
9265 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
9266 (ipa_find_reference): Update declaration.
9267 (ipa_clone_ref): Declare.
9268 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
9269 stream speculative flag.
9270 (lto_output_ref): Stream statements uids and speculation.
9271 (input_ref): Likewise.
9272 (input_edge): Stream speuclation.
9273 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
9274 (cgraph_set_call_stmt_including_clones): Handle speculation.
9275 * ipa-inline.c (heap_edge_removal_hook): New function.
9276 (inline_small_functions): Register it.
9277 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
9278 also initialize refs.
9279 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
9281 (try_make_edge_direct_simple_call): Likewise.
9282 (try_make_edge_direct_simple_call): Likewise.
9283 (update_indirect_edges_after_inlining): Likewise.
9284 (remove_described_reference): Look proper lto_stmt_uid.
9285 (propagate_controlled_uses): Likewise.
9286 (propagate_controlled_uses): Liekwise.
9287 * tree-inline.c (copy_bb): Copy speculative edges.
9288 (redirect_all_calls): New function.
9289 (copy_cfg_body): Do redirection after loop info is updated.
9290 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
9292 2013-08-09 Jan Hubicka <jh@suse.cz>
9294 * lto-streamer-out.c (output_function): Renumber PHIs.
9295 * lto-streamer-in.c (input_function): Likewise.
9297 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
9299 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
9300 (get_lane_unsigned): Likewise.
9301 (dup_lane_scalar): Likewise.
9302 (get_lane): enable for VALL.
9303 * config/aarch64/aarch64-simd.md
9304 (aarch64_dup_lane_scalar<mode>): Remove.
9305 (aarch64_get_lane_signed<mode>): Likewise.
9306 (aarch64_get_lane_unsigned<mode>): Likewise.
9307 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
9308 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
9309 (aarch64_get_lane<mode>): Enable for all vector modes.
9310 (aarch64_get_lanedi): Remove misleading constraints.
9311 * config/aarch64/arm_neon.h
9312 (__aarch64_vget_lane_any): Define.
9313 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
9314 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
9315 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
9316 * config/aarch64/iterators.md (VDQQH): New.
9320 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
9322 * configure.ac: Add GAS check for LEON instructions on SPARC.
9323 * configure: Regenerate.
9324 * config.in: Likewise.
9325 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
9327 * config/sparc/sparc.opt (LEON, LEON3): New masks.
9328 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
9330 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
9331 (AS_LEON_FLAG): New macro.
9332 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
9333 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
9334 Deal with LEON and LEON3 for the memory model.
9335 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
9336 (atomic_compare_and_swap<mode>_1): Likewise.
9337 (*atomic_compare_and_swap<mode>_1): Likewise.
9339 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9341 * config/arm/neon.md (vcond): Fix floating-point vector
9342 comparisons against 0.
9344 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
9346 * lra-constraints.c (emit_spill_move): Remove assert.
9347 (process_alt_operands): Add more debugging
9348 output. Increase reject for spilling into memory. Decrease
9349 reject for reloading scratch.
9350 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
9352 2013-08-08 Steve Ellcey <sellcey@mips.com>
9354 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
9355 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
9356 micromips incompatible. Add nan2008.
9357 (MULTILIB_DIRNAMES): Add nan2008.
9358 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9359 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
9360 and micromips incompatible. Add nan2008.
9361 (MULTILIB_DIRNAMES): Add nan2008.
9362 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9364 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
9366 PR rtl-optimization/58079
9367 * combine.c (combine_simplify_rtx): Avoid using SUBST if
9368 simplify_comparison has widened a comparison with an integer.
9370 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9372 * config/arm/neon.md (movmisalign<mode>): Disable when we
9373 don't allow unaligned accesses.
9374 (*movmisalign<mode>_neon_store): Likewise.
9375 (*movmisalign<mode>_neon_load): Likewise.
9376 (*movmisalign<mode>_neon_store): Likewise.
9377 (*movmisalign<mode>_neon_load): Likewise.
9379 2013-08-08 Jan Hubicka <jh@suse.cz>
9381 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
9382 (make_pass_rebuild_cgraph_edges): Also clear references.
9383 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
9384 * ipa-inline-transform.c (inline_transform): Remove all references
9386 * cgraphunit.c (expand_function): Remove all references after
9388 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
9389 (ipa_find_reference): Rewrite to iterator.
9390 (remove_stmt_references): Likewise.
9391 (ipa_clear_stmts_in_references): New function.
9392 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
9393 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
9395 * ipa-split.c (split_function): Remove references in split function.
9397 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9400 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
9401 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
9403 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9406 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
9407 suggested mode for the assignment isn't compatible with the
9410 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9413 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
9415 2013-08-07 Xinliang David Li <davidxl@google.com>
9417 * config/i386/i386.opt: New option -mtune-ctrl=.
9418 * config/i386/x86-tune.def: New file.
9419 * config/i386/i386.h: include x86-tune.def.
9420 * config/i386/i386.c (ix86_option_override_internal):
9421 Parsing -mtune-ctrl= option and set tune features.
9423 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
9426 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
9427 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
9430 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9432 * rtl.h (update_alignments): Declare.
9433 * final.c (grow_label_align): New function extracted from...
9434 (shorten_branches): ...here. Call it.
9435 (update_alignments): New function.
9436 * reorg.c (sibling_labels): New variable.
9437 (get_label_before): Add SIBLING parameter. If it is non-zero, push
9438 the new label along with it onto the sibling_labels vector.
9439 (fill_simple_delay_slots): Adjust call to get_label_before.
9440 (fill_slots_from_thread): Likewise.
9441 (relax_delay_slots): Likewise.
9442 (make_return_insns): Likewise.
9443 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
9445 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9447 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
9448 document its semantics.
9449 (diagnostic_report_diagnostic): Adjust accordingly.
9451 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9453 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
9454 (sparc_option_override): ...and port to new C++ pass API.
9455 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
9457 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
9459 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
9461 2013-08-06 Caroline Tice <cmtice@google.com>
9463 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
9464 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
9465 * tree-pass.h: Add pass_vtable_verify.
9466 * varasm.c (assemble_variable): Add code to properly set the comdat
9467 section and name for the .vtable_map_vars section.
9468 (assemble_vtyv_preinit_initializer): New function.
9469 (default_sectin_type_flags): Make sure .vtable_map_vars section has
9471 * output.h: Add function decl for assemble_vtv_preinit_initializer.
9472 * vtable-verify.c: New file.
9473 * vtable-verify.h: New file.
9474 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
9475 initialiation levels.
9476 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
9477 * passes.def: Insert pass_vtable_verify.
9478 * aclocal.m4: Reorder includes.
9479 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
9480 -fvtv-counts options.
9481 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
9482 as appropriate, if -fvtable-verify=... is used.
9483 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
9484 -fvtable-verify=... is used.
9485 * Makefile.in (OBJS): Add vtable-verify.o to list.
9486 (vtable-verify.o): Add new build rule.
9487 (GTFILES): Add vtable-verify.c to list.
9488 * common.opt (fvtable-verify=): New flag.
9489 (vtv_priority): Values for fvtable-verify= flag.
9490 (fvtv-counts): New flag.
9491 (fvtv-debug): New flag.
9492 * tree.h (save_vtable_map_decl): New extern function decl.
9494 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9496 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
9497 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
9498 (pass_data_rl78_devirt): ...new pass_data instance and...
9499 (make_pass_rl78_devirt): ...new function.
9500 (rl78_asm_file_start): Port pass registration to new C++ API.
9502 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9504 * coretypes.h (rtl_opt_pass): Add.
9505 (gcc::context): Add.
9506 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
9507 (epiphany_init): Port to new C++ pass API.
9508 (epiphany_optimize_mode_switching): Likewise.
9509 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
9510 (pass_manager::get_pass_mode_switching): New.
9511 (pass_manager::get_pass_peephole2): New.
9512 * mode-switching.c (pass_mode_switching): Add clone method.
9513 * recog.c (pass_peephole2): Add clone method.
9514 (pass_split_all_insns): Add clone method.
9516 2013-08-06 David Malcolm <dmalcolm@redhat.com>
9518 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
9519 (mips_option_override): ...here, porting to new C++ API for passes.
9521 2013-08-06 Jan Hubicka <jh@suse.cz>
9523 * cgraph.c (cgraph_get_body): New function based on lto.c
9525 * cgraph.h (cgraph_get_body): Declare.
9526 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
9528 * cgraphunit.c (expand_function): Get body prior expanding.
9529 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
9530 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
9532 * passes.c (do_per_function_toporder): Get body.
9533 * tree-inline.c (expand_call_inline): Get body prior inlining it.
9534 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
9536 2013-08-06 Martin Jambor <mjambor@suse.cz>
9539 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
9540 is not re-finalized. Rename second parameter to no_collect.
9542 2013-08-06 Martin Jambor <mjambor@suse.cz>
9545 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
9546 MEM_REF has proper alignment information.
9548 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
9551 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
9553 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
9554 rtx (*) (rtx, ...) with insn_gen_fn.
9555 * genoutput.c (output_insn_data): Cast gen_? function pointers to
9556 insn_gen_fn::stored_funcptr. Add initializer braces.
9558 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9560 Rewrite how instances of passes are cloned to remove assumptions
9561 about their sizes (thus allowing pass subclasses to have
9562 additional data fields, albeit non-GC-managed ones at this point).
9564 * passes.c (make_pass_instance): Now that passes have clone
9565 methods, rewrite this function to eliminate XNEW and memcpy
9566 calls that used hardcoded sizes. Since this function no longer
9567 creates pass instances, rename it to...
9568 (add_pass_instance): ...this. Document the old way that passes were
9569 numbered and flagged, and rework this function to continue using it.
9570 (next_pass_1): Add an initial_pass argument for use by
9572 (position_pass): When adding multiple instances of a pass, use
9573 the pass's clone method, rather than relying on the XNEW/memcpy
9574 within the former make_pass_instance (now add_pass_instance).
9575 (pass_manager::pass_manager): When invoking next_pass_1, also supply
9576 the initial instance of the current pass within the pass manager.
9578 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9580 This is the automated part of the conversion of passes from C
9581 structs to C++ classes.
9583 Patch autogenerated by refactor_passes.py from
9584 https://github.com/davidmalcolm/gcc-refactoring-scripts
9585 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
9587 * asan.c (pass_asan): Convert from a global struct to a subclass of
9588 gimple_opt_pass along with...
9589 (pass_data_asan): ...new pass_data instance and...
9590 (make_pass_asan): ...new function.
9591 (pass_asan_O0): Convert from a global struct to a subclass of
9592 gimple_opt_pass along with...
9593 (pass_data_asan_O0): ...new pass_data instance and...
9594 (make_pass_asan_O0): ...new function.
9595 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
9596 subclass of rtl_opt_pass along with...
9597 (pass_data_inc_dec): ...new pass_data instance and...
9598 (make_pass_inc_dec): ...new function.
9599 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
9600 a subclass of rtl_opt_pass along with...
9601 (pass_data_reorder_blocks): ...new pass_data instance and...
9602 (make_pass_reorder_blocks): ...new function.
9603 (pass_duplicate_computed_gotos): Convert from a global struct to a
9604 subclass of rtl_opt_pass along with...
9605 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
9606 (make_pass_duplicate_computed_gotos): ...new function.
9607 (pass_partition_blocks): Convert from a global struct to a subclass of
9608 rtl_opt_pass along with...
9609 (pass_data_partition_blocks): ...new pass_data instance and...
9610 (make_pass_partition_blocks): ...new function.
9611 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
9612 struct to a subclass of rtl_opt_pass along with...
9613 (pass_data_branch_target_load_optimize1): ...new pass_data instance
9615 (make_pass_branch_target_load_optimize1): ...new function.
9616 (pass_branch_target_load_optimize2): Convert from a global struct to a
9617 subclass of rtl_opt_pass along with...
9618 (pass_data_branch_target_load_optimize2): ...new pass_data instance
9620 (make_pass_branch_target_load_optimize2): ...new function.
9621 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
9622 of rtl_opt_pass along with...
9623 (pass_data_jump): ...new pass_data instance and...
9624 (make_pass_jump): ...new function.
9625 (pass_jump2): Convert from a global struct to a subclass of
9626 rtl_opt_pass along with...
9627 (pass_data_jump2): ...new pass_data instance and...
9628 (make_pass_jump2): ...new function.
9629 * cfgexpand.c (pass_expand): Convert from a global struct to a
9630 subclass of rtl_opt_pass along with...
9631 (pass_data_expand): ...new pass_data instance and...
9632 (make_pass_expand): ...new function.
9633 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
9634 of rtl_opt_pass along with...
9635 (pass_data_free_cfg): ...new pass_data instance and...
9636 (make_pass_free_cfg): ...new function.
9637 (pass_into_cfg_layout_mode): Convert from a global struct to a
9638 subclass of rtl_opt_pass along with...
9639 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
9640 (make_pass_into_cfg_layout_mode): ...new function.
9641 (pass_outof_cfg_layout_mode): Convert from a global struct to a
9642 subclass of rtl_opt_pass along with...
9643 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
9644 (make_pass_outof_cfg_layout_mode): ...new function.
9645 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
9646 struct to a subclass of gimple_opt_pass along with...
9647 (pass_data_build_cgraph_edges): ...new pass_data instance and...
9648 (make_pass_build_cgraph_edges): ...new function.
9649 (pass_rebuild_cgraph_edges): Convert from a global struct to a
9650 subclass of gimple_opt_pass along with...
9651 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
9652 (make_pass_rebuild_cgraph_edges): ...new function.
9653 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
9654 subclass of gimple_opt_pass along with...
9655 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
9657 (make_pass_remove_cgraph_callee_edges): ...new function.
9658 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
9659 struct to a subclass of rtl_opt_pass along with...
9660 (pass_data_stack_adjustments): ...new pass_data instance and...
9661 (make_pass_stack_adjustments): ...new function.
9662 * combine.c (pass_combine): Convert from a global struct to a subclass
9663 of rtl_opt_pass along with...
9664 (pass_data_combine): ...new pass_data instance and...
9665 (make_pass_combine): ...new function.
9666 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
9667 global struct to a subclass of rtl_opt_pass along with...
9668 (pass_data_compare_elim_after_reload): ...new pass_data instance
9670 (make_pass_compare_elim_after_reload): ...new function.
9671 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
9672 of rtl_opt_pass along with...
9673 (pass_data_rtl_cprop): ...new pass_data instance and...
9674 (make_pass_rtl_cprop): ...new function.
9675 * cse.c (pass_cse): Convert from a global struct to a subclass of
9676 rtl_opt_pass along with...
9677 (pass_data_cse): ...new pass_data instance and...
9678 (make_pass_cse): ...new function.
9679 (pass_cse2): Convert from a global struct to a subclass of
9680 rtl_opt_pass along with...
9681 (pass_data_cse2): ...new pass_data instance and...
9682 (make_pass_cse2): ...new function.
9683 (pass_cse_after_global_opts): Convert from a global struct to a
9684 subclass of rtl_opt_pass along with...
9685 (pass_data_cse_after_global_opts): ...new pass_data instance and...
9686 (make_pass_cse_after_global_opts): ...new function.
9687 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
9688 of rtl_opt_pass along with...
9689 (pass_data_ud_rtl_dce): ...new pass_data instance and...
9690 (make_pass_ud_rtl_dce): ...new function.
9691 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
9692 rtl_opt_pass along with...
9693 (pass_data_fast_rtl_dce): ...new pass_data instance and...
9694 (make_pass_fast_rtl_dce): ...new function.
9695 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
9696 a subclass of rtl_opt_pass along with...
9697 (pass_data_df_initialize_opt): ...new pass_data instance and...
9698 (make_pass_df_initialize_opt): ...new function.
9699 (pass_df_initialize_no_opt): Convert from a global struct to a
9700 subclass of rtl_opt_pass along with...
9701 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
9702 (make_pass_df_initialize_no_opt): ...new function.
9703 (pass_df_finish): Convert from a global struct to a subclass of
9704 rtl_opt_pass along with...
9705 (pass_data_df_finish): ...new pass_data instance and...
9706 (make_pass_df_finish): ...new function.
9707 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
9708 rtl_opt_pass along with...
9709 (pass_data_rtl_dse1): ...new pass_data instance and...
9710 (make_pass_rtl_dse1): ...new function.
9711 (pass_rtl_dse2): Convert from a global struct to a subclass of
9712 rtl_opt_pass along with...
9713 (pass_data_rtl_dse2): ...new pass_data instance and...
9714 (make_pass_rtl_dse2): ...new function.
9715 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
9716 subclass of rtl_opt_pass along with...
9717 (pass_data_dwarf2_frame): ...new pass_data instance and...
9718 (make_pass_dwarf2_frame): ...new function.
9719 * except.c (pass_set_nothrow_function_flags): Convert from a global
9720 struct to a subclass of rtl_opt_pass along with...
9721 (pass_data_set_nothrow_function_flags): ...new pass_data instance
9723 (make_pass_set_nothrow_function_flags): ...new function.
9724 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
9725 subclass of rtl_opt_pass along with...
9726 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
9728 (make_pass_convert_to_eh_region_ranges): ...new function.
9729 * final.c (pass_compute_alignments): Convert from a global struct to a
9730 subclass of rtl_opt_pass along with...
9731 (pass_data_compute_alignments): ...new pass_data instance and...
9732 (make_pass_compute_alignments): ...new function.
9733 (pass_final): Convert from a global struct to a subclass of
9734 rtl_opt_pass along with...
9735 (pass_data_final): ...new pass_data instance and...
9736 (make_pass_final): ...new function.
9737 (pass_shorten_branches): Convert from a global struct to a subclass of
9738 rtl_opt_pass along with...
9739 (pass_data_shorten_branches): ...new pass_data instance and...
9740 (make_pass_shorten_branches): ...new function.
9741 (pass_clean_state): Convert from a global struct to a subclass of
9742 rtl_opt_pass along with...
9743 (pass_data_clean_state): ...new pass_data instance and...
9744 (make_pass_clean_state): ...new function.
9745 * function.c (pass_instantiate_virtual_regs): Convert from a global
9746 struct to a subclass of rtl_opt_pass along with...
9747 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
9748 (make_pass_instantiate_virtual_regs): ...new function.
9749 (pass_leaf_regs): Convert from a global struct to a subclass of
9750 rtl_opt_pass along with...
9751 (pass_data_leaf_regs): ...new pass_data instance and...
9752 (make_pass_leaf_regs): ...new function.
9753 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
9754 subclass of rtl_opt_pass along with...
9755 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
9757 (make_pass_thread_prologue_and_epilogue): ...new function.
9758 (pass_match_asm_constraints): Convert from a global struct to a
9759 subclass of rtl_opt_pass along with...
9760 (pass_data_match_asm_constraints): ...new pass_data instance and...
9761 (make_pass_match_asm_constraints): ...new function.
9762 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
9763 subclass of rtl_opt_pass along with...
9764 (pass_data_rtl_fwprop): ...new pass_data instance and...
9765 (make_pass_rtl_fwprop): ...new function.
9766 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
9767 rtl_opt_pass along with...
9768 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
9769 (make_pass_rtl_fwprop_addr): ...new function.
9770 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
9771 rtl_opt_pass along with...
9772 (pass_data_rtl_pre): ...new pass_data instance and...
9773 (make_pass_rtl_pre): ...new function.
9774 (pass_rtl_hoist): Convert from a global struct to a subclass of
9775 rtl_opt_pass along with...
9776 (pass_data_rtl_hoist): ...new pass_data instance and...
9777 (make_pass_rtl_hoist): ...new function.
9778 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
9779 subclass of gimple_opt_pass along with...
9780 (pass_data_lower_cf): ...new pass_data instance and...
9781 (make_pass_lower_cf): ...new function.
9782 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
9783 from a global struct to a subclass of gimple_opt_pass along with...
9784 (pass_data_strength_reduction): ...new pass_data instance and...
9785 (make_pass_strength_reduction): ...new function.
9786 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
9787 of rtl_opt_pass along with...
9788 (pass_data_rtl_ifcvt): ...new pass_data instance and...
9789 (make_pass_rtl_ifcvt): ...new function.
9790 (pass_if_after_combine): Convert from a global struct to a subclass of
9791 rtl_opt_pass along with...
9792 (pass_data_if_after_combine): ...new pass_data instance and...
9793 (make_pass_if_after_combine): ...new function.
9794 (pass_if_after_reload): Convert from a global struct to a subclass of
9795 rtl_opt_pass along with...
9796 (pass_data_if_after_reload): ...new pass_data instance and...
9797 (make_pass_if_after_reload): ...new function.
9798 * init-regs.c (pass_initialize_regs): Convert from a global struct to
9799 a subclass of rtl_opt_pass along with...
9800 (pass_data_initialize_regs): ...new pass_data instance and...
9801 (make_pass_initialize_regs): ...new function.
9802 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
9803 of ipa_opt_pass_d along with...
9804 (pass_data_ipa_cp): ...new pass_data instance and...
9805 (make_pass_ipa_cp): ...new function.
9806 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
9807 global struct to a subclass of gimple_opt_pass along with...
9808 (pass_data_inline_parameters): ...new pass_data instance and...
9809 (make_pass_inline_parameters): ...new function.
9810 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
9811 subclass of gimple_opt_pass along with...
9812 (pass_data_early_inline): ...new pass_data instance and...
9813 (make_pass_early_inline): ...new function.
9814 (pass_ipa_inline): Convert from a global struct to a subclass of
9815 ipa_opt_pass_d along with...
9816 (pass_data_ipa_inline): ...new pass_data instance and...
9817 (make_pass_ipa_inline): ...new function.
9818 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
9819 struct to a subclass of gimple_opt_pass along with...
9820 (pass_data_local_pure_const): ...new pass_data instance and...
9821 (make_pass_local_pure_const): ...new function.
9822 (pass_ipa_pure_const): Convert from a global struct to a subclass of
9823 ipa_opt_pass_d along with...
9824 (pass_data_ipa_pure_const): ...new pass_data instance and...
9825 (make_pass_ipa_pure_const): ...new function.
9826 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
9827 to a subclass of ipa_opt_pass_d along with...
9828 (pass_data_ipa_reference): ...new pass_data instance and...
9829 (make_pass_ipa_reference): ...new function.
9830 * ipa-split.c (pass_split_functions): Convert from a global struct to
9831 a subclass of gimple_opt_pass along with...
9832 (pass_data_split_functions): ...new pass_data instance and...
9833 (make_pass_split_functions): ...new function.
9834 (pass_feedback_split_functions): Convert from a global struct to a
9835 subclass of gimple_opt_pass along with...
9836 (pass_data_feedback_split_functions): ...new pass_data instance and...
9837 (make_pass_feedback_split_functions): ...new function.
9838 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
9839 global struct to a subclass of simple_ipa_opt_pass along with...
9840 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
9842 (make_pass_ipa_function_and_variable_visibility): ...new function.
9843 (pass_ipa_free_inline_summary): Convert from a global struct to a
9844 subclass of simple_ipa_opt_pass along with...
9845 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
9846 (make_pass_ipa_free_inline_summary): ...new function.
9847 (pass_ipa_whole_program_visibility): Convert from a global struct to a
9848 subclass of ipa_opt_pass_d along with...
9849 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
9851 (make_pass_ipa_whole_program_visibility): ...new function.
9852 (pass_ipa_profile): Convert from a global struct to a subclass of
9853 ipa_opt_pass_d along with...
9854 (pass_data_ipa_profile): ...new pass_data instance and...
9855 (make_pass_ipa_profile): ...new function.
9856 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
9857 ipa_opt_pass_d along with...
9858 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
9859 (make_pass_ipa_cdtor_merge): ...new function.
9860 * ira.c (pass_ira): Convert from a global struct to a subclass of
9861 rtl_opt_pass along with...
9862 (pass_data_ira): ...new pass_data instance and...
9863 (make_pass_ira): ...new function.
9864 (pass_reload): Convert from a global struct to a subclass of
9865 rtl_opt_pass along with...
9866 (pass_data_reload): ...new pass_data instance and...
9867 (make_pass_reload): ...new function.
9868 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
9869 subclass of rtl_opt_pass along with...
9870 (pass_data_cleanup_barriers): ...new pass_data instance and...
9871 (make_pass_cleanup_barriers): ...new function.
9872 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
9873 of rtl_opt_pass along with...
9874 (pass_data_loop2): ...new pass_data instance and...
9875 (make_pass_loop2): ...new function.
9876 (pass_rtl_loop_init): Convert from a global struct to a subclass of
9877 rtl_opt_pass along with...
9878 (pass_data_rtl_loop_init): ...new pass_data instance and...
9879 (make_pass_rtl_loop_init): ...new function.
9880 (pass_rtl_loop_done): Convert from a global struct to a subclass of
9881 rtl_opt_pass along with...
9882 (pass_data_rtl_loop_done): ...new pass_data instance and...
9883 (make_pass_rtl_loop_done): ...new function.
9884 (pass_rtl_move_loop_invariants): Convert from a global struct to a
9885 subclass of rtl_opt_pass along with...
9886 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
9887 (make_pass_rtl_move_loop_invariants): ...new function.
9888 (pass_rtl_unswitch): Convert from a global struct to a subclass of
9889 rtl_opt_pass along with...
9890 (pass_data_rtl_unswitch): ...new pass_data instance and...
9891 (make_pass_rtl_unswitch): ...new function.
9892 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
9893 subclass of rtl_opt_pass along with...
9894 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
9896 (make_pass_rtl_unroll_and_peel_loops): ...new function.
9897 (pass_rtl_doloop): Convert from a global struct to a subclass of
9898 rtl_opt_pass along with...
9899 (pass_data_rtl_doloop): ...new pass_data instance and...
9900 (make_pass_rtl_doloop): ...new function.
9901 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
9902 a subclass of rtl_opt_pass along with...
9903 (pass_data_lower_subreg): ...new pass_data instance and...
9904 (make_pass_lower_subreg): ...new function.
9905 (pass_lower_subreg2): Convert from a global struct to a subclass of
9906 rtl_opt_pass along with...
9907 (pass_data_lower_subreg2): ...new pass_data instance and...
9908 (make_pass_lower_subreg2): ...new function.
9909 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
9910 struct to a subclass of ipa_opt_pass_d along with...
9911 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
9912 (make_pass_ipa_lto_gimple_out): ...new function.
9913 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
9914 of ipa_opt_pass_d along with...
9915 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
9916 (make_pass_ipa_lto_finish_out): ...new function.
9917 * mode-switching.c (pass_mode_switching): Convert from a global struct
9918 to a subclass of rtl_opt_pass along with...
9919 (pass_data_mode_switching): ...new pass_data instance and...
9920 (make_pass_mode_switching): ...new function.
9921 * modulo-sched.c (pass_sms): Convert from a global struct to a
9922 subclass of rtl_opt_pass along with...
9923 (pass_data_sms): ...new pass_data instance and...
9924 (make_pass_sms): ...new function.
9925 * omp-low.c (pass_expand_omp): Convert from a global struct to a
9926 subclass of gimple_opt_pass along with...
9927 (pass_data_expand_omp): ...new pass_data instance and...
9928 (make_pass_expand_omp): ...new function.
9929 (pass_lower_omp): Convert from a global struct to a subclass of
9930 gimple_opt_pass along with...
9931 (pass_data_lower_omp): ...new pass_data instance and...
9932 (make_pass_lower_omp): ...new function.
9933 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
9934 of gimple_opt_pass along with...
9935 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
9936 (make_pass_diagnose_omp_blocks): ...new function.
9937 * passes.c (pass_early_local_passes): Convert from a global struct to
9938 a subclass of simple_ipa_opt_pass along with...
9939 (pass_data_early_local_passes): ...new pass_data instance and...
9940 (make_pass_early_local_passes): ...new function.
9941 (pass_all_early_optimizations): Convert from a global struct to a
9942 subclass of gimple_opt_pass along with...
9943 (pass_data_all_early_optimizations): ...new pass_data instance and...
9944 (make_pass_all_early_optimizations): ...new function.
9945 (pass_all_optimizations): Convert from a global struct to a subclass
9946 of gimple_opt_pass along with...
9947 (pass_data_all_optimizations): ...new pass_data instance and...
9948 (make_pass_all_optimizations): ...new function.
9949 (pass_all_optimizations_g): Convert from a global struct to a subclass
9950 of gimple_opt_pass along with...
9951 (pass_data_all_optimizations_g): ...new pass_data instance and...
9952 (make_pass_all_optimizations_g): ...new function.
9953 (pass_rest_of_compilation): Convert from a global struct to a subclass
9954 of rtl_opt_pass along with...
9955 (pass_data_rest_of_compilation): ...new pass_data instance and...
9956 (make_pass_rest_of_compilation): ...new function.
9957 (pass_postreload): Convert from a global struct to a subclass of
9958 rtl_opt_pass along with...
9959 (pass_data_postreload): ...new pass_data instance and...
9960 (make_pass_postreload): ...new function.
9961 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
9962 subclass of rtl_opt_pass along with...
9963 (pass_data_gcse2): ...new pass_data instance and...
9964 (make_pass_gcse2): ...new function.
9965 * postreload.c (pass_postreload_cse): Convert from a global struct to
9966 a subclass of rtl_opt_pass along with...
9967 (pass_data_postreload_cse): ...new pass_data instance and...
9968 (make_pass_postreload_cse): ...new function.
9969 * predict.c (pass_profile): Convert from a global struct to a subclass
9970 of gimple_opt_pass along with...
9971 (pass_data_profile): ...new pass_data instance and...
9972 (make_pass_profile): ...new function.
9973 (pass_strip_predict_hints): Convert from a global struct to a subclass
9974 of gimple_opt_pass along with...
9975 (pass_data_strip_predict_hints): ...new pass_data instance and...
9976 (make_pass_strip_predict_hints): ...new function.
9977 * recog.c (pass_peephole2): Convert from a global struct to a subclass
9978 of rtl_opt_pass along with...
9979 (pass_data_peephole2): ...new pass_data instance and...
9980 (make_pass_peephole2): ...new function.
9981 (pass_split_all_insns): Convert from a global struct to a subclass of
9982 rtl_opt_pass along with...
9983 (pass_data_split_all_insns): ...new pass_data instance and...
9984 (make_pass_split_all_insns): ...new function.
9985 (pass_split_after_reload): Convert from a global struct to a subclass
9986 of rtl_opt_pass along with...
9987 (pass_data_split_after_reload): ...new pass_data instance and...
9988 (make_pass_split_after_reload): ...new function.
9989 (pass_split_before_regstack): Convert from a global struct to a
9990 subclass of rtl_opt_pass along with...
9991 (pass_data_split_before_regstack): ...new pass_data instance and...
9992 (make_pass_split_before_regstack): ...new function.
9993 (pass_split_before_sched2): Convert from a global struct to a subclass
9994 of rtl_opt_pass along with...
9995 (pass_data_split_before_sched2): ...new pass_data instance and...
9996 (make_pass_split_before_sched2): ...new function.
9997 (pass_split_for_shorten_branches): Convert from a global struct to a
9998 subclass of rtl_opt_pass along with...
9999 (pass_data_split_for_shorten_branches): ...new pass_data instance
10001 (make_pass_split_for_shorten_branches): ...new function.
10002 * ree.c (pass_ree): Convert from a global struct to a subclass of
10003 rtl_opt_pass along with...
10004 (pass_data_ree): ...new pass_data instance and...
10005 (make_pass_ree): ...new function.
10006 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
10007 subclass of rtl_opt_pass along with...
10008 (pass_data_stack_regs): ...new pass_data instance and...
10009 (make_pass_stack_regs): ...new function.
10010 (pass_stack_regs_run): Convert from a global struct to a subclass of
10011 rtl_opt_pass along with...
10012 (pass_data_stack_regs_run): ...new pass_data instance and...
10013 (make_pass_stack_regs_run): ...new function.
10014 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
10015 subclass of rtl_opt_pass along with...
10016 (pass_data_cprop_hardreg): ...new pass_data instance and...
10017 (make_pass_cprop_hardreg): ...new function.
10018 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
10019 subclass of rtl_opt_pass along with...
10020 (pass_data_reginfo_init): ...new pass_data instance and...
10021 (make_pass_reginfo_init): ...new function.
10022 * regmove.c (pass_regmove): Convert from a global struct to a subclass
10023 of rtl_opt_pass along with...
10024 (pass_data_regmove): ...new pass_data instance and...
10025 (make_pass_regmove): ...new function.
10026 * regrename.c (pass_regrename): Convert from a global struct to a
10027 subclass of rtl_opt_pass along with...
10028 (pass_data_regrename): ...new pass_data instance and...
10029 (make_pass_regrename): ...new function.
10030 * reorg.c (pass_delay_slots): Convert from a global struct to a
10031 subclass of rtl_opt_pass along with...
10032 (pass_data_delay_slots): ...new pass_data instance and...
10033 (make_pass_delay_slots): ...new function.
10034 (pass_machine_reorg): Convert from a global struct to a subclass of
10035 rtl_opt_pass along with...
10036 (pass_data_machine_reorg): ...new pass_data instance and...
10037 (make_pass_machine_reorg): ...new function.
10038 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
10039 of rtl_opt_pass along with...
10040 (pass_data_sched): ...new pass_data instance and...
10041 (make_pass_sched): ...new function.
10042 (pass_sched2): Convert from a global struct to a subclass of
10043 rtl_opt_pass along with...
10044 (pass_data_sched2): ...new pass_data instance and...
10045 (make_pass_sched2): ...new function.
10046 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
10047 to a subclass of rtl_opt_pass along with...
10048 (pass_data_stack_ptr_mod): ...new pass_data instance and...
10049 (make_pass_stack_ptr_mod): ...new function.
10050 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
10051 to a subclass of rtl_opt_pass along with...
10052 (pass_data_rtl_store_motion): ...new pass_data instance and...
10053 (make_pass_rtl_store_motion): ...new function.
10054 * tracer.c (pass_tracer): Convert from a global struct to a subclass
10055 of gimple_opt_pass along with...
10056 (pass_data_tracer): ...new pass_data instance and...
10057 (make_pass_tracer): ...new function.
10058 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
10059 to a subclass of gimple_opt_pass along with...
10060 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
10061 (make_pass_diagnose_tm_blocks): ...new function.
10062 (pass_lower_tm): Convert from a global struct to a subclass of
10063 gimple_opt_pass along with...
10064 (pass_data_lower_tm): ...new pass_data instance and...
10065 (make_pass_lower_tm): ...new function.
10066 (pass_tm_init): Convert from a global struct to a subclass of
10067 gimple_opt_pass along with...
10068 (pass_data_tm_init): ...new pass_data instance and...
10069 (make_pass_tm_init): ...new function.
10070 (pass_tm_mark): Convert from a global struct to a subclass of
10071 gimple_opt_pass along with...
10072 (pass_data_tm_mark): ...new pass_data instance and...
10073 (make_pass_tm_mark): ...new function.
10074 (pass_tm_edges): Convert from a global struct to a subclass of
10075 gimple_opt_pass along with...
10076 (pass_data_tm_edges): ...new pass_data instance and...
10077 (make_pass_tm_edges): ...new function.
10078 (pass_tm_memopt): Convert from a global struct to a subclass of
10079 gimple_opt_pass along with...
10080 (pass_data_tm_memopt): ...new pass_data instance and...
10081 (make_pass_tm_memopt): ...new function.
10082 (pass_ipa_tm): Convert from a global struct to a subclass of
10083 simple_ipa_opt_pass along with...
10084 (pass_data_ipa_tm): ...new pass_data instance and...
10085 (make_pass_ipa_tm): ...new function.
10086 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
10087 subclass of gimple_opt_pass along with...
10088 (pass_data_call_cdce): ...new pass_data instance and...
10089 (make_pass_call_cdce): ...new function.
10090 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
10091 subclass of gimple_opt_pass along with...
10092 (pass_data_build_cfg): ...new pass_data instance and...
10093 (make_pass_build_cfg): ...new function.
10094 (pass_split_crit_edges): Convert from a global struct to a subclass of
10095 gimple_opt_pass along with...
10096 (pass_data_split_crit_edges): ...new pass_data instance and...
10097 (make_pass_split_crit_edges): ...new function.
10098 (pass_warn_function_return): Convert from a global struct to a
10099 subclass of gimple_opt_pass along with...
10100 (pass_data_warn_function_return): ...new pass_data instance and...
10101 (make_pass_warn_function_return): ...new function.
10102 (pass_warn_function_noreturn): Convert from a global struct to a
10103 subclass of gimple_opt_pass along with...
10104 (pass_data_warn_function_noreturn): ...new pass_data instance and...
10105 (make_pass_warn_function_noreturn): ...new function.
10106 (pass_warn_unused_result): Convert from a global struct to a subclass
10107 of gimple_opt_pass along with...
10108 (pass_data_warn_unused_result): ...new pass_data instance and...
10109 (make_pass_warn_unused_result): ...new function.
10110 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
10111 a subclass of gimple_opt_pass along with...
10112 (pass_data_merge_phi): ...new pass_data instance and...
10113 (make_pass_merge_phi): ...new function.
10114 * tree-complex.c (pass_lower_complex): Convert from a global struct to
10115 a subclass of gimple_opt_pass along with...
10116 (pass_data_lower_complex): ...new pass_data instance and...
10117 (make_pass_lower_complex): ...new function.
10118 (pass_lower_complex_O0): Convert from a global struct to a subclass of
10119 gimple_opt_pass along with...
10120 (pass_data_lower_complex_O0): ...new pass_data instance and...
10121 (make_pass_lower_complex_O0): ...new function.
10122 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
10123 subclass of gimple_opt_pass along with...
10124 (pass_data_lower_eh): ...new pass_data instance and...
10125 (make_pass_lower_eh): ...new function.
10126 (pass_refactor_eh): Convert from a global struct to a subclass of
10127 gimple_opt_pass along with...
10128 (pass_data_refactor_eh): ...new pass_data instance and...
10129 (make_pass_refactor_eh): ...new function.
10130 (pass_lower_resx): Convert from a global struct to a subclass of
10131 gimple_opt_pass along with...
10132 (pass_data_lower_resx): ...new pass_data instance and...
10133 (make_pass_lower_resx): ...new function.
10134 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
10135 of gimple_opt_pass along with...
10136 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
10137 (make_pass_lower_eh_dispatch): ...new function.
10138 (pass_cleanup_eh): Convert from a global struct to a subclass of
10139 gimple_opt_pass along with...
10140 (pass_data_cleanup_eh): ...new pass_data instance and...
10141 (make_pass_cleanup_eh): ...new function.
10142 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
10143 to a subclass of simple_ipa_opt_pass along with...
10144 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
10145 (make_pass_ipa_lower_emutls): ...new function.
10146 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
10147 a subclass of gimple_opt_pass along with...
10148 (pass_data_if_conversion): ...new pass_data instance and...
10149 (make_pass_if_conversion): ...new function.
10150 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
10151 subclass of gimple_opt_pass along with...
10152 (pass_data_build_ssa): ...new pass_data instance and...
10153 (make_pass_build_ssa): ...new function.
10154 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
10155 global struct to a subclass of gimple_opt_pass along with...
10156 (pass_data_loop_distribution): ...new pass_data instance and...
10157 (make_pass_loop_distribution): ...new function.
10158 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
10159 subclass of gimple_opt_pass along with...
10160 (pass_data_mudflap_1): ...new pass_data instance and...
10161 (make_pass_mudflap_1): ...new function.
10162 (pass_mudflap_2): Convert from a global struct to a subclass of
10163 gimple_opt_pass along with...
10164 (pass_data_mudflap_2): ...new pass_data instance and...
10165 (make_pass_mudflap_2): ...new function.
10166 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
10167 subclass of gimple_opt_pass along with...
10168 (pass_data_mudflap_1): ...new pass_data instance and...
10169 (make_pass_mudflap_1): ...new function.
10170 (pass_mudflap_2): Convert from a global struct to a subclass of
10171 gimple_opt_pass along with...
10172 (pass_data_mudflap_2): ...new pass_data instance and...
10173 (make_pass_mudflap_2): ...new function.
10174 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
10175 gimple_opt_pass along with...
10176 (pass_data_nrv): ...new pass_data instance and...
10177 (make_pass_nrv): ...new function.
10178 (pass_return_slot): Convert from a global struct to a subclass of
10179 gimple_opt_pass along with...
10180 (pass_data_return_slot): ...new pass_data instance and...
10181 (make_pass_return_slot): ...new function.
10182 * tree-object-size.c (pass_object_sizes): Convert from a global struct
10183 to a subclass of gimple_opt_pass along with...
10184 (pass_data_object_sizes): ...new pass_data instance and...
10185 (make_pass_object_sizes): ...new function.
10186 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
10187 global struct to a subclass of gimple_opt_pass along with...
10188 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
10190 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10191 (pass_fixup_cfg): Convert from a global struct to a subclass of
10192 gimple_opt_pass along with...
10193 (pass_data_fixup_cfg): ...new pass_data instance and...
10194 (make_pass_fixup_cfg): ...new function.
10195 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
10196 (make_pass_mudflap_1): ...new function.
10197 (pass_mudflap_2): Replace declaration with that of...
10198 (make_pass_mudflap_2): ...new function.
10199 (pass_asan): Replace declaration with that of...
10200 (make_pass_asan): ...new function.
10201 (pass_asan_O0): Replace declaration with that of...
10202 (make_pass_asan_O0): ...new function.
10203 (pass_tsan): Replace declaration with that of...
10204 (make_pass_tsan): ...new function.
10205 (pass_tsan_O0): Replace declaration with that of...
10206 (make_pass_tsan_O0): ...new function.
10207 (pass_lower_cf): Replace declaration with that of...
10208 (make_pass_lower_cf): ...new function.
10209 (pass_refactor_eh): Replace declaration with that of...
10210 (make_pass_refactor_eh): ...new function.
10211 (pass_lower_eh): Replace declaration with that of...
10212 (make_pass_lower_eh): ...new function.
10213 (pass_lower_eh_dispatch): Replace declaration with that of...
10214 (make_pass_lower_eh_dispatch): ...new function.
10215 (pass_lower_resx): Replace declaration with that of...
10216 (make_pass_lower_resx): ...new function.
10217 (pass_build_cfg): Replace declaration with that of...
10218 (make_pass_build_cfg): ...new function.
10219 (pass_early_tree_profile): Replace declaration with that of...
10220 (make_pass_early_tree_profile): ...new function.
10221 (pass_cleanup_eh): Replace declaration with that of...
10222 (make_pass_cleanup_eh): ...new function.
10223 (pass_sra): Replace declaration with that of...
10224 (make_pass_sra): ...new function.
10225 (pass_sra_early): Replace declaration with that of...
10226 (make_pass_sra_early): ...new function.
10227 (pass_early_ipa_sra): Replace declaration with that of...
10228 (make_pass_early_ipa_sra): ...new function.
10229 (pass_tail_recursion): Replace declaration with that of...
10230 (make_pass_tail_recursion): ...new function.
10231 (pass_tail_calls): Replace declaration with that of...
10232 (make_pass_tail_calls): ...new function.
10233 (pass_tree_loop): Replace declaration with that of...
10234 (make_pass_tree_loop): ...new function.
10235 (pass_tree_loop_init): Replace declaration with that of...
10236 (make_pass_tree_loop_init): ...new function.
10237 (pass_lim): Replace declaration with that of...
10238 (make_pass_lim): ...new function.
10239 (pass_tree_unswitch): Replace declaration with that of...
10240 (make_pass_tree_unswitch): ...new function.
10241 (pass_predcom): Replace declaration with that of...
10242 (make_pass_predcom): ...new function.
10243 (pass_iv_canon): Replace declaration with that of...
10244 (make_pass_iv_canon): ...new function.
10245 (pass_scev_cprop): Replace declaration with that of...
10246 (make_pass_scev_cprop): ...new function.
10247 (pass_empty_loop): Replace declaration with that of...
10248 (make_pass_empty_loop): ...new function.
10249 (pass_record_bounds): Replace declaration with that of...
10250 (make_pass_record_bounds): ...new function.
10251 (pass_graphite): Replace declaration with that of...
10252 (make_pass_graphite): ...new function.
10253 (pass_graphite_transforms): Replace declaration with that of...
10254 (make_pass_graphite_transforms): ...new function.
10255 (pass_if_conversion): Replace declaration with that of...
10256 (make_pass_if_conversion): ...new function.
10257 (pass_loop_distribution): Replace declaration with that of...
10258 (make_pass_loop_distribution): ...new function.
10259 (pass_vectorize): Replace declaration with that of...
10260 (make_pass_vectorize): ...new function.
10261 (pass_slp_vectorize): Replace declaration with that of...
10262 (make_pass_slp_vectorize): ...new function.
10263 (pass_complete_unroll): Replace declaration with that of...
10264 (make_pass_complete_unroll): ...new function.
10265 (pass_complete_unrolli): Replace declaration with that of...
10266 (make_pass_complete_unrolli): ...new function.
10267 (pass_parallelize_loops): Replace declaration with that of...
10268 (make_pass_parallelize_loops): ...new function.
10269 (pass_loop_prefetch): Replace declaration with that of...
10270 (make_pass_loop_prefetch): ...new function.
10271 (pass_iv_optimize): Replace declaration with that of...
10272 (make_pass_iv_optimize): ...new function.
10273 (pass_tree_loop_done): Replace declaration with that of...
10274 (make_pass_tree_loop_done): ...new function.
10275 (pass_ch): Replace declaration with that of...
10276 (make_pass_ch): ...new function.
10277 (pass_ccp): Replace declaration with that of...
10278 (make_pass_ccp): ...new function.
10279 (pass_phi_only_cprop): Replace declaration with that of...
10280 (make_pass_phi_only_cprop): ...new function.
10281 (pass_build_ssa): Replace declaration with that of...
10282 (make_pass_build_ssa): ...new function.
10283 (pass_build_alias): Replace declaration with that of...
10284 (make_pass_build_alias): ...new function.
10285 (pass_build_ealias): Replace declaration with that of...
10286 (make_pass_build_ealias): ...new function.
10287 (pass_dominator): Replace declaration with that of...
10288 (make_pass_dominator): ...new function.
10289 (pass_dce): Replace declaration with that of...
10290 (make_pass_dce): ...new function.
10291 (pass_dce_loop): Replace declaration with that of...
10292 (make_pass_dce_loop): ...new function.
10293 (pass_cd_dce): Replace declaration with that of...
10294 (make_pass_cd_dce): ...new function.
10295 (pass_call_cdce): Replace declaration with that of...
10296 (make_pass_call_cdce): ...new function.
10297 (pass_merge_phi): Replace declaration with that of...
10298 (make_pass_merge_phi): ...new function.
10299 (pass_split_crit_edges): Replace declaration with that of...
10300 (make_pass_split_crit_edges): ...new function.
10301 (pass_pre): Replace declaration with that of...
10302 (make_pass_pre): ...new function.
10303 (pass_profile): Replace declaration with that of...
10304 (make_pass_profile): ...new function.
10305 (pass_strip_predict_hints): Replace declaration with that of...
10306 (make_pass_strip_predict_hints): ...new function.
10307 (pass_lower_complex_O0): Replace declaration with that of...
10308 (make_pass_lower_complex_O0): ...new function.
10309 (pass_lower_complex): Replace declaration with that of...
10310 (make_pass_lower_complex): ...new function.
10311 (pass_lower_vector): Replace declaration with that of...
10312 (make_pass_lower_vector): ...new function.
10313 (pass_lower_vector_ssa): Replace declaration with that of...
10314 (make_pass_lower_vector_ssa): ...new function.
10315 (pass_lower_omp): Replace declaration with that of...
10316 (make_pass_lower_omp): ...new function.
10317 (pass_diagnose_omp_blocks): Replace declaration with that of...
10318 (make_pass_diagnose_omp_blocks): ...new function.
10319 (pass_expand_omp): Replace declaration with that of...
10320 (make_pass_expand_omp): ...new function.
10321 (pass_expand_omp_ssa): Replace declaration with that of...
10322 (make_pass_expand_omp_ssa): ...new function.
10323 (pass_object_sizes): Replace declaration with that of...
10324 (make_pass_object_sizes): ...new function.
10325 (pass_strlen): Replace declaration with that of...
10326 (make_pass_strlen): ...new function.
10327 (pass_fold_builtins): Replace declaration with that of...
10328 (make_pass_fold_builtins): ...new function.
10329 (pass_stdarg): Replace declaration with that of...
10330 (make_pass_stdarg): ...new function.
10331 (pass_early_warn_uninitialized): Replace declaration with that of...
10332 (make_pass_early_warn_uninitialized): ...new function.
10333 (pass_late_warn_uninitialized): Replace declaration with that of...
10334 (make_pass_late_warn_uninitialized): ...new function.
10335 (pass_cse_reciprocals): Replace declaration with that of...
10336 (make_pass_cse_reciprocals): ...new function.
10337 (pass_cse_sincos): Replace declaration with that of...
10338 (make_pass_cse_sincos): ...new function.
10339 (pass_optimize_bswap): Replace declaration with that of...
10340 (make_pass_optimize_bswap): ...new function.
10341 (pass_optimize_widening_mul): Replace declaration with that of...
10342 (make_pass_optimize_widening_mul): ...new function.
10343 (pass_warn_function_return): Replace declaration with that of...
10344 (make_pass_warn_function_return): ...new function.
10345 (pass_warn_function_noreturn): Replace declaration with that of...
10346 (make_pass_warn_function_noreturn): ...new function.
10347 (pass_cselim): Replace declaration with that of...
10348 (make_pass_cselim): ...new function.
10349 (pass_phiopt): Replace declaration with that of...
10350 (make_pass_phiopt): ...new function.
10351 (pass_forwprop): Replace declaration with that of...
10352 (make_pass_forwprop): ...new function.
10353 (pass_phiprop): Replace declaration with that of...
10354 (make_pass_phiprop): ...new function.
10355 (pass_tree_ifcombine): Replace declaration with that of...
10356 (make_pass_tree_ifcombine): ...new function.
10357 (pass_dse): Replace declaration with that of...
10358 (make_pass_dse): ...new function.
10359 (pass_nrv): Replace declaration with that of...
10360 (make_pass_nrv): ...new function.
10361 (pass_rename_ssa_copies): Replace declaration with that of...
10362 (make_pass_rename_ssa_copies): ...new function.
10363 (pass_sink_code): Replace declaration with that of...
10364 (make_pass_sink_code): ...new function.
10365 (pass_fre): Replace declaration with that of...
10366 (make_pass_fre): ...new function.
10367 (pass_check_data_deps): Replace declaration with that of...
10368 (make_pass_check_data_deps): ...new function.
10369 (pass_copy_prop): Replace declaration with that of...
10370 (make_pass_copy_prop): ...new function.
10371 (pass_vrp): Replace declaration with that of...
10372 (make_pass_vrp): ...new function.
10373 (pass_uncprop): Replace declaration with that of...
10374 (make_pass_uncprop): ...new function.
10375 (pass_return_slot): Replace declaration with that of...
10376 (make_pass_return_slot): ...new function.
10377 (pass_reassoc): Replace declaration with that of...
10378 (make_pass_reassoc): ...new function.
10379 (pass_rebuild_cgraph_edges): Replace declaration with that of...
10380 (make_pass_rebuild_cgraph_edges): ...new function.
10381 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
10382 (make_pass_remove_cgraph_callee_edges): ...new function.
10383 (pass_build_cgraph_edges): Replace declaration with that of...
10384 (make_pass_build_cgraph_edges): ...new function.
10385 (pass_local_pure_const): Replace declaration with that of...
10386 (make_pass_local_pure_const): ...new function.
10387 (pass_tracer): Replace declaration with that of...
10388 (make_pass_tracer): ...new function.
10389 (pass_warn_unused_result): Replace declaration with that of...
10390 (make_pass_warn_unused_result): ...new function.
10391 (pass_diagnose_tm_blocks): Replace declaration with that of...
10392 (make_pass_diagnose_tm_blocks): ...new function.
10393 (pass_lower_tm): Replace declaration with that of...
10394 (make_pass_lower_tm): ...new function.
10395 (pass_tm_init): Replace declaration with that of...
10396 (make_pass_tm_init): ...new function.
10397 (pass_tm_mark): Replace declaration with that of...
10398 (make_pass_tm_mark): ...new function.
10399 (pass_tm_memopt): Replace declaration with that of...
10400 (make_pass_tm_memopt): ...new function.
10401 (pass_tm_edges): Replace declaration with that of...
10402 (make_pass_tm_edges): ...new function.
10403 (pass_split_functions): Replace declaration with that of...
10404 (make_pass_split_functions): ...new function.
10405 (pass_feedback_split_functions): Replace declaration with that of...
10406 (make_pass_feedback_split_functions): ...new function.
10407 (pass_strength_reduction): Replace declaration with that of...
10408 (make_pass_strength_reduction): ...new function.
10409 (pass_ipa_lower_emutls): Replace declaration with that of...
10410 (make_pass_ipa_lower_emutls): ...new function.
10411 (pass_ipa_function_and_variable_visibility): Replace declaration with
10413 (make_pass_ipa_function_and_variable_visibility): ...new function.
10414 (pass_ipa_tree_profile): Replace declaration with that of...
10415 (make_pass_ipa_tree_profile): ...new function.
10416 (pass_early_local_passes): Replace declaration with that of...
10417 (make_pass_early_local_passes): ...new function.
10418 (pass_ipa_whole_program_visibility): Replace declaration with that
10420 (make_pass_ipa_whole_program_visibility): ...new function.
10421 (pass_ipa_lto_gimple_out): Replace declaration with that of...
10422 (make_pass_ipa_lto_gimple_out): ...new function.
10423 (pass_ipa_increase_alignment): Replace declaration with that of...
10424 (make_pass_ipa_increase_alignment): ...new function.
10425 (pass_ipa_inline): Replace declaration with that of...
10426 (make_pass_ipa_inline): ...new function.
10427 (pass_ipa_free_lang_data): Replace declaration with that of...
10428 (make_pass_ipa_free_lang_data): ...new function.
10429 (pass_ipa_free_inline_summary): Replace declaration with that of...
10430 (make_pass_ipa_free_inline_summary): ...new function.
10431 (pass_ipa_cp): Replace declaration with that of...
10432 (make_pass_ipa_cp): ...new function.
10433 (pass_ipa_reference): Replace declaration with that of...
10434 (make_pass_ipa_reference): ...new function.
10435 (pass_ipa_pure_const): Replace declaration with that of...
10436 (make_pass_ipa_pure_const): ...new function.
10437 (pass_ipa_pta): Replace declaration with that of...
10438 (make_pass_ipa_pta): ...new function.
10439 (pass_ipa_lto_finish_out): Replace declaration with that of...
10440 (make_pass_ipa_lto_finish_out): ...new function.
10441 (pass_ipa_tm): Replace declaration with that of...
10442 (make_pass_ipa_tm): ...new function.
10443 (pass_ipa_profile): Replace declaration with that of...
10444 (make_pass_ipa_profile): ...new function.
10445 (pass_ipa_cdtor_merge): Replace declaration with that of...
10446 (make_pass_ipa_cdtor_merge): ...new function.
10447 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
10449 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10450 (pass_init_datastructures): Replace declaration with that of...
10451 (make_pass_init_datastructures): ...new function.
10452 (pass_fixup_cfg): Replace declaration with that of...
10453 (make_pass_fixup_cfg): ...new function.
10454 (pass_expand): Replace declaration with that of...
10455 (make_pass_expand): ...new function.
10456 (pass_instantiate_virtual_regs): Replace declaration with that of...
10457 (make_pass_instantiate_virtual_regs): ...new function.
10458 (pass_rtl_fwprop): Replace declaration with that of...
10459 (make_pass_rtl_fwprop): ...new function.
10460 (pass_rtl_fwprop_addr): Replace declaration with that of...
10461 (make_pass_rtl_fwprop_addr): ...new function.
10462 (pass_jump): Replace declaration with that of...
10463 (make_pass_jump): ...new function.
10464 (pass_jump2): Replace declaration with that of...
10465 (make_pass_jump2): ...new function.
10466 (pass_lower_subreg): Replace declaration with that of...
10467 (make_pass_lower_subreg): ...new function.
10468 (pass_cse): Replace declaration with that of...
10469 (make_pass_cse): ...new function.
10470 (pass_fast_rtl_dce): Replace declaration with that of...
10471 (make_pass_fast_rtl_dce): ...new function.
10472 (pass_ud_rtl_dce): Replace declaration with that of...
10473 (make_pass_ud_rtl_dce): ...new function.
10474 (pass_rtl_dce): Replace declaration with that of...
10475 (make_pass_rtl_dce): ...new function.
10476 (pass_rtl_dse1): Replace declaration with that of...
10477 (make_pass_rtl_dse1): ...new function.
10478 (pass_rtl_dse2): Replace declaration with that of...
10479 (make_pass_rtl_dse2): ...new function.
10480 (pass_rtl_dse3): Replace declaration with that of...
10481 (make_pass_rtl_dse3): ...new function.
10482 (pass_rtl_cprop): Replace declaration with that of...
10483 (make_pass_rtl_cprop): ...new function.
10484 (pass_rtl_pre): Replace declaration with that of...
10485 (make_pass_rtl_pre): ...new function.
10486 (pass_rtl_hoist): Replace declaration with that of...
10487 (make_pass_rtl_hoist): ...new function.
10488 (pass_rtl_store_motion): Replace declaration with that of...
10489 (make_pass_rtl_store_motion): ...new function.
10490 (pass_cse_after_global_opts): Replace declaration with that of...
10491 (make_pass_cse_after_global_opts): ...new function.
10492 (pass_rtl_ifcvt): Replace declaration with that of...
10493 (make_pass_rtl_ifcvt): ...new function.
10494 (pass_into_cfg_layout_mode): Replace declaration with that of...
10495 (make_pass_into_cfg_layout_mode): ...new function.
10496 (pass_outof_cfg_layout_mode): Replace declaration with that of...
10497 (make_pass_outof_cfg_layout_mode): ...new function.
10498 (pass_loop2): Replace declaration with that of...
10499 (make_pass_loop2): ...new function.
10500 (pass_rtl_loop_init): Replace declaration with that of...
10501 (make_pass_rtl_loop_init): ...new function.
10502 (pass_rtl_move_loop_invariants): Replace declaration with that of...
10503 (make_pass_rtl_move_loop_invariants): ...new function.
10504 (pass_rtl_unswitch): Replace declaration with that of...
10505 (make_pass_rtl_unswitch): ...new function.
10506 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
10507 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10508 (pass_rtl_doloop): Replace declaration with that of...
10509 (make_pass_rtl_doloop): ...new function.
10510 (pass_rtl_loop_done): Replace declaration with that of...
10511 (make_pass_rtl_loop_done): ...new function.
10512 (pass_web): Replace declaration with that of...
10513 (make_pass_web): ...new function.
10514 (pass_cse2): Replace declaration with that of...
10515 (make_pass_cse2): ...new function.
10516 (pass_df_initialize_opt): Replace declaration with that of...
10517 (make_pass_df_initialize_opt): ...new function.
10518 (pass_df_initialize_no_opt): Replace declaration with that of...
10519 (make_pass_df_initialize_no_opt): ...new function.
10520 (pass_reginfo_init): Replace declaration with that of...
10521 (make_pass_reginfo_init): ...new function.
10522 (pass_inc_dec): Replace declaration with that of...
10523 (make_pass_inc_dec): ...new function.
10524 (pass_stack_ptr_mod): Replace declaration with that of...
10525 (make_pass_stack_ptr_mod): ...new function.
10526 (pass_initialize_regs): Replace declaration with that of...
10527 (make_pass_initialize_regs): ...new function.
10528 (pass_combine): Replace declaration with that of...
10529 (make_pass_combine): ...new function.
10530 (pass_if_after_combine): Replace declaration with that of...
10531 (make_pass_if_after_combine): ...new function.
10532 (pass_ree): Replace declaration with that of...
10533 (make_pass_ree): ...new function.
10534 (pass_partition_blocks): Replace declaration with that of...
10535 (make_pass_partition_blocks): ...new function.
10536 (pass_match_asm_constraints): Replace declaration with that of...
10537 (make_pass_match_asm_constraints): ...new function.
10538 (pass_regmove): Replace declaration with that of...
10539 (make_pass_regmove): ...new function.
10540 (pass_split_all_insns): Replace declaration with that of...
10541 (make_pass_split_all_insns): ...new function.
10542 (pass_fast_rtl_byte_dce): Replace declaration with that of...
10543 (make_pass_fast_rtl_byte_dce): ...new function.
10544 (pass_lower_subreg2): Replace declaration with that of...
10545 (make_pass_lower_subreg2): ...new function.
10546 (pass_mode_switching): Replace declaration with that of...
10547 (make_pass_mode_switching): ...new function.
10548 (pass_sms): Replace declaration with that of...
10549 (make_pass_sms): ...new function.
10550 (pass_sched): Replace declaration with that of...
10551 (make_pass_sched): ...new function.
10552 (pass_ira): Replace declaration with that of...
10553 (make_pass_ira): ...new function.
10554 (pass_reload): Replace declaration with that of...
10555 (make_pass_reload): ...new function.
10556 (pass_clean_state): Replace declaration with that of...
10557 (make_pass_clean_state): ...new function.
10558 (pass_branch_prob): Replace declaration with that of...
10559 (make_pass_branch_prob): ...new function.
10560 (pass_value_profile_transformations): Replace declaration with that
10562 (make_pass_value_profile_transformations): ...new function.
10563 (pass_postreload_cse): Replace declaration with that of...
10564 (make_pass_postreload_cse): ...new function.
10565 (pass_gcse2): Replace declaration with that of...
10566 (make_pass_gcse2): ...new function.
10567 (pass_split_after_reload): Replace declaration with that of...
10568 (make_pass_split_after_reload): ...new function.
10569 (pass_branch_target_load_optimize1): Replace declaration with that
10571 (make_pass_branch_target_load_optimize1): ...new function.
10572 (pass_thread_prologue_and_epilogue): Replace declaration with that
10574 (make_pass_thread_prologue_and_epilogue): ...new function.
10575 (pass_stack_adjustments): Replace declaration with that of...
10576 (make_pass_stack_adjustments): ...new function.
10577 (pass_peephole2): Replace declaration with that of...
10578 (make_pass_peephole2): ...new function.
10579 (pass_if_after_reload): Replace declaration with that of...
10580 (make_pass_if_after_reload): ...new function.
10581 (pass_regrename): Replace declaration with that of...
10582 (make_pass_regrename): ...new function.
10583 (pass_cprop_hardreg): Replace declaration with that of...
10584 (make_pass_cprop_hardreg): ...new function.
10585 (pass_reorder_blocks): Replace declaration with that of...
10586 (make_pass_reorder_blocks): ...new function.
10587 (pass_branch_target_load_optimize2): Replace declaration with that
10589 (make_pass_branch_target_load_optimize2): ...new function.
10590 (pass_leaf_regs): Replace declaration with that of...
10591 (make_pass_leaf_regs): ...new function.
10592 (pass_split_before_sched2): Replace declaration with that of...
10593 (make_pass_split_before_sched2): ...new function.
10594 (pass_compare_elim_after_reload): Replace declaration with that of...
10595 (make_pass_compare_elim_after_reload): ...new function.
10596 (pass_sched2): Replace declaration with that of...
10597 (make_pass_sched2): ...new function.
10598 (pass_stack_regs): Replace declaration with that of...
10599 (make_pass_stack_regs): ...new function.
10600 (pass_stack_regs_run): Replace declaration with that of...
10601 (make_pass_stack_regs_run): ...new function.
10602 (pass_df_finish): Replace declaration with that of...
10603 (make_pass_df_finish): ...new function.
10604 (pass_compute_alignments): Replace declaration with that of...
10605 (make_pass_compute_alignments): ...new function.
10606 (pass_duplicate_computed_gotos): Replace declaration with that of...
10607 (make_pass_duplicate_computed_gotos): ...new function.
10608 (pass_variable_tracking): Replace declaration with that of...
10609 (make_pass_variable_tracking): ...new function.
10610 (pass_free_cfg): Replace declaration with that of...
10611 (make_pass_free_cfg): ...new function.
10612 (pass_machine_reorg): Replace declaration with that of...
10613 (make_pass_machine_reorg): ...new function.
10614 (pass_cleanup_barriers): Replace declaration with that of...
10615 (make_pass_cleanup_barriers): ...new function.
10616 (pass_delay_slots): Replace declaration with that of...
10617 (make_pass_delay_slots): ...new function.
10618 (pass_split_for_shorten_branches): Replace declaration with that of...
10619 (make_pass_split_for_shorten_branches): ...new function.
10620 (pass_split_before_regstack): Replace declaration with that of...
10621 (make_pass_split_before_regstack): ...new function.
10622 (pass_convert_to_eh_region_ranges): Replace declaration with that
10624 (make_pass_convert_to_eh_region_ranges): ...new function.
10625 (pass_shorten_branches): Replace declaration with that of...
10626 (make_pass_shorten_branches): ...new function.
10627 (pass_set_nothrow_function_flags): Replace declaration with that of...
10628 (make_pass_set_nothrow_function_flags): ...new function.
10629 (pass_dwarf2_frame): Replace declaration with that of...
10630 (make_pass_dwarf2_frame): ...new function.
10631 (pass_final): Replace declaration with that of...
10632 (make_pass_final): ...new function.
10633 (pass_rtl_seqabstr): Replace declaration with that of...
10634 (make_pass_rtl_seqabstr): ...new function.
10635 (pass_release_ssa_names): Replace declaration with that of...
10636 (make_pass_release_ssa_names): ...new function.
10637 (pass_early_inline): Replace declaration with that of...
10638 (make_pass_early_inline): ...new function.
10639 (pass_inline_parameters): Replace declaration with that of...
10640 (make_pass_inline_parameters): ...new function.
10641 (pass_update_address_taken): Replace declaration with that of...
10642 (make_pass_update_address_taken): ...new function.
10643 (pass_convert_switch): Replace declaration with that of...
10644 (make_pass_convert_switch): ...new function.
10645 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
10646 to a subclass of simple_ipa_opt_pass along with...
10647 (pass_data_ipa_tree_profile): ...new pass_data instance and...
10648 (make_pass_ipa_tree_profile): ...new function.
10649 * tree-sra.c (pass_sra_early): Convert from a global struct to a
10650 subclass of gimple_opt_pass along with...
10651 (pass_data_sra_early): ...new pass_data instance and...
10652 (make_pass_sra_early): ...new function.
10653 (pass_sra): Convert from a global struct to a subclass of
10654 gimple_opt_pass along with...
10655 (pass_data_sra): ...new pass_data instance and...
10656 (make_pass_sra): ...new function.
10657 (pass_early_ipa_sra): Convert from a global struct to a subclass of
10658 gimple_opt_pass along with...
10659 (pass_data_early_ipa_sra): ...new pass_data instance and...
10660 (make_pass_early_ipa_sra): ...new function.
10661 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
10662 subclass of gimple_opt_pass along with...
10663 (pass_data_ccp): ...new pass_data instance and...
10664 (make_pass_ccp): ...new function.
10665 (pass_fold_builtins): Convert from a global struct to a subclass of
10666 gimple_opt_pass along with...
10667 (pass_data_fold_builtins): ...new pass_data instance and...
10668 (make_pass_fold_builtins): ...new function.
10669 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
10670 subclass of gimple_opt_pass along with...
10671 (pass_data_copy_prop): ...new pass_data instance and...
10672 (make_pass_copy_prop): ...new function.
10673 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
10674 global struct to a subclass of gimple_opt_pass along with...
10675 (pass_data_rename_ssa_copies): ...new pass_data instance and...
10676 (make_pass_rename_ssa_copies): ...new function.
10677 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
10678 subclass of gimple_opt_pass along with...
10679 (pass_data_dce): ...new pass_data instance and...
10680 (make_pass_dce): ...new function.
10681 (pass_dce_loop): Convert from a global struct to a subclass of
10682 gimple_opt_pass along with...
10683 (pass_data_dce_loop): ...new pass_data instance and...
10684 (make_pass_dce_loop): ...new function.
10685 (pass_cd_dce): Convert from a global struct to a subclass of
10686 gimple_opt_pass along with...
10687 (pass_data_cd_dce): ...new pass_data instance and...
10688 (make_pass_cd_dce): ...new function.
10689 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
10690 subclass of gimple_opt_pass along with...
10691 (pass_data_dominator): ...new pass_data instance and...
10692 (make_pass_dominator): ...new function.
10693 (pass_phi_only_cprop): Convert from a global struct to a subclass of
10694 gimple_opt_pass along with...
10695 (pass_data_phi_only_cprop): ...new pass_data instance and...
10696 (make_pass_phi_only_cprop): ...new function.
10697 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
10698 subclass of gimple_opt_pass along with...
10699 (pass_data_dse): ...new pass_data instance and...
10700 (make_pass_dse): ...new function.
10701 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
10702 a subclass of gimple_opt_pass along with...
10703 (pass_data_forwprop): ...new pass_data instance and...
10704 (make_pass_forwprop): ...new function.
10705 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
10706 struct to a subclass of gimple_opt_pass along with...
10707 (pass_data_tree_ifcombine): ...new pass_data instance and...
10708 (make_pass_tree_ifcombine): ...new function.
10709 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
10710 subclass of gimple_opt_pass along with...
10711 (pass_data_ch): ...new pass_data instance and...
10712 (make_pass_ch): ...new function.
10713 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
10714 subclass of gimple_opt_pass along with...
10715 (pass_data_tree_loop): ...new pass_data instance and...
10716 (make_pass_tree_loop): ...new function.
10717 (pass_tree_loop_init): Convert from a global struct to a subclass of
10718 gimple_opt_pass along with...
10719 (pass_data_tree_loop_init): ...new pass_data instance and...
10720 (make_pass_tree_loop_init): ...new function.
10721 (pass_lim): Convert from a global struct to a subclass of
10722 gimple_opt_pass along with...
10723 (pass_data_lim): ...new pass_data instance and...
10724 (make_pass_lim): ...new function.
10725 (pass_tree_unswitch): Convert from a global struct to a subclass of
10726 gimple_opt_pass along with...
10727 (pass_data_tree_unswitch): ...new pass_data instance and...
10728 (make_pass_tree_unswitch): ...new function.
10729 (pass_predcom): Convert from a global struct to a subclass of
10730 gimple_opt_pass along with...
10731 (pass_data_predcom): ...new pass_data instance and...
10732 (make_pass_predcom): ...new function.
10733 (pass_vectorize): Convert from a global struct to a subclass of
10734 gimple_opt_pass along with...
10735 (pass_data_vectorize): ...new pass_data instance and...
10736 (make_pass_vectorize): ...new function.
10737 (pass_graphite): Convert from a global struct to a subclass of
10738 gimple_opt_pass along with...
10739 (pass_data_graphite): ...new pass_data instance and...
10740 (make_pass_graphite): ...new function.
10741 (pass_graphite_transforms): Convert from a global struct to a subclass
10742 of gimple_opt_pass along with...
10743 (pass_data_graphite_transforms): ...new pass_data instance and...
10744 (make_pass_graphite_transforms): ...new function.
10745 (pass_check_data_deps): Convert from a global struct to a subclass of
10746 gimple_opt_pass along with...
10747 (pass_data_check_data_deps): ...new pass_data instance and...
10748 (make_pass_check_data_deps): ...new function.
10749 (pass_iv_canon): Convert from a global struct to a subclass of
10750 gimple_opt_pass along with...
10751 (pass_data_iv_canon): ...new pass_data instance and...
10752 (make_pass_iv_canon): ...new function.
10753 (pass_scev_cprop): Convert from a global struct to a subclass of
10754 gimple_opt_pass along with...
10755 (pass_data_scev_cprop): ...new pass_data instance and...
10756 (make_pass_scev_cprop): ...new function.
10757 (pass_record_bounds): Convert from a global struct to a subclass of
10758 gimple_opt_pass along with...
10759 (pass_data_record_bounds): ...new pass_data instance and...
10760 (make_pass_record_bounds): ...new function.
10761 (pass_complete_unroll): Convert from a global struct to a subclass of
10762 gimple_opt_pass along with...
10763 (pass_data_complete_unroll): ...new pass_data instance and...
10764 (make_pass_complete_unroll): ...new function.
10765 (pass_complete_unrolli): Convert from a global struct to a subclass of
10766 gimple_opt_pass along with...
10767 (pass_data_complete_unrolli): ...new pass_data instance and...
10768 (make_pass_complete_unrolli): ...new function.
10769 (pass_parallelize_loops): Convert from a global struct to a subclass
10770 of gimple_opt_pass along with...
10771 (pass_data_parallelize_loops): ...new pass_data instance and...
10772 (make_pass_parallelize_loops): ...new function.
10773 (pass_loop_prefetch): Convert from a global struct to a subclass of
10774 gimple_opt_pass along with...
10775 (pass_data_loop_prefetch): ...new pass_data instance and...
10776 (make_pass_loop_prefetch): ...new function.
10777 (pass_iv_optimize): Convert from a global struct to a subclass of
10778 gimple_opt_pass along with...
10779 (pass_data_iv_optimize): ...new pass_data instance and...
10780 (make_pass_iv_optimize): ...new function.
10781 (pass_tree_loop_done): Convert from a global struct to a subclass of
10782 gimple_opt_pass along with...
10783 (pass_data_tree_loop_done): ...new pass_data instance and...
10784 (make_pass_tree_loop_done): ...new function.
10785 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
10786 struct to a subclass of gimple_opt_pass along with...
10787 (pass_data_cse_reciprocals): ...new pass_data instance and...
10788 (make_pass_cse_reciprocals): ...new function.
10789 (pass_cse_sincos): Convert from a global struct to a subclass of
10790 gimple_opt_pass along with...
10791 (pass_data_cse_sincos): ...new pass_data instance and...
10792 (make_pass_cse_sincos): ...new function.
10793 (pass_optimize_bswap): Convert from a global struct to a subclass of
10794 gimple_opt_pass along with...
10795 (pass_data_optimize_bswap): ...new pass_data instance and...
10796 (make_pass_optimize_bswap): ...new function.
10797 (pass_optimize_widening_mul): Convert from a global struct to a
10798 subclass of gimple_opt_pass along with...
10799 (pass_data_optimize_widening_mul): ...new pass_data instance and...
10800 (make_pass_optimize_widening_mul): ...new function.
10801 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
10802 subclass of gimple_opt_pass along with...
10803 (pass_data_phiopt): ...new pass_data instance and...
10804 (make_pass_phiopt): ...new function.
10805 (pass_cselim): Convert from a global struct to a subclass of
10806 gimple_opt_pass along with...
10807 (pass_data_cselim): ...new pass_data instance and...
10808 (make_pass_cselim): ...new function.
10809 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
10810 subclass of gimple_opt_pass along with...
10811 (pass_data_phiprop): ...new pass_data instance and...
10812 (make_pass_phiprop): ...new function.
10813 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
10814 subclass of gimple_opt_pass along with...
10815 (pass_data_pre): ...new pass_data instance and...
10816 (make_pass_pre): ...new function.
10817 (pass_fre): Convert from a global struct to a subclass of
10818 gimple_opt_pass along with...
10819 (pass_data_fre): ...new pass_data instance and...
10820 (make_pass_fre): ...new function.
10821 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
10822 subclass of gimple_opt_pass along with...
10823 (pass_data_reassoc): ...new pass_data instance and...
10824 (make_pass_reassoc): ...new function.
10825 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
10826 subclass of gimple_opt_pass along with...
10827 (pass_data_sink_code): ...new pass_data instance and...
10828 (make_pass_sink_code): ...new function.
10829 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
10830 subclass of gimple_opt_pass along with...
10831 (pass_data_strlen): ...new pass_data instance and...
10832 (make_pass_strlen): ...new function.
10833 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
10834 struct to a subclass of gimple_opt_pass along with...
10835 (pass_data_build_alias): ...new pass_data instance and...
10836 (make_pass_build_alias): ...new function.
10837 (pass_build_ealias): Convert from a global struct to a subclass of
10838 gimple_opt_pass along with...
10839 (pass_data_build_ealias): ...new pass_data instance and...
10840 (make_pass_build_ealias): ...new function.
10841 (pass_ipa_pta): Convert from a global struct to a subclass of
10842 simple_ipa_opt_pass along with...
10843 (pass_data_ipa_pta): ...new pass_data instance and...
10844 (make_pass_ipa_pta): ...new function.
10845 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
10846 subclass of gimple_opt_pass along with...
10847 (pass_data_uncprop): ...new pass_data instance and...
10848 (make_pass_uncprop): ...new function.
10849 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
10850 global struct to a subclass of gimple_opt_pass along with...
10851 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
10852 (make_pass_late_warn_uninitialized): ...new function.
10853 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
10854 to a subclass of gimple_opt_pass along with...
10855 (pass_data_init_datastructures): ...new pass_data instance and...
10856 (make_pass_init_datastructures): ...new function.
10857 (pass_early_warn_uninitialized): Convert from a global struct to a
10858 subclass of gimple_opt_pass along with...
10859 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
10860 (make_pass_early_warn_uninitialized): ...new function.
10861 (pass_update_address_taken): Convert from a global struct to a
10862 subclass of gimple_opt_pass along with...
10863 (pass_data_update_address_taken): ...new pass_data instance and...
10864 (make_pass_update_address_taken): ...new function.
10865 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
10866 struct to a subclass of gimple_opt_pass along with...
10867 (pass_data_release_ssa_names): ...new pass_data instance and...
10868 (make_pass_release_ssa_names): ...new function.
10869 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
10870 subclass of gimple_opt_pass along with...
10871 (pass_data_stdarg): ...new pass_data instance and...
10872 (make_pass_stdarg): ...new function.
10873 * tree-switch-conversion.c (pass_convert_switch): Convert from a
10874 global struct to a subclass of gimple_opt_pass along with...
10875 (pass_data_convert_switch): ...new pass_data instance and...
10876 (make_pass_convert_switch): ...new function.
10877 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
10878 to a subclass of gimple_opt_pass along with...
10879 (pass_data_tail_recursion): ...new pass_data instance and...
10880 (make_pass_tail_recursion): ...new function.
10881 (pass_tail_calls): Convert from a global struct to a subclass of
10882 gimple_opt_pass along with...
10883 (pass_data_tail_calls): ...new pass_data instance and...
10884 (make_pass_tail_calls): ...new function.
10885 * tree-vect-generic.c (pass_lower_vector): Convert from a global
10886 struct to a subclass of gimple_opt_pass along with...
10887 (pass_data_lower_vector): ...new pass_data instance and...
10888 (make_pass_lower_vector): ...new function.
10889 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
10890 gimple_opt_pass along with...
10891 (pass_data_lower_vector_ssa): ...new pass_data instance and...
10892 (make_pass_lower_vector_ssa): ...new function.
10893 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
10894 to a subclass of gimple_opt_pass along with...
10895 (pass_data_slp_vectorize): ...new pass_data instance and...
10896 (make_pass_slp_vectorize): ...new function.
10897 (pass_ipa_increase_alignment): Convert from a global struct to a
10898 subclass of simple_ipa_opt_pass along with...
10899 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
10900 (make_pass_ipa_increase_alignment): ...new function.
10901 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
10902 gimple_opt_pass along with...
10903 (pass_data_vrp): ...new pass_data instance and...
10904 (make_pass_vrp): ...new function.
10905 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
10906 subclass of simple_ipa_opt_pass along with...
10907 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
10908 (make_pass_ipa_free_lang_data): ...new function.
10909 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
10910 gimple_opt_pass along with...
10911 (pass_data_tsan): ...new pass_data instance and...
10912 (make_pass_tsan): ...new function.
10913 (pass_tsan_O0): Convert from a global struct to a subclass of
10914 gimple_opt_pass along with...
10915 (pass_data_tsan_O0): ...new pass_data instance and...
10916 (make_pass_tsan_O0): ...new function.
10917 * var-tracking.c (pass_variable_tracking): Convert from a global
10918 struct to a subclass of rtl_opt_pass along with...
10919 (pass_data_variable_tracking): ...new pass_data instance and...
10920 (make_pass_variable_tracking): ...new function.
10921 * web.c (pass_web): Convert from a global struct to a subclass of
10922 rtl_opt_pass along with...
10923 (pass_data_web): ...new pass_data instance and...
10924 (make_pass_web): ...new function.
10925 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
10926 declaration with that of...
10927 (make_pass_mode_switch_use): ...new function.
10928 (pass_resolve_sw_modes): Replace declaration with that of...
10929 (make_pass_resolve_sw_modes): ...new function.
10930 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
10931 from a global struct to a subclass of rtl_opt_pass along with...
10932 (pass_data_mode_switch_use): ...new pass_data instance and...
10933 (make_pass_mode_switch_use): ...new function.
10934 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
10935 from a global struct to a subclass of rtl_opt_pass along with...
10936 (pass_data_resolve_sw_modes): ...new pass_data instance and...
10937 (make_pass_resolve_sw_modes): ...new function.
10938 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
10939 struct to a subclass of rtl_opt_pass along with...
10940 (pass_data_insert_vzeroupper): ...new pass_data instance and...
10941 (make_pass_insert_vzeroupper): ...new function.
10942 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
10943 global struct to a subclass of rtl_opt_pass along with...
10944 (pass_data_work_around_errata): ...new pass_data instance and...
10945 (make_pass_work_around_errata): ...new function.
10946 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
10947 struct to a subclass of rtl_opt_pass along with...
10948 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
10949 (make_pass_mips_machine_reorg2): ...new function.
10951 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10953 * passes.c (pass_manager::operator new): New.
10955 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10957 Handwritten part of conversion of passes to C++ classes.
10959 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
10960 (toplev.o): Add dep on PASS_MANAGER_H.
10961 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
10962 of early local pases to reflect this moving from a global to a
10963 member of gcc::pass_manager.
10964 (cgraph_add_new_function): Likewise.
10965 * lto-cgraph.c (lto_output_node): Update for conversion of
10966 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
10967 * passes.c (opt_pass::clone): New.
10968 (opt_pass::gate): New.
10969 (opt_pass::execute): New.
10970 (opt_pass::opt_pass): New.
10971 (pass_manager::execute_early_local_passes): New.
10972 (pass_manager::execute_pass_mode_switching): new.
10973 (finish_optimization_passes): Convert to...
10974 (pass_manager::finish_optimization_passes): ...this.
10975 (finish_optimization_passes): Update for conversion of passes to
10977 (register_dump_files_1): Use has_gate since we cannot portably
10978 check a vtable entry against NULL.
10979 (dump_one_pass): Likewise.
10980 (ipa_write_summaries_2): Likewise.
10981 (ipa_write_optimization_summaries_1): Likewise.
10982 (ipa_read_summaries_1): Likewise.
10983 (ipa_read_optimization_summaries_1): Likewise.
10984 (execute_ipa_stmt_fixups): Likewise.
10985 (pass_manager::pass_manager): Rewrite pass-creation, invoking
10986 pass-creation functions rather than wiring up globals, and
10987 storing the results in fields of pass_manager generated using
10988 pass-instances.def.
10989 (pass_manager::dump_profile_report): Update for conversion of
10990 passes to C++ classes.
10991 (pass_manager::execute_ipa_summary_passes): Likewise.
10992 (execute_one_ipa_transform_pass): Likewise.
10993 (execute_one_pass): Use has_gate and has_execute since we cannot
10994 portably check a vtable entry against NULL.
10995 * pass_manager.h (pass_manager::finish_optimization_passes): New.
10996 (pass_manager): Use pass-instances.def to add fields for the
10997 various pass instances.
10998 * toplev.c (finalize): Update for move of
10999 finish_optimization_passes to a method of gcc::pass_manager.
11000 * toplev.h (finish_optimization_passes): Move to method of class
11002 * tree-pass.h (struct pass_data): New.
11003 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
11004 (opt_pass::gate): Convert to virtual function.
11005 (opt_pass::~opt_pass): New.
11006 (opt_pass::clone): New.
11007 (opt_pass::execute): Convert to virtual function.
11008 (opt_pass::opt_pass): New.
11009 (opt_pass::ctxt_): new.
11010 (gimple_opt_pass): Convert to subclass of opt_pass.
11011 (gimple_opt_pass::gimple_opt_pass): New.
11012 (rtl_opt_pass): Convert to subclass of opt_pass.
11013 (rtl_opt_pass::rtl_opt_pass): New.
11014 (ipa_opt_pass_d): Convert to subclass of opt_pass.
11015 (ipa_opt_pass_d::ipa_opt_pass_d): New.
11016 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
11017 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
11018 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
11019 invocation of pass_mode_switching to reflect this moving from a
11020 global to a member of gcc::pass_manager.
11021 (ix86_option_override): Rework how pass_insert_vzeroupper is
11022 added to the pass_manager to reflect autogenerated changes.
11023 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
11025 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
11027 PR rtl-optimization/57708
11028 * recog.c (peep2_find_free_register): Validate all regs in a
11031 2013-08-05 Jan Hubicka <jh@suse.cz>
11034 * cgraph.c (verify_cgraph_node): Accept local flags from other
11036 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
11037 (function_and_variable_visibility): Likewise.
11038 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
11040 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11042 * graph.c (init_graph_slim_pretty_print): Remove.
11043 (print_graph_cfg): Do not call it. Use local pretty printer.
11044 (start_graph_dump): Likewise.
11046 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11048 * gimple-pretty-print.c (buffer): Remove.
11049 (initialized): Likewise.
11050 (maybe_init_pretty_print): Likewise.
11051 (print_gimple_stmt): Do not call it. Use non-static local
11052 pretty_printer variable.
11053 (print_gimple_expr): Likewise.
11054 (print_gimple_seq): Likewise.
11055 (gimple_dump_bb): Likewise.
11057 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11059 * asan.c (asan_pp): Remove.
11060 (asan_pp_initialized): Likewise.
11061 (asan_pp_initialize): Likewise.
11062 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
11063 (asan_emit_stack_protection): Tidy. Use local pretty printer.
11064 (asan_add_global): Likewise.
11066 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
11068 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
11069 * diagnostic.h (diagnostic_flush_buffer): Adjust.
11070 * pretty-print.c (pp_formatted_text_data): Likewise.
11071 (pp_indent): Rename from pp_base_indent.
11072 (pp_format): Rename from pp_base_format.
11073 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
11074 (pp_format_verbatim): Rename from pp_base_format_verbatim.
11075 (pp_flush): Rename from pp_base_flush.
11076 (pp_set_line_maximum_length): Rename from
11077 pp_base_set_line_maximum_length.
11078 (pp_clear_output_area): Rename from pp_base_clear_output_area.
11079 (pp_set_prefix): Rename from pp_base_set_prefix.
11080 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
11081 (pp_emit_prefix): Rename from pp_base_emit_prefix.
11082 (pp_append_text): Rename from pp_base_append_text.
11083 (pp_formatted_text): Rename from pp_base_formatted_text.
11084 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
11085 (pp_remaining_character_count_for_line): Rename from
11086 pp_base_remaining_character_count_for_line.
11087 (pp_newline): Rename from pp_base_newline.
11088 (pp_character): Rename from pp_base_character.
11089 (pp_string): Rename from pp_base_string.
11090 (pp_maybe_space): Rename from pp_base_maybe_space.
11091 * asan.c (asan_pp_string): Adjust.
11092 (asan_emit_stack_protection): Likewise.
11093 (asan_add_global): Likewise.
11094 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
11095 * tree-mudflap.c (mf_varname_tree): Likewise.
11096 * tree-pretty-print.c (pp_tree_identifier): Rename from
11097 pp_base_tree_identifier.
11098 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
11099 Declare as function.
11101 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11103 * pretty-print.h (pp_bar_bar): New.
11104 (pp_ampersand_ampersand): Likewise.
11105 (pp_less_equal): Likewise.
11106 (pp_greater_equal): Likewise.
11107 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
11108 printer functions instead of pp_string or operators and punctuators.
11109 (dump_gimple_call): Likewise.
11110 (dump_gimple_omp_for): Likewise.
11111 (dump_gimple_transaction): Likewise.
11112 (dump_gimple_phi): Likewise.
11113 (pp_gimple_stmt_1): Likewise.
11114 * sched-vis.c (print_insn): Likewise.
11115 * tree-mudflap.c (mf_varname_tree): Likewise.
11116 * tree-pretty-print.c (dump_block_node): Likewise.
11117 (dump_generic_node): Likewise.
11119 2013-08-02 Jan Hubicka <jh@suse.cz>
11121 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
11123 * lto-streamer-out.c (tree_is_indexable): Results decls and
11124 parm decls are not indexable.
11125 (DFS_write_tree_body): Do not follow args and results.
11126 (hash_tree): Likewise.
11127 (output_functions): Rearrange so struct function is needed
11128 only when real body is output; be able to also ouptut abstract
11129 functions; output DECL_ARGUMENTS and DECL_RESULT.
11130 (lto_output): When not in WPA, ale store abstract functions.
11131 (write_symbol): Do not care about RESULT_DECL.
11132 (output_symbol_p): Handle correctly sbtract decls.
11133 * lto-streamer-in.c (input_function): Rearrange so struct
11134 function can be NULL at entry; allow streaming of
11135 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
11136 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
11137 sanity check during LTO.
11138 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
11139 RESULT_DECl and DECL_ARGUMENTS.
11140 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
11143 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11145 * pretty-print.h (pp_underscore): New.
11147 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
11148 printer functions instead of pp_character.
11149 (dump_binary_rhs): Likewise.
11150 (dump_ternary_rhs): Likewise.
11151 (dump_gimple_call_args): Likewise.
11152 (pp_points_to_solution): Likewise.
11153 (dump_gimple_call): Likewise.
11154 (dump_gimple_switch): Likewise.
11155 (dump_gimple_cond): Likewise.
11156 (dump_gimple_bind): Likewise.
11157 (dump_gimple_try): Likewise.
11158 (dump_gimple_omp_for): Likewise.
11159 (dump_gimple_omp_continue): Likewise.
11160 (dump_gimple_omp_single): Likewise.
11161 (dump_gimple_omp_sections): Likewise.
11162 (dump_gimple_omp_block): Likewise.
11163 (dump_gimple_omp_critical): Likewise.
11164 (dump_gimple_transaction): Likewise.
11165 (dump_gimple_asm): Likewise.
11166 (dump_gimple_phi): Likewise.
11167 (dump_gimple_omp_parallel): Likewise.
11168 (dump_gimple_omp_task): Likewise.
11169 (dump_gimple_omp_atomic_load): Likewise.
11170 (dump_gimple_omp_atomic_store): Likewise.
11171 (dump_gimple_mem_ops): Likewise.
11172 (pp_gimple_stmt_1): Likewise.
11173 (pp_cfg_jump): Likewise.
11174 (dump_implicit_edges): Likewise.
11175 (gimple_dump_bb_for_graph): Likewise.
11176 * graph.c (draw_cfg_node): Likewise.
11177 * langhooks.c (lhd_print_error_function): Likewise.
11178 * sched-vis.c (print_exp): Likewise.
11179 (print_value): Likewise.
11180 (print_pattern): Likewise.
11181 (print_insn): Likewise.
11182 (rtl_dump_bb_for_graph): Likewise.
11183 * tree-pretty-print.c (dump_function_declaration): Likewise.
11184 (dump_array_domain): Likewise.
11185 (dump_omp_clause): Likewise.
11186 (dump_location): Likewise.
11187 (dump_generic_node): Likewise.
11188 (print_struct_decl): Likewise.
11189 * diagnostic.c (diagnostic_show_locus): Use pp_space.
11191 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11193 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
11194 candidate table when replacing a candidate statement.
11195 (replace_rhs_if_not_dup): Likewise.
11196 (replace_one_candidate): Likewise.
11198 2013-08-02 Jan Hubicka <jh@suse.cz>
11199 Martin Liska <marxin.liska@gmail.com>
11201 * cgraphunit.c (add_new_function): Fix logic when adding from
11203 (assemble_thunk): Rename to ...
11204 (expand_thunk); .. this one; export; get it working with
11205 general functions; make produced gimple valid.
11206 * cgraph.h (expand_thunk): Declare.
11208 2013-08-02 Jan Hubicka <jh@suse.cz>
11210 * ipa-cp.c (gather_context_independent_values): Use
11211 ipa_get_param_move_cost.
11212 (get_replacement_map): Remove PARAM; move parameter folding
11214 (create_specialized_node): Update.
11215 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
11216 assert that we have gimple body; update move_cost.
11217 (count_formal_params): Assert that we have gimple body.
11218 (ipa_dump_param): New function.
11219 (ipa_alloc_node_params): Break out from ...
11220 (ipa_initialize_node_params): ... here.
11221 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
11222 (ipa_write_node_info): Stream move costs.
11223 (ipa_read_node_info): Read move costs.
11224 (ipa_update_after_lto_read): Do not recompute node params.
11225 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
11226 (ipa_get_param): Check we are not in WPA.
11227 (ipa_get_param_move_cost): New.
11228 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
11229 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
11230 parm numbers to be present.
11232 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11234 PR rtl-optimization/58048
11235 * lra-constraints.c (process_alt_operands): Don't check asm
11236 operand on register.
11238 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
11240 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
11241 the implied StoreLoad barrier for atomic operations if before.
11243 2013-08-02 Jan Hubicka <jh@suse.cz>
11244 Martin Liska <marxin.liska@gmail.com>
11246 * cgraph.c (cgraph_function_body_availability): Do not check
11248 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
11249 symtab_node_availability): Declare.
11250 * ipa.c (can_replace_by_local_alias): New.
11251 (function_and_variable_visibility): Use it.
11252 * symtab.c (symtab_for_node_and_aliases,
11253 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
11255 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11257 PR rtl-optimization/57963
11258 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
11259 (lra_constraints): Use them.
11261 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
11263 * config/arm/types.md (define_attr "type"): Add "load_acq"
11265 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
11267 (cortex_a53_store1): Likewise.
11269 2013-08-01 Jan Hubicka <jh@suse.cz>
11271 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
11272 partitions are not needed.
11274 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
11276 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
11277 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
11278 MAYBE_NON_Q_CLASS_P where appropriate.
11280 2013-08-01 Jan Hubicka <jh@suse.cz>
11282 * cgraph.h (release_function_body): Declare.
11283 * tree.c (free_lang_data_in_decl): Free, parameters and return values
11284 of unused delcarations.
11286 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11288 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
11289 RTL form when subtracting a constant.
11291 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11293 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
11294 Generate canonical plus rtx with negated immediate instead of minus
11296 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
11298 2013-08-01 Jan Hubicka <jh@suse.cz>
11300 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
11301 (cgraph_release_function_body): Likewise.
11302 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
11303 * cgraph.h (cgrpah_node): Rename abstract_and_needed
11304 to used_as_abstract_origin.
11305 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
11306 symbols used as abstract origins.
11307 * cgraphunit.c (analyze_functions): Update.
11308 * ipa.c (symtab_remove_unreachable_nodes): Recompute
11309 used_as_abstract_origin.
11310 * tree-inline.c (tree_function_versioning): Update
11311 used_as_abstract_origin; be ready for DECL_RESULT and
11312 DECL_ARGUMENTS to be NULL.
11314 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
11315 for abstract functions.
11316 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
11319 2013-08-01 Jan Hubicka <jh@suse.cz>
11321 * profile.c (compute_value_histograms): Fix thinko.
11323 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11325 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
11326 aarch-common-protos.h to extra_headers.
11327 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11328 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
11329 * config/aarch64/t-aarch64 (aarch-common.o): Define.
11331 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11333 * config/aarch64/aarch64.md (define_attr "type"): Delete.
11334 Include "../arm/types.md". Define "type" attribute for all patterns.
11335 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
11338 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
11340 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
11341 to support power8 load fusion.
11342 (fusion_gpr_mem_load): Likewise.
11344 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
11346 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
11347 declarations for power8 load fusion.
11348 (emit_fusion_gpr_load): Likewise.
11350 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
11351 tuning for power8, turn on fusion mode by default. Turn on sign
11352 extending fusion mode if normal fusion mode is on, and we are at
11354 (fusion_gpr_load_p): New function, return true if we can fuse an
11355 addis instruction with a dependent load to a GPR.
11356 (emit_fusion_gpr_load): Emit the instructions for power8 load
11359 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
11360 (VSX load fusion peepholes): New peepholes to fuse together an
11361 addi instruction with a VSX load instruction.
11363 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
11364 peepholes to fuse an addis instruction with a load to a GPR base
11365 register. If we are supporting sign extending fusions, convert
11366 sign extending loads to zero extending loads and add an explicit
11369 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11371 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
11372 aarch-common-protos.h to extra_headers.
11373 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11374 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
11375 (arm_early_store_addr_dep): Likewise.
11376 (arm_no_early_alu_shift_dep): Likewise.
11377 (arm_no_early_alu_shift_value_dep): Likewise.
11378 (arm_no_early_mul_dep): Likewise.
11379 (arm_no_early_store_addr_dep): Likewise.
11380 (arm_mac_accumulator_is_mul_result): Likewise.
11381 (arm_mac_accumulator_is_result): Likewise.
11382 * config/arm/aarch-common.c: ... here. New file.
11383 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
11385 (arm_early_store_addr_dep): Likewise.
11386 (arm_no_early_alu_shift_dep): Likewise.
11387 (arm_no_early_alu_shift_value_dep): Likewise.
11388 (arm_no_early_mul_dep): Likewise.
11389 (arm_no_early_store_addr_dep): Likewise.
11390 (arm_mac_accumulator_is_mul_result): Likewise.
11391 (arm_mac_accumulator_is_result): Likewise.
11392 * config/arm/aarch-common-protos.h: ... here. New file.
11393 * config/arm/t-arm (aarch-common.o): Define.
11395 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11397 * config/arm/arm.md: Include new file "types.md".
11398 (define_attr "type"): Move from here to ...
11399 (define_attr "mul32"): Likewise.
11400 (define_attr "mul64"): Likewise.
11401 * config/arm/types.md: ... here. New file.
11403 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
11405 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
11406 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
11408 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11410 * gen-pass-instances.awk: Fix offset of substr().
11412 2013-07-31 David Malcolm <dmalcolm@redhat.com>
11414 * Makefile.in (pass-instances.def): New.
11415 (passes.o): Replace dependency on passes.def with one on
11418 * gen-pass-instances.awk: New.
11420 * passes.c (pass_manager::pass_manager): Use pass-instances.def
11421 rather than passes.def, updating local definition of NEXT_PASS
11422 macro to add an extra NUM parameter (currently unused).
11424 2013-07-30 David Malcolm <dmalcolm@redhat.com>
11426 * Makefile.in (PASS_MANAGER_H): New.
11427 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
11428 (passes.o): Likewise.
11429 (statistics.o): Likewise.
11430 (cgraphunit.o): Likewise.
11431 (context.o): Depend on PASS_MANAGER_H.
11433 * pass_manager.h: New.
11435 * cgraphunit.c (cgraph_add_new_function): Update for moves
11436 of globals to fields of pass_manager.
11437 (analyze_function): Likewise.
11438 (expand_function): Likewise.
11439 (ipa_passes): Likewise.
11440 (compile): Likewise.
11442 * context.c (context::context): New.
11443 * context.h (context::context): New.
11444 (context::get_passes): New.
11445 (context::passes_): New.
11447 * lto-cgraph.c (input_node): Update for moves of globals to
11448 fields of pass_manager.
11450 * passes.c (all_passes): Remove, in favor of a field of the
11451 same name within the new class pass_manager.
11452 (all_small_ipa_passes): Likewise.
11453 (all_lowering_passes): Likewise.
11454 (all_regular_ipa_passes): Likewise.
11455 (all_late_ipa_passes): Likewise.
11456 (all_lto_gen_passes): Likewise.
11457 (passes_by_id): Likewise.
11458 (passes_by_id_size): Likewise.
11459 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11460 the new class pass_manager.
11461 (set_pass_for_id): Convert to...
11462 (pass_manager::set_pass_for_id): ...method.
11463 (get_pass_for_id): Convert to...
11464 (pass_manager::get_pass_for_id): ...method.
11465 (register_one_dump_file): Move body of implementation into...
11466 (pass_manager::register_one_dump_file): ...here.
11467 (register_dump_files_1): Convert to...
11468 (pass_manager::register_dump_files_1): ...method.
11469 (register_dump_files): Convert to...
11470 (pass_manager::register_dump_files): ...method.
11471 (create_pass_tab): Update for moves of globals to fields of
11473 (dump_passes): Move body of implementation into...
11474 (pass_manager::dump_passes): ...here.
11475 (register_pass): Move body of implementation into...
11476 (pass_manager::register_pass): ...here.
11477 (init_optimization_passes): Convert into...
11478 (pass_manager::pass_manager): ...constructor for new
11479 pass_manager class, and initialize the pass_lists array.
11480 (check_profile_consistency): Update for moves of globals to
11481 fields of pass_manager.
11482 (dump_profile_report): Move body of implementation into...
11483 (pass_manager::dump_profile_report): ...here.
11484 (ipa_write_summaries_1): Update for moves of pass lists from
11485 being globals to fields of pass_manager.
11486 (ipa_write_optimization_summaries): Likewise.
11487 (ipa_read_summaries): Likewise.
11488 (ipa_read_optimization_summaries): Likewise.
11489 (execute_all_ipa_stmt_fixups): Likewise.
11491 * statistics.c (statistics_fini): Update for moves of globals to
11492 fields of pass_manager.
11494 * toplev.c (general_init): Replace call to
11495 init_optimization_passes with construction of the pass_manager
11498 * tree-pass.h (all_passes): Remove, in favor of a field of the
11499 same name within the new class pass_manager.
11500 (all_small_ipa_passes): Likewise.
11501 (all_lowering_passes): Likewise.
11502 (all_regular_ipa_passes): Likewise.
11503 (all_lto_gen_passes): Likewise.
11504 (all_late_ipa_passes): Likewise.
11505 (passes_by_id): Likewise.
11506 (passes_by_id_size): Likewise.
11507 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11508 the new class pass_manager.
11509 (get_pass_for_id): Remove.
11511 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11513 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
11516 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11518 * arm.md (mulhi3): New expand pattern.
11520 2013-07-30 Jan Hubicka <jh@suse.cz>
11521 Martin Liska <marxin.liska@gmail.com>
11523 * profile.c (compute_value_histograms): Do not ICE when
11524 there is mismatch only on some counters.
11526 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11528 PR rtl-optimization/57637
11529 * function.c (move_insn_for_shrink_wrap): Also check the
11530 GEN set of the LIVE problem for the liveness analysis
11531 if it exists, otherwise give up.
11533 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11535 PR tree-optimization/57993
11536 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
11537 replaced statement in the candidate table.
11538 (phi_add_costs): Return infinite cost when the hidden basis does
11539 not dominate all phis on which the candidate is dependent.
11540 (replace_one_candidate): Record replaced statement in the
11543 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
11545 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
11546 (ashlv2si3): New expander.
11547 (*ashlv2si3_i): New define_insn_and_split.
11548 * predicates.md (float_operation): Allow patterns with three
11549 basic sub-patterns.
11551 PR rtl-optimization/58021
11552 * mode-switching.c (create_pre_exit): Always split off preceding
11553 insns if we are not at the basic block head.
11555 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
11557 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
11558 (UCLIBC_DYNAMIC_LINKER): New macro.
11559 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
11561 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
11562 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
11564 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
11565 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
11566 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
11567 for SF and DF modes. Use ieee_quad_format for TF mode.
11568 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
11569 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
11570 (mips_option_override): Handle `-mnan=legacy'.
11571 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
11572 `-mabs=2008' and `-mnan=2008'.
11573 (OPTION_DEFAULT_SPECS): Add "nan" default.
11574 (ASM_SPEC): Handle `-mnan='.
11575 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
11576 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
11577 comment accordingly.
11578 (neg<mode>2): Likewise.
11579 * config/mips/mips.opt (mabs, mnan): New options.
11580 * doc/install.texi (Configuration): Document `--with-nan=' option.
11581 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
11583 (MIPS Options): Document them.
11584 * config.gcc <mips*-*-*>: Handle `--with-nan='.
11585 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
11586 * configure: Regenerate.
11587 * config.in: Regenerate.
11589 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11591 * config/i386/i386.md (float post-reload splitters): Do not check
11592 for subregs of SSE registers.
11594 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11595 H.J. Lu <hongjiu.lu@intel.com>
11599 * config/i386/i386.md (post-reload splitter
11600 to avoid partial SSE reg dependency stalls): New pattern.
11602 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
11604 * config/s390/s390.md ("movcc"): Swap load and store instructions.
11606 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
11608 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11609 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
11611 2013-07-26 Cary Coutant <ccoutant@google.com>
11613 * dwarf2out.c (die_checksum_ordered): Don't include template
11614 instantiations in signature.
11615 (is_template_parameter): New function.
11616 (is_template_instantiation): New function.
11617 (generate_skeleton_bottom_up): Don't include template instantiations
11619 (generate_skeleton): Likewise.
11620 (break_out_comdat_types): Move recursive call to break out nested
11622 (prune_unused_types_mark_generic_parms_dies): Call
11623 is_template_parameter.
11625 2013-07-26 Ian Bolton <ian.bolton@arm.com>
11627 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
11628 uses vector registers.
11629 * config/aarch64/iterators.md: Add attributes rtn and vas.
11631 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11632 Richard Earnshaw <richard.earnshaw@arm.com>
11634 * combine.c (simplify_comparison): Re-canonicalize operands
11636 * config/arm/arm.md (movcond_addsi): New splitter.
11638 2013-07-25 Sterling Augustine <saugustine@google.com>
11640 * dwarf2out.c (size_of_pubnames): Move code to...
11641 (include_pubname_in_output): ...here. New.
11642 (want_pubnames): Rearrange.
11643 (output_pubnames): Call include_pubname_in_output. Move assertion.
11645 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11647 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
11649 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11652 * doc/extend.texi: Remove obsolete builtins. Fix
11653 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
11655 2013-07-25 Jan Hubicka <jh@suse.cz>
11657 * cgraph.c (release_function_body): Break out from ...
11658 (cgraph_release_function_body): ... this one; also release DECL_RESULT
11659 and DECL_ARGUMENTS.
11660 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
11661 old_tree in the map.
11662 (create_specialized_node): Update.
11663 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
11665 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
11666 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
11667 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
11668 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
11671 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11673 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
11674 addsi3_carryin_alt2_<optab>): Correct output template.
11676 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11678 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
11679 Adjust for arm_restrict_it.
11680 Remove trailing whitespace.
11682 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
11684 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
11685 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
11687 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
11689 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
11691 PR rtl-optimization/57960
11692 * lra-constraints.c (process_alt_operands): Use the right mode
11693 when checking strict_low.
11695 2013-07-25 Jan Hubicka <jh@suse.cz>
11697 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
11698 * cgraph.c (cgraph_remove_node): Do not release function body
11699 when in cgraph streaming.
11700 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
11701 in other partitions are not considered reachable; fix handling of
11704 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11706 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
11708 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11713 * config/arm/arm.md ("*sibcall_insn): Replace use of
11714 Ss with US. Adjust output for v5 and v4t.
11715 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
11717 * config/arm/constraints.md ("Ss"): Rename to US.
11719 2013-07-25 Terry Guo <terry.guo@arm.com>
11721 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
11722 shift_add/shift_sub0/shift_sub1 RTXs.
11724 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
11725 Anton Blanchard <anton@au1.ibm.com>
11727 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
11728 (altivec_vpks<VI_char>ss): Likewise.
11729 (altivec_vpks<VI_char>us): Likewise.
11730 (altivec_vpku<VI_char>us): Likewise.
11731 (altivec_vpku<VI_char>um): Likewise.
11733 2013-07-24 David Malcolm <dmalcolm@redhat.com>
11735 Introduce context class.
11737 * Makefile.in (CONTEXT_H): New.
11738 (OBJS): Add context.o.
11739 (toplev.o): Add CONTEXT_H to dependencies.
11742 * toplev.c (general_init): Create the singleton gcc::context instance.
11748 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
11750 PR rtl-optimization/57968
11751 * mode-switching.c (create_pre_exit): Allow instructions that
11752 don't set a return register to need a non-exit mode.
11754 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11755 Anton Blanchard <anton@au1.ibm.com>
11757 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
11758 operands to vperm for little endian.
11759 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
11760 of lvsl to create the control mask for a vperm for little endian.
11762 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11763 Anton Blanchard <anton@au1.ibm.com>
11765 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
11766 two operands for little-endian.
11768 2013-07-23 Steve Ellcey <sellcey@mips.com>
11770 * config/mips/mips.c (mips_case_values_threshold): New.
11771 (TARGET_CASE_VALUES_THRESHOLD): Define.
11773 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11774 Anton Blanchard <anton@au1.ibm.com>
11776 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
11777 selection of field for vector splat in little endian mode.
11779 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
11781 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
11782 expanders to rs6000.md.
11783 (ior<mode>3): Likewise.
11784 (and<mode>3): Likewise.
11785 (one_cmpl<mode>2): Likewise.
11786 (nor<mode>3): Likewise.
11787 (andc<mode>3): Likewise.
11788 (eqv<mode>3): Likewise.
11789 (nand<mode>3): Likewise.
11790 (orc<mode>3): Likewise.
11792 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
11795 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
11796 to split multi-word logical operations.
11797 (rs6000_split_logical_di): Likewise.
11798 (rs6000_split_logical): Likewise.
11800 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
11801 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
11802 and allow TImode operations in 32-bit.
11803 (vsx_and<mode>3_64bit): Likewise.
11804 (vsx_ior<mode>3_32bit): Likewise.
11805 (vsx_ior<mode>3_64bit): Likewise.
11806 (vsx_xor<mode>3_32bit): Likewise.
11807 (vsx_xor<mode>3_64bit): Likewise.
11808 (vsx_one_cmpl<mode>2_32bit): Likewise.
11809 (vsx_one_cmpl<mode>2_64bit): Likewise.
11810 (vsx_nor<mode>3_32bit): Likewise.
11811 (vsx_nor<mode>3_64bit): Likewise.
11812 (vsx_andc<mode>3_32bit): Likewise.
11813 (vsx_andc<mode>3_64bit): Likewise.
11814 (vsx_eqv<mode>3_32bit): Likewise.
11815 (vsx_eqv<mode>3_64bit): Likewise.
11816 (vsx_nand<mode>3_32bit): Likewise.
11817 (vsx_nand<mode>3_64bit): Likewise.
11818 (vsx_orc<mode>3_32bit): Likewise.
11819 (vsx_orc<mode>3_64bit): Likewise.
11821 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
11822 logical types in GPRs.
11824 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
11825 logical insns to rs6000.md, and allow TImode operations in
11827 (altivec_ior<mode>3): Likewise.
11828 (altivec_xor<mode>3): Likewise.
11829 (altivec_one_cmpl<mode>2): Likewise.
11830 (altivec_nor<mode>3): Likewise.
11831 (altivec_andc<mode>3): Likewise.
11833 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
11834 attributes for moving the 128-bit logical operations into
11836 (BOOL_REGS_OUTPUT): Likewise.
11837 (BOOL_REGS_OP1): Likewise.
11838 (BOOL_REGS_OP2): Likewise.
11839 (BOOL_REGS_UNARY): Likewise.
11840 (BOOL_REGS_AND_CR0): Likewise.
11841 (one_cmpl<mode>2): Add support for DI logical operations on
11842 32-bit, splitting the operations to 32-bit.
11843 (anddi3): Likewise.
11844 (iordi3): Likewise.
11845 (xordi3): Likewise.
11846 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
11847 changes to combine the 32/64-bit code, allow logical operations on
11848 TI mode in 32-bit, and to use similar match_operator patterns like
11849 scalar mode uses. Combine the Altivec and VSX code for logical
11850 operations, and move it here.
11851 (ior<mode>3, 128-bit types): Likewise.
11852 (xor<mode>3, 128-bit types): Likewise.
11853 (one_cmpl<mode>3, 128-bit types): Likewise.
11854 (nor<mode>3, 128-bit types): Likewise.
11855 (andc<mode>3, 128-bit types): Likewise.
11856 (eqv<mode>3, 128-bit types): Likewise.
11857 (nand<mode>3, 128-bit types): Likewise.
11858 (orc<mode>3, 128-bit types): Likewise.
11859 (and<mode>3_internal): Likewise.
11860 (bool<mode>3_internal): Likewise.
11861 (boolc<mode>3_internal1): Likewise.
11862 (boolc<mode>3_internal2): Likewise.
11863 (boolcc<mode>3_internal1): Likewise.
11864 (boolcc<mode>3_internal2): Likewise.
11865 (eqv<mode>3_internal1): Likewise.
11866 (eqv<mode>3_internal2): Likewise.
11867 (one_cmpl1<mode>3_internal): Likewise.
11869 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
11871 * config/microblaze/microblaze.c (microblaze_expand_prologue):
11872 Rename flag_stack_usage to flag_stack_usage_info.
11874 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
11876 * config/microblaze/sync.md: New file.
11877 * config/microblaze/microblaze.md: Include sync.md
11878 * config/microblaze/microblaze.c: Add print_operand 'y'.
11879 * config/microblaze/constraints.md: Add memory_contraint
11880 'Q' which is a single register.
11882 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
11884 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
11886 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
11888 * reload.c (find_reloads): Exit loop once we find this operand
11889 cannot be reloaded somehow for this alternative.
11891 * reload.c (find_reloads): Exit loop once we find a hard register.
11893 * rtlanal.c (computed_jump_p): Exit loop once we find label
11896 * i386.c (ix86_pad_returns): Exit loop after setting replace.
11898 * cfgloopmanip.c (remove_path): Exit loop after setting
11901 * gensupport.c (subst_dup): Avoid loop if code is not
11902 MATCH_DUP nor MATCH_OP_DUP.
11904 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
11906 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
11908 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11910 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
11911 true for SP_REGNUM if mode == ptr_mode.
11912 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
11913 with value R0_REGNUM + 31.
11915 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11917 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
11918 pad pointer-typed argument downward.
11920 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11922 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
11923 and __ILP32__ when the ILP32 model is in use.
11925 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11927 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
11928 (aarch64_load_symref_appropriately): In the case of
11929 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
11930 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
11931 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
11932 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
11933 if 'mode' doesn't equal to 'ptr_mode'.
11934 (aarch64_output_mi_thunk): Add an assertion on the alignment of
11935 'vcall_offset'; change to call aarch64_emit_move differently depending
11936 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
11937 to calculate the upper bound of 'vcall_offset'.
11938 (aarch64_cannot_force_const_mem): Change to also return true if
11940 (aarch64_legitimize_reload_address): In the case of large
11941 displacements, add new local variable 'xmode' and an assertion
11942 based on it; change to use 'xmode' to generate the new rtx and
11944 (aarch64_asm_trampoline_template): Change to generate the template
11945 differently depending on TARGET_ILP32 or not; change to use
11946 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
11947 (aarch64_trampoline_size): Removed.
11948 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
11949 and replace immediate literals with it. Change to use 'ptr_mode'
11950 instead of 'DImode' and call convert_memory_address if the mode
11951 of 'fnaddr' doesn't equal to 'ptr_mode'.
11952 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
11954 (aarch64_elf_asm_destructor): Likewise.
11955 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
11956 on TARGET_ILP32 instead of aarch64_trampoline_size.
11957 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
11958 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
11959 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
11960 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
11961 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
11962 (storewb_pair<GPI:mode>_<P:mode>): ... this.
11963 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
11964 depending on the value of 'mode'.
11965 (add_losym_<mode>): New.
11966 (ldr_got_small_<mode>): New, based on ldr_got_small.
11967 (ldr_got_small): Remove.
11968 (ldr_got_small_sidi): New.
11969 * config/aarch64/iterators.md (P): New.
11970 (PTR): Change to 'ptr_mode' in the condition.
11972 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11974 * config.gcc (aarch64*-*-*): Support --with-abi.
11975 (aarch64*-*-elf): Support --with-multilib-list.
11976 (aarch64*-*-linux*): Likewise.
11977 (supported_defaults): Add abi to aarch64*-*-*.
11978 * configure.ac: Mention AArch64 for --with-multilib-list.
11979 * configure: Re-generated.
11980 * config/aarch64/biarchilp32.h: New file.
11981 * config/aarch64/biarchlp64.h: New file.
11982 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
11984 (MULTILIB_DEFAULTS): Ditto.
11985 (DRIVER_SELF_SPECS): Ditto.
11986 (ASM_SPEC): Update to also substitute -mabi.
11987 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
11988 file whose name depends on -mabi= and -mbig-endian.
11989 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
11991 (POINTER_SIZE): New define.
11992 (POINTERS_EXTEND_UNSIGNED): Ditto.
11993 (enum aarch64_abi_type): New enumeration tag.
11994 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
11995 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
11996 (TARGET_ILP32): New define.
11997 * config/aarch64/aarch64.opt (mabi): New.
11998 (aarch64_abi): New.
11999 (ilp32, lp64): New values for -mabi.
12000 * config/aarch64/t-aarch64 (comma): New define.
12001 (MULTILIB_OPTIONS): Ditto.
12002 (MULTILIB_DIRNAMES): Ditto.
12003 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
12004 * doc/invoke.texi: Document -mabi for AArch64.
12006 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
12008 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
12010 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12011 Anton Blanchard <anton@au1.ibm.com>
12013 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
12014 endianness when selecting field to splat.
12016 2013-07-22 Eric Christopher <echristo@gmail.com>
12018 * dwarf2out.c (die_odr_checksum): New function to use
12019 CHECKSUM_ macros and ULEB128 for DIE tag.
12020 (generate_type_signature): Use.
12022 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
12024 * config.gcc (sparc*-*-*): Accept leon3 processor.
12025 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
12026 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
12027 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
12028 * config/sparc/sparc.opt (enum processor_type): Add leon3.
12029 (mfix-ut699): Adjust comment.
12030 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
12031 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
12032 (CPP_CPU_SPEC): Likewise.
12033 (ASM_CPU_SPEC): Likewise.
12034 * config/sparc/sparc.c (leon3_cost): New constant.
12035 (sparc_option_override): Add leon3 support.
12036 (mem_ref): New function.
12037 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
12038 (sparc_do_work_around_errata): Look into the instruction in the delay
12039 slot and adjust accordingly. Add fix for the data cache nullify issues
12040 of the UT699. Change insertion position for the NOP.
12041 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
12042 (leon3_load): New reservation.
12043 (leon_store): Bump latency to 2.
12044 (grfpu): New automaton.
12045 (grfpu_alu): New unit.
12046 (grfpu_ds): Likewise.
12047 (leon_fp_alu): Adjust.
12048 (leon_fp_mult): Delete.
12049 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
12050 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
12051 * config/sparc/sparc.md (cpu): Add leon3.
12052 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
12053 (swapsi): Likewise.
12054 (atomic_test_and_set): Likewise.
12055 (ldstub): Likewise.
12057 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
12059 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
12060 default for R5900 targets.
12061 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
12062 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
12063 * config/mips/mips.c (mips_option_override): Report an error for
12064 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
12065 for -march=r5900 -mhard-float.
12067 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
12069 * df-problems.c (can_move_insns_across): Exit loop once we
12070 find a non-fixed, non-global register.
12072 * ipa-pure-const.c (propagate_nothrow): Exit loop after
12075 * omega.c (omega_eliminate_red): Break after setting red_found.
12076 (omega_problem_has_red_equations): Similarly after setting found.
12077 (omega_query_variable): Similarly after setting coupled.
12079 2013-07-22 Marek Polacek <polacek@redhat.com>
12081 * gimplify.c: Don't include gimple.h twice.
12083 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12085 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
12086 instead of TARGET_THUMB1.
12087 (Pz): New constraint.
12088 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
12090 (compare_negsi_si): Likewise.
12091 (compare_addsi2_op0): Likewise.
12092 (compare_addsi2_op1): Likewise.
12093 (addsi3_carryin_<optab>): Likewise.
12094 (addsi3_carryin_alt2_<optab>): Likewise.
12095 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
12096 for arm_restrict_it.
12097 (subsi3_carryin): Likewise.
12098 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
12099 (minmax_arithsi): Disable for arm_restrict_it.
12100 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
12101 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
12102 (satsi_<SAT:code>_shift): Likewise.
12103 (arm_shiftsi3): Add alternative for 16-bit encoding.
12104 (arm32_movhf): Disable for arm_restrict_it.
12105 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
12106 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
12108 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
12110 * config/arm/arm.md (attribute "insn"): Delete.
12111 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
12112 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
12113 (not_shiftsi): Update for attribute change.
12114 (not_shiftsi_compare0): Likewise.
12115 (not_shiftsi_compare0_scratch): Likewise.
12116 (arm_one_cmplsi2): Likewise.
12117 (thumb1_one_cmplsi2): Likewise.
12118 (notsi_compare0): Likewise.
12119 (notsi_compare0_scratch): Likewise.
12120 (thumb1_movdi_insn): Likewise.
12121 (arm_movsi_insn): Likewise.
12122 (movhi_insn_arch4): Likewise.
12123 (movhi_bytes): Likewise.
12124 (arm_movqi_insn): Likewise.
12125 (thumb1_movqi_insn): Likewise.
12126 (arm32_movhf): Likewise.
12127 (thumb1_movhf): Likewise.
12128 (arm_movsf_soft_insn): Likewise.
12129 (thumb1_movsf_insn): Likewise.
12130 (thumb_movdf_insn): Likewise.
12131 (movsicc_insn): Likewise.
12132 (movsfcc_soft_insn): Likewise.
12133 (and_scc): Likewise.
12134 (cond_move): Likewise.
12135 (if_move_not): Likewise.
12136 (if_not_move): Likewise.
12137 (if_shift_move): Likewise.
12138 (if_move_shift): Likewise.
12139 (if_shift_shift): Likewise.
12140 (if_not_arith): Likewise.
12141 (if_arith_not): Likewise.
12142 (cond_move_not): Likewise.
12143 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
12144 (neon_mov<mode>): Likewise.
12145 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
12146 (thumb2_movsi_vfp): Likewise.
12147 (movsf_vfp): Likewise.
12148 (thumb2_movsf_vfp): Likewise.
12149 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12151 (cortexa7_older_only): Likewise.
12152 (cortexa7_younger): Likewise.
12153 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
12154 (1020alu_shift_op): Likewise.
12155 (1020alu_shift_reg_op): Likewise.
12156 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
12157 (alu_shift_op): Likewise.
12158 (alu_shift_reg_op): Likewise.
12159 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
12160 (11_alu_shift_op): Likewise.
12161 (11_alu_shift_reg_op): Likewise.
12162 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
12163 (9_alu_shift_reg_op): Likewise.
12164 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12166 (cortex_a15_alu_shift): Likewise.
12167 (cortex_a15_alu_shift_reg): Likewise.
12168 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
12169 (cortex_a5_alu_shift): Likewise.
12170 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
12172 (cortex_a53_alu_shift): Likewise.
12173 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12175 (cortex_a7_alu_reg): Likewise.
12176 (cortex_a7_alu_shift): Likewise.
12177 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12178 (cortex_a8_alu_shift): Likewise.
12179 (cortex_a8_alu_shift_reg): Likewise.
12180 (cortex_a8_mov): Likewise.
12181 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
12182 (cortex_a9_dp_shift): Likewise.
12183 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
12184 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
12185 (cortex_r4_mov): Likewise.
12186 (cortex_r4_alu_shift): Likewise.
12187 (cortex_r4_alu_shift_reg): Likewise.
12188 * config/arm/fa526.md (526_alu_op): Update for attribute change.
12189 (526_alu_shift_op): Likewise.
12190 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
12191 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
12192 (626te_alu_shift_op): Likewise.
12193 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
12194 (726te_alu_op): Likewise.
12195 (726te_alu_shift_op): Likewise.
12196 (726te_alu_shift_reg_op): Likewise.
12197 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
12198 (mp626_alu_shift_op): Likewise.
12199 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
12200 (pj4_alu_e1_conds): Likewise.
12201 (pj4_alu): Likewise.
12202 (pj4_alu_conds): Likewise.
12203 (pj4_shift): Likewise.
12204 (pj4_shift_conds): Likewise.
12205 (pj4_alu_shift): Likewise.
12206 (pj4_alu_shift_conds): Likewise.
12208 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12210 * config/arm/predicates.md (shiftable_operator_strict_it):
12212 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
12213 Disable cond_exec version for arm_restrict_it.
12214 (thumb2_smaxsi3): Convert to generate cond_exec.
12215 (thumb2_sminsi3): Likewise.
12216 (thumb32_umaxsi3): Likewise.
12217 (thumb2_uminsi3): Likewise.
12218 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
12219 (thumb2_neg_abssi2): Likewise.
12220 (thumb2_mov_scc): Add alternative for 16-bit encoding.
12221 (thumb2_movsicc_insn): Adjust alternatives.
12222 (thumb2_mov_negscc): Disable for arm_restrict_it.
12223 (thumb2_mov_negscc_strict_it): New pattern.
12224 (thumb2_mov_notscc_strict_it): New pattern.
12225 (thumb2_mov_notscc): Disable for arm_restrict_it.
12226 (thumb2_ior_scc): Likewise.
12227 (thumb2_ior_scc_strict_it): New pattern.
12228 (thumb2_cond_move): Adjust for arm_restrict_it.
12229 (thumb2_cond_arith): Disable for arm_restrict_it.
12230 (thumb2_cond_arith_strict_it): New pattern.
12231 (thumb2_cond_sub): Adjust for arm_restrict_it.
12232 (thumb2_movcond): Likewise.
12233 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
12234 (thumb2_zero_extendhisi2_v6): Likewise.
12235 (thumb2_zero_extendqisi2_v6): Likewise.
12236 (orsi_notsi_si): Likewise.
12237 (orsi_not_shiftsi_si): Likewise.
12239 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
12241 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
12242 instruction sequence is 1 byte shorter.
12244 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
12246 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
12247 it is not needed after split.
12249 2013-07-20 Iain Sandoe <iain@codesourcery.com>
12252 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
12253 second label for nonlocal goto receivers. Don't output pic base labels
12254 unless we're producing PIC; mark that action unreachable().
12255 (ix86_save_reg): If the function contains a nonlocal label, save the
12257 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
12258 * config/darwin.c (emitted_pic_label_num): New GTY.
12259 (update_pic_label_number_if_needed): New.
12260 (machopic_output_function_base_name): Adjust for nonlocal receiver
12262 (machopic_should_output_picbase_label): New.
12263 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
12264 (nonlocal_goto_receiver): New insn and split.
12266 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
12268 * config/aarch64/aarch64-builtins.c
12269 (aarch64_fold_builtin): Fold abs in all modes.
12270 * config/aarch64/aarch64-simd-builtins.def
12271 (abs): Enable for all modes.
12272 * config/aarch64/arm_neon.h
12273 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
12274 (vabs_f64): Add missing intrinsic.
12276 2013-07-19 Ian Bolton <ian.bolton@arm.com>
12278 * config/aarch64/arm_neon.h (vabs_s64): New function
12280 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
12283 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
12284 * config/avr/avr.md (adjust_len): Add `round'.
12285 * config/avr/avr-protos.h (avr_out_round): New prototype.
12286 (avr_out_plus): Add `out_label' argument.
12287 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
12288 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
12289 Handle the case where `insn' is just a pattern.
12290 (avr_out_bitop): Handle the case where `insn' is just a pattern.
12291 (avr_out_round): New function.
12292 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
12294 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
12296 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12297 Add check for flag_stack_usage to handle -fstack-usage support
12299 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
12301 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
12302 interaction for new Power8 flags and VSX.
12304 2013-07-18 Sriraman Tallam <tmsriram@google.com>
12306 PR middle-end/57698
12307 * tree-inline.c (expand_call_inline): Emit errors during
12308 early_inlining only if optimization is not turned on.
12310 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12314 * passes.c (init_optimization_passes): Move the construction of
12315 the pass hierarchy into a new passes.def file.
12317 * Makefile.in (passes.o): Add dependency on passes.def.
12319 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12321 * passes.c (init_optimization_passes): Introduce macros for
12322 constructing the tree of passes (INSERT_PASSES_AFTER,
12323 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
12324 TERMINATE_PASS_LIST).
12326 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
12327 Wei Mi <wmi@google.com>
12329 PR rtl-optimization/57878
12330 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
12332 (reload_pseudo_compare_func): Check nregs first for reload
12335 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12337 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
12339 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
12341 * read-rtl.c (validate_const_int): Once an invalid character is
12342 seen, quit the loop.
12344 * gengtype.c (write_roots): Similarly once we find the "deletable"
12345 or "if_marked" option.
12347 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12349 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
12350 "xtab" and "sat". Move value "clz" from here to ...
12351 (attriubte "type"): ... here.
12352 (satsi_<SAT:code>): Delete "insn" attribute.
12353 (satsi_<SAT:code>_shift): Likewise.
12354 (arm_zero_extendqisi2addsi): Likewise.
12355 (arm_extendqisi2addsi): Likewise.
12356 (clzsi2): Update for attribute changes.
12357 (rbitsi2): Likewise.
12358 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
12360 (arm_usatsihi): Likewise.
12361 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12363 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12365 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
12366 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
12367 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
12368 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
12369 in alphabetical order.
12370 (attribute "core_cycles"): Update for attribute changes.
12371 (arm_addsi3): Likewise.
12372 (addsi3_compare0): Likewise.
12373 (addsi3_compare0_scratch): Likewise.
12374 (addsi3_compare_op1): Likewise.
12375 (addsi3_compare_op2): Likewise.
12376 (compare_addsi2_op0): Likewise.
12377 (compare_addsi2_op1): Likewise.
12378 (addsi3_carryin_shift_<optab>): Likewise.
12379 (subsi3_carryin_shift): Likewise.
12380 (rsbsi3_carryin_shift): Likewise.
12381 (arm_subsi3_insn): Likewise.
12382 (subsi3_compare0): Likewise.
12383 (subsi3_compare): Likewise.
12384 (arm_andsi3_insn): Likewise.
12385 (thumb1_andsi3_insn): Likewise.
12386 (andsi3_compare0): Likewise.
12387 (andsi3_compare0_scratch): Likewise.
12388 (zeroextractsi_compare0_scratch
12389 (andsi_not_shiftsi_si): Likewise.
12390 (iorsi3_insn): Likewise.
12391 (iorsi3_compare0): Likewise.
12392 (iorsi3_compare0_scratch): Likewise.
12393 (arm_xorsi3): Likewise.
12394 (thumb1_xorsi3_insn): Likewise.
12395 (xorsi3_compare0): Likewise.
12396 (xorsi3_compare0_scratch): Likewise.
12397 (satsi_<SAT:code>_shift): Likewise.
12399 (arm_shiftsi3): Likewise.
12400 (shiftsi3_compare0): Likewise.
12401 (not_shiftsi): Likewise.
12402 (not_shiftsi_compare0): Likewise.
12403 (not_shiftsi_compare0_scratch): Likewise.
12404 (arm_one_cmplsi2): Likewise.
12405 (thumb_one_complsi2): Likewise.
12406 (notsi_compare0): Likewise.
12407 (notsi_compare0_scratch): Likewise.
12408 (thumb1_zero_extendhisi2): Likewise.
12409 (arm_zero_extendhisi2): Likewise.
12410 (arm_zero_extendhisi2_v6): Likewise.
12411 (arm_zero_extendhisi2addsi): Likewise.
12412 (thumb1_zero_extendqisi2): Likewise.
12413 (thumb1_zero_extendqisi2_v6): Likewise.
12414 (arm_zero_extendqisi2): Likewise.
12415 (arm_zero_extendqisi2_v6): Likewise.
12416 (arm_zero_extendqisi2addsi): Likewise.
12417 (thumb1_extendhisi2): Likewise.
12418 (arm_extendhisi2): Likewise.
12419 (arm_extendhisi2_v6): Likewise.
12420 (arm_extendqisi): Likewise.
12421 (arm_extendqisi_v6): Likewise.
12422 (arm_extendqisi2addsi): Likewise.
12423 (thumb1_extendqisi2): Likewise.
12424 (thumb1_movdi_insn): Likewise.
12425 (arm_movsi_insn): Likewise.
12426 (movsi_compare0): Likewise.
12427 (movhi_insn_arch4): Likewise.
12428 (movhi_bytes): Likewise.
12429 (arm_movqi_insn): Likewise.
12430 (thumb1_movqi_insn): Likewise.
12431 (arm32_movhf): Likewise.
12432 (thumb1_movhf): Likewise.
12433 (arm_movsf_soft_insn): Likewise.
12434 (thumb1_movsf_insn): Likewise.
12435 (movdf_soft_insn): Likewise.
12436 (thumb_movdf_insn): Likewise.
12437 (arm_cmpsi_insn): Likewise.
12438 (cmpsi_shiftsi): Likewise.
12439 (cmpsi_shiftsi_swp): Likewise.
12440 (arm_cmpsi_negshiftsi_si): Likewise.
12441 (movsicc_insn): Likewise.
12442 (movsfcc_soft_insn): Likewise.
12443 (arith_shiftsi): Likewise.
12444 (arith_shiftsi_compare0
12445 (arith_shiftsi_compare0_scratch
12446 (sub_shiftsi): Likewise.
12447 (sub_shiftsi_compare0
12448 (sub_shiftsi_compare0_scratch
12449 (and_scc): Likewise.
12450 (cond_move): Likewise.
12451 (if_plus_move): Likewise.
12452 (if_move_plus): Likewise.
12453 (if_move_not): Likewise.
12454 (if_not_move): Likewise.
12455 (if_shift_move): Likewise.
12456 (if_move_shift): Likewise.
12457 (if_shift_shift): Likewise.
12458 (if_not_arith): Likewise.
12459 (if_arith_not): Likewise.
12460 (cond_move_not): Likewise.
12461 (thumb1_ashlsi3): Set type attribute.
12462 (thumb1_ashrsi3): Likewise.
12463 (thumb1_lshrsi3): Likewise.
12464 (thumb1_rotrsi3): Likewise.
12465 (shiftsi3_compare0_scratch): Likewise.
12466 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
12467 (neon_mov<mode>): Likewise.
12468 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
12470 (thumb2_movsi_insn): Likewise.
12471 (thumb2_cmpsi_neg_shiftsi): Likewise.
12472 (thumb2_extendqisi_v6): Likewise.
12473 (thumb2_zero_extendhisi2_v6): Likewise.
12474 (thumb2_zero_extendqisi2_v6): Likewise.
12475 (thumb2_shiftsi3_short): Likewise.
12476 (thumb2_addsi3_compare0_scratch): Likewise.
12477 (orsi_not_shiftsi_si): Likewise.
12478 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
12479 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
12481 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
12482 (1020alu_shift_op): Likewise.
12483 (1020alu_shift_reg_op): Likewise.
12484 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
12485 (alu_shift_op): Likewise.
12486 (alu_shift_reg_op): Likewise.
12487 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
12488 (11_alu_shift_op): Likewise.
12489 (11_alu_shift_reg_op): Likewise.
12490 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
12491 (9_alu_shift_reg_op): Likewise.
12492 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12494 (cortex_a15_alu_shift): Likewise.
12495 (cortex_a15_alu_shift_reg): Likewise.
12496 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
12498 (cortex_a5_alu_shift): Likewise.
12499 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
12501 (cortex_a53_alu_shift): Likewise.
12502 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12504 (cortex_a7_alu_reg): Likewise.
12505 (cortex_a7_alu_shift): Likewise.
12506 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
12508 (cortex_a8_alu_shift): Likewise.
12509 (cortex_a8_alu_shift_reg): Likewise.
12510 (cortex_a8_mov): Likewise.
12511 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
12512 (cortex_a9_dp_shift): Likewise.
12513 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
12515 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
12517 (cortex_r4_mov): Likewise.
12518 (cortex_r4_alu_shift): Likewise.
12519 (cortex_r4_alu_shift_reg): Likewise.
12520 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
12521 (526_alu_shift_op): Likewise.
12522 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
12523 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
12524 (626te_alu_shift_op): Likewise.
12525 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
12526 (726te_alu_op): Likewise.
12527 (726te_alu_shift_op): Likewise.
12528 (726te_alu_shift_reg_op): Likewise.
12529 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
12530 (mp626_alu_shift_op): Likewise.
12531 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
12532 (pj4_alu_e1_conds): Likewise.
12533 (pj4_alu): Likewise.
12534 (pj4_alu_conds): Likewise.
12535 (pj4_shift): Likewise.
12536 (pj4_shift_conds): Likewise.
12537 (pj4_alu_shift): Likewise.
12538 (pj4_alu_shift_conds): Likewise.
12539 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12541 (cortexa7_older_only): Likewise.
12542 (cortexa7_younger): Likewise.
12544 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12546 * ipa-pure-const.c (generate_summary): Rename to...
12547 (pure_const_generate_summary): ... this.
12549 2013-07-17 Iain Sandoe <iain@codesourcery.com>
12551 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
12553 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
12556 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
12559 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12561 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
12562 enabled without -march=zEC12.
12563 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
12566 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12568 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
12569 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
12570 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
12571 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
12572 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
12573 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
12574 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
12576 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
12577 and ISA_HAS_NMADD3_NMSUB3.
12578 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
12579 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
12580 (nmsub4<mode>, nmsub3<mode>): Likewise.
12581 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
12583 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12585 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
12586 TARGET_MIPS5400 checking.
12588 2013-07-16 Jakub Jelinek <jakub@redhat.com>
12589 Peter Bergner <bergner@vnet.ibm.com>
12591 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
12592 registers in the comment.
12593 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
12594 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
12595 rather than FIRST_PSEUDO_REGISTERS.
12597 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
12599 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
12600 enable extra ISA flags with TARGET_HTM.
12602 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12604 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
12607 2013-07-15 Cong Hou <congh@google.com>
12609 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
12610 in compare function for sorting.
12612 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
12614 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
12615 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
12616 * config/rs6000/rs6000.opt: Add -mhtm option.
12617 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
12618 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
12619 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12620 __HTM__ if the HTM instructions are available.
12621 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
12622 htm_spr_reg_operand): New define_predicates.
12623 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
12624 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
12626 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
12627 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
12628 HTM builtin functions.
12629 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
12630 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
12631 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
12632 (rs6000_builtin_mask_calculate): Likewise.
12633 (rs6000_option_override_internal): Likewise.
12634 (bdesc_htm): Add new HTM builtin support.
12635 (htm_spr_num): New function.
12636 (htm_spr_regno): Likewise.
12637 (rs6000_htm_spr_icode): Likewise.
12638 (htm_expand_builtin): Likewise.
12639 (htm_init_builtins): Likewise.
12640 (rs6000_expand_builtin): Add support for HTM builtin functions.
12641 (rs6000_init_builtins): Likewise.
12642 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
12644 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
12645 (TARGET_HTM, MASK_HTM): Define macros.
12646 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
12647 (FIXED_REGISTERS): Likewise.
12648 (CALL_USED_REGISTERS): Likewise.
12649 (CALL_REALLY_USED_REGISTERS): Likewise.
12650 (REG_ALLOC_ORDER): Likewise.
12651 (enum reg_class): Likewise.
12652 (REG_CLASS_NAMES): Likewise.
12653 (REG_CLASS_CONTENTS): Likewise.
12654 (REGISTER_NAMES): Likewise.
12655 (ADDITIONAL_REGISTER_NAMES): Likewise.
12656 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
12657 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
12658 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
12659 * config/rs6000/htm.md: New file.
12660 * config/rs6000/htmintrin.h: New file.
12661 * config/rs6000/htmxlintrin.h: New file.
12663 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
12665 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
12666 Define SYMBOL_TINY_GOT, update comment.
12667 * config/aarch64/aarch64.c
12668 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
12669 (aarch64_expand_mov_immediate): Likewise.
12670 (aarch64_print_operand): Likewise.
12671 (aarch64_classify_symbol): Likewise.
12672 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
12673 (ldr_got_tiny): Define.
12675 2013-07-13 Tobias Grosser <tobias@grosser.es>
12677 PR tree-optimization/54094
12678 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
12679 scheduling dimension for the parallelism check from the polyhedral
12680 information in the AST.
12681 * graphite-dependences.c (carries_deps): Do not assume the schedule is
12684 2013-07-13 Jason Merrill <jason@redhat.com>
12686 * print-tree.c (debug_vec_tree): Use debug_raw.
12687 (debug_raw (vec<tree, va_gc> &)): New.
12688 (debug_raw (vec<tree, va_gc> *)): New.
12689 * tree.h: Declare them.
12691 2013-07-13 Bin Cheng <bin.cheng@arm.com>
12693 * ifcvt.c (ifcvt_after_combine): New static variable.
12694 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
12696 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
12697 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
12698 rest_of_handle_if_after_reload): Pass new argument for if_convert.
12700 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
12702 * config/mips/mips.c (mips_expand_call): Remove empty statement.
12704 2013-07-12 Michael Matz <matz@suse.de>
12706 PR middle-end/55771
12707 * convert.c (convert_to_real): Reject non-float inner types.
12709 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
12711 * config/aarch64/aarch64-protos.h
12712 (aarch64_simd_immediate_valid_for_move): Remove.
12713 * config/aarch64/aarch64.c (simd_immediate_info): New member.
12714 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
12716 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
12718 2013-07-11 Steve Ellcey <sellcey@mips.com>
12720 * config/mips/mips.c (mips_conditional_register_usage): Do not
12721 use t[0-7] registers in MIPS16 mode when optimizing for size.
12723 2013-07-11 Sriraman Tallam <tmsriram@google.com>
12725 * config/i386/i386.c (dispatch_function_versions): Fix array
12726 indexing of function_version_info to match actual_versions.
12728 2013-07-11 Teresa Johnson <tejohnson@google.com>
12730 * vec.h (struct va_gc): Move release out-of-line.
12731 (va_gc::release): Call ggc_free on released vec.
12733 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12735 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
12736 Require GOT register as additional operand in UNSPEC.
12737 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
12738 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12739 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12740 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
12741 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
12742 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
12743 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
12745 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
12748 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
12749 name seen by assembler/linker rather if available.
12751 2013-07-11 Andreas Schwab <schwab@suse.de>
12753 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
12755 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
12757 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
12759 2013-07-10 Joseph Myers <joseph@codesourcery.com>
12761 * doc/tm.texi.in: Move hook documentation to ....
12762 * target.def: ... here.
12764 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
12765 text on @hook line.
12766 * doc/tm.texi: Regenerate.
12768 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
12771 * doc/invoke.texi: Document -Wconditionally-supported.
12773 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12776 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
12779 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12782 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
12783 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
12784 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
12785 Remove duplicate devices.
12786 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
12787 * config/avr/t-multilib: Regenerate.
12788 * config/avr/avr-tables.opt: Regenerate.
12789 * doc/avr-mmcu.texi: Regenerate.
12791 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12794 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
12796 2013-07-10 Graham Stott <graham.stott@btinternet.com>
12798 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
12799 the cost of MULT when optimizing for size.
12801 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12803 * config/cr16/cr16-protos.h: Don't include target.h.
12805 2013-07-09 Joseph Myers <joseph@codesourcery.com>
12807 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
12808 adjust register size for TDmode and TFmode for VSX registers.
12810 2013-07-08 Kai Tietz <ktietz@redhat.com>
12813 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
12814 hook_bool_const_tree_true.
12816 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12818 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
12819 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
12820 * config/s390/s390.md: Define FPR*_REGNUM constants.
12821 Fix FPR2_REGNUM constant (18 -> 17).
12822 ("*trunc<BFP:mode><DFP_ALL:mode>2")
12823 ("*trunc<DFP_ALL:mode><BFP:mode>2")
12824 ("trunc<BFP:mode><DFP_ALL:mode>2")
12825 ("trunc<DFP_ALL:mode><BFP:mode>2")
12826 ("*extend<BFP:mode><DFP_ALL:mode>2")
12827 ("*extend<DFP_ALL:mode><BFP:mode>2")
12828 ("extend<BFP:mode><DFP_ALL:mode>2")
12829 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
12832 2013-07-08 Graham Stott <graham.stott@btinternet.com>
12834 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
12836 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12838 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
12839 and cfun_fpr_bit_p to cfun_fpr_save_p.
12840 (s390_frame_area, s390_register_info, s390_frame_info)
12841 (s390_emit_prologue, s390_emit_epilogue)
12842 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
12844 * config/s390/s390.h: Define *_REGNUM macros for floating point
12847 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
12849 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
12851 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
12853 PR rtl-optimization/57786
12854 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
12855 and break out of the loop when it is set to false.
12857 2013-07-08 Jakub Jelinek <jakub@redhat.com>
12860 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
12861 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
12862 (const_int 63)) 0)).
12863 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
12864 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
12865 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
12867 PR rtl-optimization/57829
12868 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
12869 mask bits outside of mode are just sign-extension from mode to HWI.
12871 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
12873 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
12874 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
12875 adjust_address instead of change_address to keep info about alignment.
12876 (emit_strmov): Remove.
12877 (emit_memmov): New function.
12878 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
12879 (expand_movmem_epilogue): Likewise and return updated rtx for
12881 (expand_constant_movmem_prologue): Likewise and return updated rtx for
12882 destination and source.
12883 (decide_alignment): Refactor, handle vector_loop.
12884 (ix86_expand_movmem): Likewise.
12885 (ix86_expand_setmem): Likewise.
12886 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
12888 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
12890 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
12891 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
12893 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
12895 * config/i386/sse.md (sse_movlhps): Change alternative 3
12896 of operand 2 to "m".
12898 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
12901 * config/i386/sse.md (iptr): New mode attribute.
12902 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
12903 (<sse>_vm<plusminus_insn><mode>3): Ditto.
12904 (<sse>_vmmul<mode>3): Ditto.
12905 (<sse>_vmdiv<mode>3): Ditto.
12906 (sse_vmrcpv4sf2): Ditto.
12907 (<sse>_vmsqrt<mode>2): Ditto.
12908 (sse_vmrsqrtv4sf2): Ditto.
12909 (<sse>_vm<code><mode>3): Ditto.
12910 (avx_vmcmp<mode>3): Ditto.
12911 (<sse>_vmmaskcmp<mode>3): Ditto.
12912 (<sse>_comi): Ditto.
12913 (<sse>_ucomi): Ditto.
12914 (*xop_vmfrcz_<mode>): Ditto.
12915 (*fmai_fmadd_<mode>): Ditto.
12916 (*fmai_fmsub_<mode>): Ditto.
12917 (*fmai_fnmadd_<mode>): Ditto.
12918 (*fmai_fnmsub_<mode>): Ditto.
12919 (*fma4i_vmfmadd_<mode>): Ditto.
12920 (*fma4i_vmfmsub_<mode>): Ditto.
12921 (*fma4i_vmfnmadd_<mode>): Ditto.
12922 (*fma4i_vmfnmsub_<mode>): Ditto.
12923 (*xop_vmfrcz_<mode>): Ditto.
12924 (sse_cvtps2pi): Ditto.
12925 (sse_cvttps2pi): Ditto.
12926 (sse_cvtss2si): Ditto.
12927 (sse_cvtss2si_2): Ditto.
12928 (sse_cvtss2siq_2): Ditto.
12929 (sse_cvttss2si): Ditto.
12930 (sse_cvttss2siq): Ditto.
12931 (sse_cvtsd2si): Ditto.
12932 (sse_cvtsd2si_2): Ditto.
12933 (sse_cvtsd2siq_2): Ditto.
12934 (sse_cvttsd2si): Ditto.
12935 (sse_cvttsd2siq): Ditto.
12936 (sse_cvtsd2ss): Ditto.
12937 (sse_cvtss2sd): Ditto.
12938 (avx2_pbroadcast<mode>): Ditto.
12939 (avx2_pbroadcast<mode>_1): Ditto.
12940 (*avx_vperm_broadcast_v4sf): Ditto.
12942 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
12943 (sse_movlhps): Ditto.
12944 (sse_storehps): Ditto.
12945 (sse_loadhps): Ditto.
12946 (sse_storelps): Ditto.
12947 (sse_loadlps): Ditto.
12948 (*vec_concatv4sf): Ditto.
12949 (*vec_interleave_highv2df): Ditto.
12950 (*vec_interleave_lowv2df): Ditto.
12951 (*vec_extractv2df_1_sse): Ditto.
12952 (*vec_extractv2df_0_sse): Ditto.
12953 (sse2_storelpd): Ditto.
12954 (sse2_loadlpd): Ditto.
12955 (sse2_movsd): Ditto.
12956 (*vec_concatv4si): Ditto.
12957 (vec_concatv2di): Ditto.
12959 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
12960 for Intel asm dialect.
12961 (mmx_punpcklwd): Ditto.
12962 (mmx_punpckldq): Ditto.
12964 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
12965 for intel assembler dialect.
12967 2013-07-06 Jakub Jelinek <jakub@redhat.com>
12970 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
12971 for BUILT_IN_C{LZ,LRSB}*.
12972 * tree.h (CASE_INT_FN): Add FN##IMAX case.
12973 * tree-vrp.c (extract_range_basic): Handle
12974 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
12975 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
12976 fall thru to code calling set_value*.
12977 * builtins.c (expand_builtin): Remove *IMAX cases.
12978 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
12979 if width is bigger than 2*HWI.
12981 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
12983 PR rtl-optimization/55342
12984 * lra-int.h (lra_subreg_reload_pseudos): New.
12985 * lra.c: Add undoing optional reloads to the block diagram.
12986 (lra_subreg_reload_pseudos): New.
12987 (lra_optional_reload_pseudos): Change comments.
12988 (lra): Init and clear lra_subreg_reload_pseudos. Clear
12989 lra_optional_reload_pseudos after undo transformations.
12990 * lra-assigns.c (pseudo_prefix_title): New.
12991 (lra_setup_reg_renumber): Use it.
12992 (spill_for): Ditto. Check subreg reload pseudos too.
12993 (assign_by_spills): Consider subreg reload pseudos too.
12994 * lra-constraints.c (simplify_operand_subreg): Use
12995 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
12996 (curr_insn_transform): Recognize and do optional reloads.
12997 (undo_optional_reloads): New.
12998 (lra_undo_inheritance): Call undo_optional_reloads.
13000 2013-07-05 Thomas Quinot <quinot@adacore.com>
13002 * tree-complex.c (expand_complex_operations_1): Fix typo.
13004 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
13006 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
13007 (tune_params): New member 'const vec_costs'.
13008 * config/aarch64/aarch64.c (generic_vector_cost): New.
13009 (generic_tunings): New member 'generic_vector_cost'.
13010 (aarch64_builtin_vectorization_cost): New.
13011 (aarch64_add_stmt_cost): New.
13012 (TARGET_VECTORIZE_ADD_STMT_COST): New.
13013 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
13015 2013-07-03 Jakub Jelinek <jakub@redhat.com>
13018 * config/i386/predicates.md (vsib_address_operand): Disallow
13019 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
13021 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
13023 PR middle-end/55030
13024 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
13025 expand_builtin_setjmp_receiver.
13026 (expand_label): Adjust, call expand_builtin_setjmp_receiver
13027 with NULL for the label parameter.
13028 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
13029 the frame-pointer. Adjust comments.
13030 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
13031 only if LABEL is non-NULL.
13033 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
13035 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
13036 (ARM_ABI_AAPCS64): Ditto.
13038 (ARM_DEFAULT_ABI): Ditto.
13040 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
13042 * config/aarch64/aarch64-builtins.c
13043 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
13044 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
13046 * config/aarch64/aarch64-simd.md
13047 (aarch64_ld1<VALL:mode>): New.
13048 (aarch64_st1<VALL:mode>): Likewise.
13049 * config/aarch64/arm_neon.h
13050 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
13052 2013-07-02 Sriraman Tallam <tmsriram@google.com>
13054 * config/i386/i386.c (gate_insert_vzeroupper): Check if
13056 (ix86_option_override_internal):Turn on all -mavx target flags by
13057 default as they are dependent on AVX anyway.
13059 2013-07-02 Cary Coutant <ccoutant@google.com>
13061 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
13062 deterministic hash.
13063 (loc_checksum_ordered): Likewise.
13064 (hash_loc_operands): Remove inline keyword.
13066 2013-07-02 Jakub Jelinek <jakub@redhat.com>
13068 PR tree-optimization/57741
13069 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
13070 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
13071 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
13072 Allow REAL_CST step_exprs if flag_associative_math.
13073 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
13075 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13077 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
13079 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13081 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
13083 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13085 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
13087 (iorsi3_insn): Likewise.
13088 (arm_xorsi3): Likewise.
13090 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
13092 * arm.md (attribute "wtype"): Delete. Move attribute values from here
13094 (attribute "type"): ... here, and prefix with "wmmx_".
13095 (attribute "core_cycles"): Update for attribute changes.
13096 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
13097 (tbcstv4hi): Likewise.
13098 (tbcstv2si): Likewise.
13099 (iwmmxt_iordi3): Likewise.
13100 (iwmmxt_xordi3): Likewise.
13101 (iwmmxt_anddi3): Likewise.
13102 (iwmmxt_nanddi3): Likewise.
13103 (iwmmxt_arm_movdi): Likewise.
13104 (iwmmxt_movsi_insn): Likewise.
13105 (mov<mode>_internal): Likewise.
13106 (and<mode>3_iwmmxt): Likewise.
13107 (ior<mode>3_iwmmxt): Likewise.
13108 (xor<mode>3_iwmmxt): Likewise.
13109 (add<mode>3_iwmmxt): Likewise.
13110 (ssaddv8qi3): Likewise.
13111 (ssaddv4hi3): Likewise.
13112 (ssaddv2si3): Likewise.
13113 (usaddv8qi3): Likewise.
13114 (usaddv4hi3): Likewise.
13115 (usaddv2si3): Likewise.
13116 (sub<mode>3_iwmmxt): Likewise.
13117 (sssubv8qi3): Likewise.
13118 (sssubv4hi3): Likewise.
13119 (sssubv2si3): Likewise.
13120 (ussubv8qi3): Likewise.
13121 (ussubv4hi3): Likewise.
13122 (ussubv2si3): Likewise.
13123 (mulv4hi3_iwmmxt): Likewise.
13124 (smulv4hi3_highpart): Likewise.
13125 (umulv4hi3_highpart): Likewise.
13126 (iwmmxt_wmacs): Likewise.
13127 (iwmmxt_wmacsz): Likewise.
13128 (iwmmxt_wmacu): Likewise.
13129 (iwmmxt_wmacuz): Likewise.
13130 (iwmmxt_clrdi): Likewise.
13131 (iwmmxt_clrv8qi): Likewise.
13132 (iwmmxt_clr4hi): Likewise.
13133 (iwmmxt_clr2si): Likewise.
13134 (iwmmxt_uavgrndv8qi3): Likewise.
13135 (iwmmxt_uavgrndv4hi3): Likewise.
13136 (iwmmxt_uavgv8qi3): Likewise.
13137 (iwmmxt_uavgv4hi3): Likewise.
13138 (iwmmxt_tinsrb): Likewise.
13139 (iwmmxt_tinsrh): Likewise.
13140 (iwmmxt_tinsrw): Likewise.
13141 (iwmmxt_textrmub): Likewise.
13142 (iwmmxt_textrmsb): Likewise.
13143 (iwmmxt_textrmuh): Likewise.
13144 (iwmmxt_textrmsh): Likewise.
13145 (iwmmxt_textrmw): Likewise.
13146 (iwmxxt_wshufh): Likewise.
13147 (eqv8qi3): Likewise.
13148 (eqv4hi3): Likewise.
13149 (eqv2si3): Likewise.
13150 (gtuv8qi3): Likewise.
13151 (gtuv4hi3): Likewise.
13152 (gtuv2si3): Likewise.
13153 (gtv8qi3): Likewise.
13154 (gtv4hi3): Likewise.
13155 (gtv2si3): Likewise.
13156 (smax<mode>3_iwmmxt): Likewise.
13157 (umax<mode>3_iwmmxt): Likewise.
13158 (smin<mode>3_iwmmxt): Likewise.
13159 (umin<mode>3_iwmmxt): Likewise.
13160 (iwmmxt_wpackhss): Likewise.
13161 (iwmmxt_wpackwss): Likewise.
13162 (iwmmxt_wpackdss): Likewise.
13163 (iwmmxt_wpackhus): Likewise.
13164 (iwmmxt_wpackwus): Likewise.
13165 (iwmmxt_wpackdus): Likewise.
13166 (iwmmxt_wunpckihb): Likewise.
13167 (iwmmxt_wunpckihh): Likewise.
13168 (iwmmxt_wunpckihw): Likewise.
13169 (iwmmxt_wunpckilb): Likewise.
13170 (iwmmxt_wunpckilh): Likewise.
13171 (iwmmxt_wunpckilw): Likewise.
13172 (iwmmxt_wunpckehub): Likewise.
13173 (iwmmxt_wunpckehuh): Likewise.
13174 (iwmmxt_wunpckehuw): Likewise.
13175 (iwmmxt_wunpckehsb): Likewise.
13176 (iwmmxt_wunpckehsh): Likewise.
13177 (iwmmxt_wunpckehsw): Likewise.
13178 (iwmmxt_wunpckelub): Likewise.
13179 (iwmmxt_wunpckeluh): Likewise.
13180 (iwmmxt_wunpckeluw): Likewise.
13181 (iwmmxt_wunpckelsb): Likewise.
13182 (iwmmxt_wunpckelsh): Likewise.
13183 (iwmmxt_wunpckelsw): Likewise.
13184 (ror<mode>3): Likewise.
13185 (ashr<mode>3_iwmmxt): Likewise.
13186 (lshr<mode>3_iwmmxt): Likewise.
13187 (ashl<mode>3_iwmmxt): Likewise.
13188 (ror<mode>3_di): Likewise.
13189 (ashr<mode>3_di): Likewise.
13190 (lshr<mode>3_di): Likewise.
13191 (ashl<mode>3_di): Likewise.
13192 (iwmmxt_wmadds): Likewise.
13193 (iwmmxt_wmaddu): Likewise.
13194 (iwmmxt_tmia): Likewise.
13195 (iwmmxt_tmiaph): Likewise.
13196 (iwmmxt_tmiabb): Likewise.
13197 (iwmmxt_tmiatb): Likewise.
13198 (iwmmxt_tmiabt): Likewise.
13199 (iwmmxt_tmiatt): Likewise.
13200 (iwmmxt_tmovmskb): Likewise.
13201 (iwmmxt_tmovmskh): Likewise.
13202 (iwmmxt_tmovmskw): Likewise.
13203 (iwmmxt_waccb): Likewise.
13204 (iwmmxt_wacch): Likewise.
13205 (iwmmxt_waccw): Likewise.
13206 (iwmmxt_waligni): Likewise.
13207 (iwmmxt_walignr): Likewise.
13208 (iwmmxt_walignr0): Likewise.
13209 (iwmmxt_walignr1): Likewise.
13210 (iwmmxt_walignr2): Likewise.
13211 (iwmmxt_walignr3): Likewise.
13212 (iwmmxt_wsadb): Likewise.
13213 (iwmmxt_wsadh): Likewise.
13214 (iwmmxt_wsadbz): Likewise.
13215 (iwmmxt_wsadhz): Likewise.
13216 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
13217 (iwmmxt_wabsdiffb): Likewise.
13218 (iwmmxt_wabsdiffh): Likewise.
13219 (iwmmxt_wabsdiffw): Likewise.
13220 (iwmmxt_waddsubhx): Likewise
13221 (iwmmxt_wsubaddhx): Likewise.
13222 (addc<mode>3): Likewise.
13223 (iwmmxt_avg4): Likewise.
13224 (iwmmxt_avg4r): Likewise.
13225 (iwmmxt_wmaddsx): Likewise.
13226 (iwmmxt_wmaddux): Likewise.
13227 (iwmmxt_wmaddsn): Likewise.
13228 (iwmmxt_wmaddun): Likewise.
13229 (iwmmxt_wmulwsm): Likewise.
13230 (iwmmxt_wmulwum): Likewise.
13231 (iwmmxt_wmulsmr): Likewise.
13232 (iwmmxt_wmulumr): Likewise.
13233 (iwmmxt_wmulwsmr): Likewise.
13234 (iwmmxt_wmulwumr): Likewise.
13235 (iwmmxt_wmulwl): Likewise.
13236 (iwmmxt_wqmulm): Likewise.
13237 (iwmmxt_wqmulwm): Likewise.
13238 (iwmmxt_wqmulmr): Likewise.
13239 (iwmmxt_wqmulwmr): Likewise.
13240 (iwmmxt_waddbhusm): Likewise.
13241 (iwmmxt_waddbhusl): Likewise.
13242 (iwmmxt_wqmiabb): Likewise.
13243 (iwmmxt_wqmiabt): Likewise.
13244 (iwmmxt_wqmiatb): Likewise.
13245 (iwmmxt_wqmiatt): Likewise.
13246 (iwmmxt_wqmiabbn): Likewise.
13247 (iwmmxt_wqmiabtn): Likewise.
13248 (iwmmxt_wqmiatbn): Likewise.
13249 (iwmmxt_wqmiattn): Likewise.
13250 (iwmmxt_wmiabb): Likewise.
13251 (iwmmxt_wmiabt): Likewise.
13252 (iwmmxt_wmiatb): Likewise.
13253 (iwmmxt_wmiatt): Likewise.
13254 (iwmmxt_wmiabbn): Likewise.
13255 (iwmmxt_wmiabtn): Likewise.
13256 (iwmmxt_wmiatbn): Likewise.
13257 (iwmmxt_wmiattn): Likewise.
13258 (iwmmxt_wmiawbb): Likewise.
13259 (iwmmxt_wmiawbt): Likewise.
13260 (iwmmxt_wmiawtb): Likewise.
13261 (iwmmxt_wmiawtt): Likewise.
13262 (iwmmxt_wmiawbbn): Likewise.
13263 (iwmmxt_wmiawbtn): Likewise.
13264 (iwmmxt_wmiawtbn): Likewise.
13265 (iwmmxt_wmiawttn): Likewise.
13266 (iwmmxt_wmerge): Likewise.
13267 (iwmmxt_tandc<mode>3): Likewise.
13268 (iwmmxt_torc<mode>3): Likewise.
13269 (iwmmxt_torvsc<mode>3): Likewise.
13270 (iwmmxt_textrc<mode>3): Likewise.
13271 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
13272 (wmmxt_pack): Likewise.
13273 (wmmxt_mult_c1): Likewise.
13274 (wmmxt_mult_c2): Likewise.
13275 (wmmxt_alu_c1): Likewise.
13276 (wmmxt_alu_c2): Likewise.
13277 (wmmxt_alu_c3): Likewise.
13278 (wmmxt_transfer_c1): Likewise.
13279 (wmmxt_transfer_c2): Likewise.
13280 (wmmxt_transfer_c3): Likewise.
13281 (marvell_f_iwmmxt_wstr): Likewise.
13282 (marvell_f_iwmmxt_wldr): Likewise.
13284 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
13286 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
13288 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13291 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13292 * lra-constraints.c (need_for_split_p): Check call used hard regs
13293 living through calls.
13295 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
13296 call used regs for call insn.
13298 2013-06-28 Jakub Jelinek <jakub@redhat.com>
13301 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
13302 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
13304 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
13306 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
13308 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13310 * lra-constraints.c (need_for_split_p): Check call used hard regs
13311 living through calls.
13313 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
13316 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
13317 to tie with any other modes. Eliminate Altivec vector mode tests,
13318 since these are a subset of ALTIVEC or VSX vector modes. Simplify
13319 code, to return 0 if testing MODE2 for a condition, if we've
13320 already tested MODE1 for the same condition.
13322 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13324 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
13327 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13329 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
13330 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
13332 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13334 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
13336 (aarch64_symbolic_constant_p): Remove.
13337 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
13338 static. Fix line length and white space.
13339 (aarch64_symbolic_constant_p): Remove.
13340 * config/aarch64/predicates.md (aarch64_valid_symref):
13341 Use aarch64_classify_symbol_expression.
13343 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13345 * config/arm/constraints.md (Ts): New constraint.
13346 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
13348 (compare_scc): Use "Ts" constraint for operand 0.
13349 (ior_scc_scc): Likewise.
13350 (and_scc_scc): Likewise.
13351 (and_scc_scc_nodom): Likewise.
13352 (ior_scc_scc_cmp): Likewise for operand 7.
13353 (and_scc_scc_cmp): Likewise.
13354 * config/arm/thumb2.md (thumb2_movsi_insn):
13355 Add alternatives for 16-bit encodings.
13356 (thumb2_movhi_insn): Likewise.
13357 (thumb2_movsicc_insn): Likewise.
13358 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
13359 (thumb2_negscc): Use "Ts" constraint.
13360 Move mvn instruction outside cond_exec block.
13361 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
13362 for 16-bit encodings.
13364 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13366 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
13368 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
13369 (mulsi3subsi): Likewise.
13370 (mulsidi3adddi): Likewise.
13371 (mulsidi3_v6): Likewise.
13372 (umulsidi3_v6): Likewise.
13373 (umulsidi3adddi_v6): Likewise.
13374 (smulsi3_highpart_v6): Likewise.
13375 (umulsi3_highpart_v6): Likewise.
13376 (mulhisi3tb): Likewise.
13377 (mulhisi3bt): Likewise.
13378 (mulhisi3tt): Likewise.
13379 (maddhisi4): Likewise.
13380 (maddhisi4tb): Likewise.
13381 (maddhisi4tt): Likewise.
13382 (maddhidi4): Likewise.
13383 (maddhidi4tb): Likewise.
13384 (maddhidi4tt): Likewise.
13385 (zeroextractsi_compare0_scratch): Likewise.
13386 (insv_zero): Likewise.
13387 (insv_t2): Likewise.
13388 (anddi_notzesidi_di): Likewise.
13389 (anddi_notsesidi_di): Likewise.
13390 (andsi_notsi_si): Likewise.
13391 (iordi_zesidi_di): Likewise.
13392 (xordi_zesidi_di): Likewise.
13393 (andsi_iorsi3_notsi): Likewise.
13394 (smax_0): Likewise.
13395 (smax_m1): Likewise.
13396 (smin_0): Likewise.
13397 (not_shiftsi): Likewise.
13398 (unaligned_loadsi): Likewise.
13399 (unaligned_loadhis): Likewise.
13400 (unaligned_loadhiu): Likewise.
13401 (unaligned_storesi): Likewise.
13402 (unaligned_storehi): Likewise.
13403 (extv_reg): Likewise.
13404 (extzv_t2): Likewise.
13405 (divsi3): Likewise.
13406 (udivsi3): Likewise.
13407 (arm_zero_extendhisi2addsi): Likewise.
13408 (arm_zero_extendqisi2addsi): Likewise.
13409 (compareqi_eq0): Likewise.
13410 (arm_extendhisi2_v6): Likewise.
13411 (arm_extendqisi2addsi): Likewise.
13412 (arm_movt): Likewise.
13413 (thumb2_ldrd): Likewise.
13414 (thumb2_ldrd_base): Likewise.
13415 (thumb2_ldrd_base_neg): Likewise.
13416 (thumb2_strd): Likewise.
13417 (thumb2_strd_base): Likewise.
13418 (thumb2_strd_base_neg): Likewise.
13419 (arm_negsi2): Add alternative for 16-bit encoding.
13420 (arm_one_cmplsi2): Likewise.
13422 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13424 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
13425 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
13426 (movdfcc): Likewise.
13427 * config/arm/vfp.md (*thumb2_movsf_vfp):
13428 Disable predication for arm_restrict_it.
13429 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
13430 (*thumb2_movdfcc_vfp): Likewise.
13431 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
13432 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
13433 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
13434 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
13435 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
13436 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
13437 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
13438 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
13439 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
13440 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
13441 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
13442 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
13443 Disable predication for arm_restrict_it.
13445 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
13447 * config/i386/bmiintrin.h (_bextr_u32): New.
13448 (_bextr_u64): Ditto.
13450 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
13452 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
13453 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
13454 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
13455 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
13456 * config/mips/n32-elf.h: ...this new file.
13458 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
13461 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
13462 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
13464 2013-06-27 Catherine Moore <clm@codesourcery.com>
13466 * config/mips/mips-tables.opt: Regenerate.
13467 * config/mips/mips-cpus.def: Add m14ke and m14kec.
13468 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
13469 * doc/invoke.texi: Add -m14kc.
13471 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13474 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
13475 constraints of operand 1 and 2.
13478 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
13479 to match RTL canonicalization. Swap predicates and
13480 constraints of operand 1 and 2.
13482 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
13484 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
13485 Process OP_INOUT regs for splitting too.
13487 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13489 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
13490 decl before the loop, initialize to NULL.
13491 (vectorizable_load): Initialize ptr_incr to NULL.
13493 2013-06-27 Martin Jambor <mjambor@suse.cz>
13496 * ipa-ref.h (ipa_maybe_record_reference): Declare.
13497 * ipa-ref.c (ipa_maybe_record_reference): New function.
13498 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
13499 * ipa-cp.c (create_specialized_node): Record potential references from
13501 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
13503 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13505 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
13506 parameter 'mode' of type 'enum machine_mode mode'; change to pass
13507 'mode' to force_reg.
13508 (aarch64_add_offset): Update calls to aarch64_force_temporary.
13509 (aarch64_expand_mov_immediate): Likewise.
13511 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13513 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
13514 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
13516 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13518 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
13519 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13520 (struct machine_function): Add tbegin_p.
13521 (s390_canonicalize_comparison): Fold CC mode compares to
13522 conditional jump if possible.
13523 (s390_emit_jump): Return the emitted jump.
13524 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
13525 Handle CCRAWmode compares.
13526 (s390_option_override): Default to -mhtm if available.
13527 (s390_reg_clobbered_rtx): Handle floating point regs as well.
13528 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
13529 FPRs instead of df_regs_ever_live_p.
13530 (s390_optimize_nonescaping_tx): New function.
13531 (s390_init_frame_layout): Extend clobbered_regs array to cover
13533 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
13534 (s390_expand_tbegin): New function.
13535 (enum s390_builtin): New enum definition.
13536 (code_for_builtin): New array definition.
13537 (s390_init_builtins): New function.
13538 (s390_expand_builtin): New function.
13539 (TARGET_INIT_BUILTINS): Define.
13540 (TARGET_EXPAND_BUILTIN): Define.
13541 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
13542 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
13543 (s390_alc_comparison): Likewise.
13544 * config/s390/s390-modes.def: Add CCRAWmode.
13545 * config/s390/s390.h (processor_flags): Add PF_TX.
13546 (TARGET_CPU_HTM): Define macro.
13547 (TARGET_HTM): Define macro.
13548 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
13549 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
13550 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13551 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
13552 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
13553 (TBEGIN_MASK, TBEGINC_MASK): New constants.
13554 ("*cc_to_int"): Move up.
13555 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
13556 constants other than 0.
13557 ("*ccraw_to_int"): New insn and splitter definition.
13558 ("tbegin", "tbegin_nofloat", "tbegin_retry")
13559 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
13560 ("tx_assist"): New expander.
13561 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
13562 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
13563 * config/s390/s390.opt: Add -mhtm option.
13564 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
13565 * config/s390/htmxlintrin.h: New file.
13566 * config/s390/htmintrin.h: New file.
13567 * config/s390/s390intrin.h: New file.
13568 * doc/extend.texi: Document htm builtins.
13569 * config.gcc: Add the new header files to extra_headers.
13571 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
13573 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
13574 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
13576 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
13577 Pat Haugen <pthaugen@us.ibm.com>
13578 Peter Bergner <bergner@vnet.ibm.com>
13580 * config/rs6000/power8.md: New.
13581 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
13582 setting for power8 entry.
13583 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
13584 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
13585 test for Power4/Power5 only.
13586 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
13588 (force_new_group): Adjust comment.
13589 * config/rs6000/rs6000.md: Include power8.md.
13591 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
13593 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
13594 * config/arm/arm-protos.h (arm_max_conditional_execute): New
13596 (tune_params): Update comment.
13597 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
13598 (arm_max_conditional_execute): New function.
13599 (thumb2_final_prescan_insn): Use max_insn_skipped and
13600 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
13602 2013-06-25 Jakub Jelinek <jakub@redhat.com>
13604 PR tree-optimization/57705
13605 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
13606 SSA_NAME step, provided that it is not defined inside the loop.
13607 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
13608 (get_initial_def_for_induction): Handle SSA_NAME IV step.
13610 2013-06-25 Martin Jambor <mjambor@suse.cz>
13612 PR middle-end/57670
13613 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
13614 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
13616 (ipa_note_param_call): Initialize member_ptr flag.
13617 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
13618 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
13619 (ipa_write_indirect_edge_info): Stream member_ptr flag.
13620 (ipa_read_indirect_edge_info): Likewise.
13622 2013-06-25 Richard Biener <rguenther@suse.de>
13624 PR middle-end/56977
13625 * passes.c (init_optimization_passes): Move pass_fold_builtins
13626 and pass_dce earlier with -Og.
13628 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
13630 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
13631 <BIT_FIELD_REF>: Remove trailing TAB.
13632 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
13635 2013-06-24 Martin Jambor <mjambor@suse.cz>
13637 PR tree-optimization/57358
13638 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
13639 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
13640 (ipa_analyze_params_uses): Generate pessimistic info when true.
13642 2013-06-24 Martin Jambor <mjambor@suse.cz>
13644 PR tree-optimization/57539
13645 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
13646 global.inlined_to of the new node to it. All callers changed.
13647 * ipa-inline-transform.c (clone_inlined_nodes): New variable
13648 inlining_into, pass it to cgraph_clone_node.
13649 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
13650 ipa_free_edge_args_substructures.
13651 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
13652 rdesc linked list. Do not assert rdesc edges have inlined caller.
13653 Assert we have found an rdesc in the rdesc list.
13655 2013-06-24 Richard Biener <rguenther@suse.de>
13657 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
13658 (pointer_set_lookup): Declare.
13659 (class pointer_map): New template class implementing a
13660 generic pointer to T map.
13661 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
13662 pointer_map<T>::contains, pointer_map<T>::insert,
13663 pointer_map<T>::traverse): New functions.
13664 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
13665 (pointer_set_lookup): New function.
13666 (pointer_set_contains): Use pointer_set_lookup.
13667 (pointer_set_insert): Likewise.
13668 (insert_aux): Remove.
13669 (struct pointer_map_t): Embed a pointer_set_t.
13670 (pointer_map_create): Adjust.
13671 (pointer_map_destroy): Likewise.
13672 (pointer_map_contains): Likewise.
13673 (pointer_map_insert): Likewise.
13674 (pointer_map_traverse): Likewise.
13675 * tree-streamer.h (struct streamer_tree_cache_d): Use a
13676 pointer_map<unsigned> instead of a pointer_map_t.
13677 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
13678 (streamer_tree_cache_lookup): Likewise.
13679 (streamer_tree_cache_create): Likewise.
13680 (streamer_tree_cache_delete): Likewise.
13681 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
13682 pointer_map<unsigned> instead of a pointer_map_t.
13683 (lto_init_tree_ref_encoder): Adjust.
13684 (lto_destroy_tree_ref_encoder): Likewise.
13685 * lto-section-out.c (lto_output_decl_index): Likewise.
13686 (lto_record_function_out_decl_state): Likewise.
13687 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
13689 2013-06-24 Richard Biener <rguenther@suse.de>
13691 PR tree-optimization/57488
13692 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
13694 2013-06-24 Alan Modra <amodra@gmail.com>
13696 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
13697 (gen_easy_altivec_constant): Likewise.
13698 * config/rs6000/predicates.md (easy_vector_constant_add_self,
13699 easy_vector_constant_msb): Likewise.
13701 2013-06-23 Jakub Jelinek <jakub@redhat.com>
13704 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
13705 add missing return true.
13707 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
13710 * config/sh/predicates.md (general_extend_operand): Invoke
13711 general_movsrc_operand for memory operands.
13712 (general_movsrc_operand): Allow reg+reg addressing, do not use
13713 general_operand for memory operands.
13715 2013-06-23 Sriraman Tallam <tmsriram@google.com>
13717 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
13718 when current target options does not apply.
13719 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
13720 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
13721 * config/i386/bmiintrin.h: Pass appropriate target
13722 attributes to header.
13723 * config/i386/mmintrin.h: Ditto.
13724 * config/i386/nmmintrin.h: Ditto.
13725 * config/i386/avx2intrin.h: Ditto.
13726 * config/i386/fxsrintrin.h: Ditto.
13727 * config/i386/tbmintrin.h: Ditto.
13728 * config/i386/xsaveintrin.h: Ditto.
13729 * config/i386/f16cintrin.h: Ditto.
13730 * config/i386/xtestintrin.h: Ditto.
13731 * config/i386/xsaveoptintrin.h: Ditto.
13732 * config/i386/bmi2intrin.h: Ditto.
13733 * config/i386/lzcntintrin.h: Ditto.
13734 * config/i386/smmintrin.h: Ditto.
13735 * config/i386/wmmintrin.h: Ditto.
13736 * config/i386/x86intrin.h: Remove all header include guards.
13737 * config/i386/prfchwintrin.h: Ditto.
13738 * config/i386/pmmintrin.h: Ditto.
13739 * config/i386/tmmintrin.h: Ditto.
13740 * config/i386/xmmintrin.h: Ditto.
13741 * config/i386/popcntintrin.h: Ditto.
13742 * config/i386/rdseedintrin.h: Ditto.
13743 * config/i386/ammintrin.h: Ditto.
13744 * config/i386/emmintrin.h: Ditto.
13745 * config/i386/immintrin.h: Remove all header include guards.
13746 * config/i386/fma4intrin.h: Ditto.
13747 * config/i386/lwpintrin.h: Ditto.
13748 * config/i386/xopintrin.h: Ditto.
13749 * config/i386/ia32intrin.h: Ditto.
13750 * config/i386/avxintrin.h: Ditto.
13751 * config/i386/rtmintrin.h: Ditto.
13752 * config/i386/fmaintrin.h: Ditto.
13753 * config/i386/mm3dnow.h: Ditto.
13755 2013-06-22 Sriraman Tallam <tmsriram@google.com>
13757 * common/config/i386/i386-common.c: Handle LZCNT.
13759 2013-06-22 Andi Kleen <ak@linux.intel.com>
13761 * doc/extend.texi: Use __atomic_store_n instead of
13762 __atomic_store in HLE example.
13764 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
13766 * config/sh/sh.c: Remove <cstdlib> workaround.
13768 2013-06-21 Andi Kleen <ak@linux.intel.com>
13770 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
13772 2013-06-21 Andi Kleen <ak@linux.intel.com>
13774 * doc/extend.texi: Document that __atomic_clear and
13775 __atomic_test_and_set should only be used with bool.
13777 2013-06-20 Jan Hubicka <jh@suse.cz>
13779 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
13780 types_same_for_odr.
13781 * tree.c (decls_same_for_odr): New function.
13782 (same_for_edr): New function.
13783 (types_same_for_odr): New function.
13784 (get_binfo_at_offset): Use it.
13785 * tree.h (types_same_for_odr): Declare.
13787 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
13788 Jason Merrill <jason@redhat.com>
13790 * system.h: Include <cstdlib> as well as <stdlib.h>.
13792 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
13795 * config/i386/i386.c (construct_container): Report error if
13796 long double is used with disabled x87 float returns.
13798 2013-06-20 Jan Hubicka <jh@suse.cz>
13800 * lto-cgraph.c (input_symtab): Do not set cgraph state.
13802 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
13804 PR rtl-optimization/57425
13805 PR rtl-optimization/57569
13806 * alias.c (write_dependence_p): Remove parameters mem_mode and
13807 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
13808 Changed all callers.
13809 (canon_anti_dependence): Get comments and semantics in sync.
13810 Add parameter mem_canonicalized. Changed all callers.
13811 * rtl.h (canon_anti_dependence): Update prototype.
13813 2013-06-20 Richard Biener <rguenther@suse.de>
13815 * data-streamer-in.c (streamer_read_uhwi): Optimize single
13816 byte case, inline streamer_read_uchar and defer section
13819 2013-06-20 Richard Biener <rguenther@suse.de>
13821 PR tree-optimization/57584
13822 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
13823 SSA names into the expanded expression that take part in
13824 abnormal coalescing.
13826 2013-06-19 Sharad Singhai <singhai@google.com>
13828 * gcov.c (print_usage): Handle new option.
13829 (process_args): Ditto.
13830 (get_gcov_intermediate_filename): New function.
13831 (output_intermediate_file): New function.
13832 (output_gcov_file): New function
13833 (generate_results): Handle new option.
13834 (release_function): Relase demangled name.
13835 (read_graph_file): Handle demangled name.
13836 (output_lines): Ditto.
13837 * doc/gcov.texi: Document gcov intermediate format.
13839 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
13842 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
13843 (lra_emit_add): Use the functions. Add comment about Y as an
13846 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
13849 * collect2.c (collect_atexit): New.
13850 (collect_exit): Delete.
13851 (main): Register collect_atexit with atexit.
13852 (collect_wait): Change collect_exit to exit.
13854 * collect2.h (collect_exit): Delete.
13855 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
13857 2013-06-19 Wei Mi <wmi@google.com>
13859 PR rtl-optimization/57518
13860 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
13861 if regno is used in paradoxical subreg.
13862 (update_equiv_regs): Check pdx_subregs[regno] before
13863 set a reg to be equivalent with a mem.
13865 2013-06-19 Matthias Klose <doko@ubuntu.com>
13868 * file-find.h (find_a_file): Add a mode parameter.
13869 * file-find.c (find_a_file): Likewise.
13870 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
13871 with X_OK for the executables.
13872 * collect2.c (main): Call find_a_file with X_OK.
13874 2013-06-19 Steve Ellcey <sellcey@mips.com>
13877 * config/mips/mips.md (casesi_internal_mips16_<mode>):
13878 Use NEXT_INSN instead of next_real_insn.
13880 2013-06-19 Jan Hubicka <jh@suse.cz>
13882 * cgraph.h (const_value_known_p): Replace by ...
13883 (ctor_for_folding): .. this one.
13884 * cgraphunit.c (process_function_and_variable_attributes): Use it.
13885 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
13886 * expr.c (expand_expr_real_1): Likewise.
13887 (string_constant): Likewise.
13888 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
13889 * ipa.c (process_references): Likewise.
13890 (symtab_remove_unreachable_nodes): Likewise.
13891 * ipa-inline-analysis.c (param_change_prob): Likewise.
13892 * gimple-fold.c (canonicalize_constructor_val): Likewise.
13893 (get_base_constructor): Likwise.
13894 * varpool.c (varpool_remove_node): Likewise.
13895 (varpool_remove_initializer): LIkewise.
13896 (dump_varpool_node): LIkwise.
13897 (const_value_known_p): Rewrite to ...
13898 (ctor_for_folding): ... this one.
13900 2013-06-19 Jakub Jelinek <jakub@redhat.com>
13903 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
13904 PERSONALITY in $PATH derived prefixes.
13906 2013-06-19 Jeff Law <law@redhat.com>
13908 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
13911 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
13912 (simplify_bitwise_binary): Use it to simpify certain binary ops on
13915 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
13917 * config/arm/vfp.md: Move VFP instruction classification documentation
13919 * config/arm/arm.md: ... here. Update instruction classification
13922 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
13924 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
13925 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
13926 pattern. Use more efficient sequences on ARMv5 and Thumb2.
13928 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
13931 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13932 with NEXT_INSN. Use tablejump_p to check for jump table data
13935 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
13938 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
13939 that now in C++ the value is correct per the C++ standards.
13941 2013-06-19 Richard Biener <rguenther@suse.de>
13943 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
13944 for global context.
13946 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13949 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13952 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13953 with next_active_insn.
13955 2013-06-18 Sriraman Tallam <tmsriram@google.com>
13957 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
13958 functions are inlined during failures to flag an error.
13959 * tree-inline.c (expand_call_inline): Allow the error to be flagged
13960 in early inline pass.
13962 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
13964 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
13967 2013-06-18 Julian Brown <julian@codesourcery.com>
13969 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
13970 Permit virtual register pre-reload if !strict.
13971 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
13973 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
13975 * config/arm/neon.md (movmisalign<mode>): Use
13976 neon_perm_struct_or_reg_operand instead of
13977 neon_struct_or_register_operand.
13978 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
13979 neon_permissive_struct_operand instead of neon_struct_operand.
13980 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
13981 neon_vector_mem_operand.
13982 * config/arm/predicates.md (neon_struct_operand): Adjust call to
13983 neon_vector_mem_operand.
13984 (neon_permissive_struct_operand): New.
13985 (neon_struct_or_register_operand): Rename to...
13986 (neon_perm_struct_or_reg_operand): This. Adjust call to
13987 neon_vector_mem_operand.
13989 2013-06-18 Richard Biener <rguenther@suse.de>
13991 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
13992 * lto-streamer.h: Include pointer-set.h.
13993 (struct lto_decl_slot): Remove.
13994 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
13995 Remove next_index entry.
13996 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
13997 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
13998 (lto_init_tree_ref_encoder): Adjust.
13999 (lto_destroy_tree_ref_encoder): Likewise.
14000 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14001 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14002 (lto_output_decl_index): Adjust.
14003 (lto_new_out_decl_state): Likewise.
14004 (lto_record_function_out_decl_state): Likewise.
14005 * lto-streamer-out.c (copy_function): Likewise.
14007 2013-06-18 Richard Biener <rguenther@suse.de>
14009 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
14010 * cgraphunit.c: Include cfgloop.h.
14011 (init_lowered_empty_function): Initialize the loop tree.
14012 (assemble_thunk): Insert new BBs into loops.
14014 2013-06-18 Richard Biener <rguenther@suse.de>
14016 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
14017 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
14018 the map from cache entry to cache index optional.
14019 (streamer_tree_cache_replace_tree): Adjust accordingly.
14020 (streamer_tree_cache_append): Likewise.
14021 (streamer_tree_cache_delete): Likewise.
14022 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
14023 streamer cache map from cache entry to cache index.
14024 * lto-streamer-out.c (create_output_block): Adjust.
14026 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
14028 * config/arm/arm.md (attribute "insn"): Move multiplication and
14029 division attributes to...
14030 (attribute "type"): ... here. Remove mult.
14031 (attribute "mul32"): New attribute.
14032 (attribute "mul64"): Add umaal.
14033 (*arm_mulsi3): Update attributes.
14034 (*arm_mulsi3_v6): Likewise.
14035 (*thumb_mulsi3): Likewise.
14036 (*thumb_mulsi3_v6): Likewise.
14037 (*mulsi3_compare0): Likewise.
14038 (*mulsi3_compare0_v6): Likewise.
14039 (*mulsi_compare0_scratch): Likewise.
14040 (*mulsi_compare0_scratch_v6): Likewise.
14041 (*mulsi3addsi): Likewise.
14042 (*mulsi3addsi_v6): Likewise.
14043 (*mulsi3addsi_compare0): Likewise.
14044 (*mulsi3addsi_compare0_v6): Likewise.
14045 (*mulsi3addsi_compare0_scratch): Likewise.
14046 (*mulsi3addsi_compare0_scratch_v6): Likewise.
14047 (*mulsi3subsi): Likewise.
14048 (*mulsidi3adddi): Likewise.
14049 (*mulsi3addsi_v6): Likewise.
14050 (*mulsidi3adddi_v6): Likewise.
14051 (*mulsidi3_nov6): Likewise.
14052 (*mulsidi3_v6): Likewise.
14053 (*umulsidi3_nov6): Likewise.
14054 (*umulsidi3_v6): Likewise.
14055 (*umulsidi3adddi): Likewise.
14056 (*umulsidi3adddi_v6): Likewise.
14057 (*smulsi3_highpart_nov6): Likewise.
14058 (*smulsi3_highpart_v6): Likewise.
14059 (*umulsi3_highpart_nov6): Likewise.
14060 (*umulsi3_highpart_v6): Likewise.
14061 (mulhisi3): Likewise.
14062 (*mulhisi3tb): Likewise.
14063 (*mulhisi3bt): Likewise.
14064 (*mulhisi3tt): Likewise.
14065 (maddhisi4): Likewise.
14066 (*maddhisi4tb): Likewise.
14067 (*maddhisi4tt): Likewise.
14068 (maddhidi4): Likewise.
14069 (*maddhidi4tb): Likewise.
14070 (*maddhidi4tt): Likewise.
14071 (divsi3): Likewise.
14072 (udivsi3): Likewise.
14073 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
14074 (thumb2_mulsi_short_compare0): Likewise.
14075 (thumb2_mulsi_short_compare0_scratch): Likewise.
14076 * config/arm/arm1020e.md (1020mult1): Update attribute change.
14077 (1020mult2): Likewise.
14078 (1020mult3): Likewise.
14079 (1020mult4): Likewise.
14080 (1020mult5): Likewise.
14081 (1020mult6): Likewise.
14082 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
14084 (cortex_a15_mult64): Likewise.
14085 (cortex_a15_sdiv): Likewise.
14086 (cortex_a15_udiv): Likewise.
14087 * config/arm/arm1026ejs.md (mult1): Update attribute change.
14093 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
14094 (pj4_ir_div): Likewise.
14095 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
14096 (11_mult2): Likewise.
14097 (11_mult3): Likewise.
14098 (11_mult4): Likewise.
14099 (11_mult5): Likewise.
14100 (11_mult6): Likewise.
14101 (11_mult7): Likewise.
14102 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
14103 (cortex_a8_mla): Likewise.
14104 (cortex_a8_mull): Likewise.
14105 (cortex_a8_smulwy): Likewise.
14106 (cortex_a8_smlald): Likewise.
14107 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
14108 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
14109 (cortex_r4_mul_3): Likewise.
14110 (cortex_r4_mla_4): Likewise.
14111 (cortex_r4_mla_3): Likewise.
14112 (cortex_r4_smlald): Likewise.
14113 (cortex_r4_mull): Likewise.
14114 (cortex_r4_sdiv): Likewise.
14115 (cortex_r4_udiv): Likewise.
14116 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
14117 (cortex_a7_idiv): Likewise.
14118 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
14119 (9_mult2): Likewise.
14120 (9_mult3): Likewise.
14121 (9_mult4): Likewise.
14122 (9_mult5): Likewise.
14123 (9_mult6): Likewise.
14124 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
14125 (cortex_a53_sdiv): Likewise.
14126 (cortex_a53_udiv): Likewise.
14127 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
14128 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
14129 (mp626_mult2): Likewise.
14130 (mp626_mult3): Likewise.
14131 (mp626_mult4): Likewise.
14132 * config/arm/fa526.md (526_mult1): Update attribute change.
14133 (526_mult2): Likewise.
14134 * config/arm/arm-generic.md (mult): Update attribute change.
14135 (mult_ldsched_strongarm): Likewise.
14136 (mult_ldsched): Likewise.
14137 (multi_cycle): Likewise.
14138 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
14139 * config/arm/fa606te.md (606te_mult1): Update attribute change.
14140 (606te_mult2): Likewise.
14141 (606te_mult3): Likewise.
14142 (606te_mult4): Likewise.
14143 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
14144 (cortex_a9_mac16): Likewise.
14145 (cortex_a9_multiply): Likewise.
14146 (cortex_a9_mac): Likewise.
14147 (cortex_a9_multiply_long): Likewise.
14148 * config/arm/fa626te.md (626te_mult1): Update attribute change.
14149 (626te_mult2): Likewise.
14150 (626te_mult3): Likewise.
14151 (626te_mult4): Likewise.
14153 2013-06-18 Richard Biener <rguenther@suse.de>
14156 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
14158 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14161 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14162 with next_active_insn.
14164 2013-06-18 Alan Modra <amodra@gmail.com>
14166 * config/rs6000/rs6000.h (enum data_align): New.
14167 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
14168 (DATA_ABI_ALIGNMENT): Define.
14169 (CONSTANT_ALIGNMENT): Correct comment.
14170 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
14171 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
14173 2013-06-17 David Malcolm <dmalcolm@redhat.com>
14175 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
14176 ATTRIBUTE_UNUSED marking.
14178 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
14180 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
14181 alternative and update.
14182 (aarch64_dup_lanedi): Delete.
14183 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
14184 * config/aarch64/aarch64-simd-builtins.def: Update.
14186 2013-06-17 Richard Biener <rguenther@suse.de>
14188 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
14189 (lto_input_scc): Declare.
14190 (lto_input_tree_1): Likewise.
14191 (struct lto_stats_d): Add num_tree_bodies_output and
14192 num_pickle_refs_output.
14193 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
14194 (lto_read_tree_1): Split out from ...
14195 (lto_read_tree): ... this.
14196 (lto_input_scc): New function.
14197 (lto_input_tree_1): Split out from ...
14198 (lto_input_tree): ... this. Handle LTO_tree_scc.
14199 (lto_data_in_create): Create the streamer cache without hashes.
14200 * lto-streamer-out.c (create_output_block): Create the streamer
14201 cache with hashes when not doing WPA.
14202 (lto_write_tree_1): Split out from ...
14203 (lto_write_tree): ... this.
14204 (get_symbol_initial_value): New function.
14205 (lto_output_tree_1): Split out from ...
14206 (lto_output_tree): ... this. Write trees as series of SCCs
14207 using a DFS walk via DFS_write_tree.
14208 (struct sccs, struct scc_entry): New types.
14209 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
14210 (DFS_write_tree_body): New function.
14211 (DFS_write_tree): Likewise.
14212 (hash_tree): Likewise.
14213 (scc_entry_compare): Likewise.
14214 (hash_scc): Likewise.
14215 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
14216 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
14217 TREE_CHAIN as regular reference.
14218 (streamer_read_integer_cst): Remove.
14219 (streamer_get_pickled_tree): Adjust.
14220 * tree-streamer-out.c (streamer_write_chain): Disable streaming
14221 of DECL_EXTERNALs in BLOCK_VARS for now.
14222 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
14224 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
14225 Add hash value argument and record that if hashes are recorded
14227 (streamer_tree_cache_insert_1): Adjust.
14228 (streamer_tree_cache_insert): Likewise.
14229 (streamer_tree_cache_insert_at): Rename to ...
14230 (streamer_tree_cache_replace_tree): ... this and adjust.
14231 (streamer_tree_cache_append): Adjust.
14232 (record_common_node): Likewise.
14233 (streamer_tree_cache_create): Add argument whether to
14234 record hash values together with trees.
14235 (streamer_tree_cache_delete): Adjust.
14236 * tree-streamer.h (struct streamer_tree_cache_d): Add
14238 (streamer_read_integer_cst): Remove.
14239 (streamer_tree_cache_insert): Adjust.
14240 (streamer_tree_cache_append): Likewise.
14241 (streamer_tree_cache_insert_at): Rename to ...
14242 (streamer_tree_cache_replace_tree): ... this and adjust.
14243 (streamer_tree_cache_create): Add argument whether to record hashes.
14244 (streamer_tree_cache_get): Rename to ...
14245 (streamer_tree_cache_get_tree): ... this.
14246 (streamer_tree_cache_get_hash): New function.
14247 * tree.c (cache_integer_cst): New function.
14248 * tree.h (cache_integer_cst): Declare.
14249 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
14250 * lto-symtab.c (lto_varpool_replace_node): Only release
14251 DECL_INITIAL of non-prevailing decls.
14252 * varpool.c (varpool_remove_initializer): Do not release
14253 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
14255 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14257 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
14258 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
14259 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
14260 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
14261 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
14262 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
14263 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
14264 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
14265 instead of TARGET_64BIT.
14266 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
14267 Require ISA_HAS_<D>DIV.
14269 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
14271 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
14272 (mips*-*-linux*): Move default with_llsc setting to where other
14274 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
14276 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
14277 Likewise. Remove default with_tune setting. Move default float
14278 setting to its own block. Handle with_llsc in the same block as above.
14280 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
14282 PR rtl-optimization/57425
14283 PR rtl-optimization/57569
14284 * alias.c (write_dependence_p): Add new parameters mem_mode,
14285 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
14286 Changed all callers.
14287 (canon_anti_dependence): New function.
14288 * cse.c (check_dependence): Use canon_anti_dependence.
14289 * cselib.c (cselib_invalidate_mem): Likewise.
14290 * rtl.h (canon_anti_dependence): Declare.
14292 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14294 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
14295 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
14296 ".set mips3" for 64-bit targets.
14298 2013-06-15 Dehao Chen <dehao@google.com>
14300 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
14301 * gimple-low.c (gimple_check_call_matching_types): Likewise.
14302 (gimple_check_call_args): Likewise.
14303 * value-prof.c (check_ic_target): Likewise.
14304 * ipa-inline.c (early_inliner): Likewise.
14305 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
14306 * cgraph.c (cgraph_create_edge_1): Likewise.
14307 (cgraph_make_edge_direct): Likewise.
14309 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
14312 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
14313 rs6000_output_move_128bit to handle emitting quad memory
14314 operations. Set attribute length to 8 bytes.
14316 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
14318 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
14320 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
14321 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
14322 (aarch64_<su>mlsl<mode>): Likewise.
14324 2013-06-14 Mike Stump <mikestump@comcast.net>
14326 * Makefile.in (TARGET_H): Add insn-codes.h.
14328 2013-06-14 Alan Modra <amodra@gmail.com>
14330 PR middle-end/57134
14331 PR middle-end/57586
14332 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
14333 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
14334 bitfield expansion when EXPAND_MEMORY.
14335 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
14337 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
14339 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
14340 test for clearing quad memory on 32-bit later.
14342 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14344 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
14345 (fold_negate_expr): Likewise.
14346 (fold_real_zero_addition_p): Handle vectors.
14347 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
14349 2013-06-14 Alan Modra <amodra@gmail.com>
14351 * varasm.c (force_const_mem): Revert 2013-06-07 change.
14353 2013-06-13 Jan Hubicka <jh@suse.cz>
14355 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
14356 Local comdats are not externally visible.
14357 * symtab.c (dump_symtab_base): Dump externally visible.
14358 (verify_symtab_base): Verify back links in the symtab hash.
14360 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14362 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
14363 CONVERT_EXPR as equal nodes.
14365 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14367 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
14369 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14371 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
14372 Generalize to complex and vector.
14373 * tree.c (build_all_ones_cst): New function.
14374 * tree.h (build_all_ones_cst): Declare it.
14376 2013-06-13 Alan Modra <amodra@gmail.com>
14378 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
14379 * config/rs6000/rs6000.md (signbittf2): New insn.
14380 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
14381 (abstf2_internal, cmptf_internal2): Likewise.
14382 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
14384 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
14385 Pat Haugen <pthaugen@us.ibm.com>
14386 Peter Bergner <bergner@vnet.ibm.com>
14388 * config/rs6000/rs6000.c (emit_load_locked): Add support for
14389 power8 byte, half-word, and quad-word atomic instructions.
14390 (emit_store_conditional): Likewise.
14391 (rs6000_expand_atomic_compare_and_swap): Likewise.
14392 (rs6000_expand_atomic_op): Likewise.
14394 * config/rs6000/sync.md (larx): Add new modes for power8.
14396 (AINT): New mode iterator to include TImode as well as normal
14397 integer modes on power8.
14398 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
14399 that VSX registers are not considered. Use AINT mode iterator
14400 instead of INT1 to allow inclusion of quad word atomic operations
14402 (load_locked<mode>): Likewise.
14403 (store_conditional<mode>): Likewise.
14404 (atomic_compare_and_swap<mode>): Likewise.
14405 (atomic_exchange<mode>): Likewise.
14406 (atomic_nand<mode>): Likewise.
14407 (atomic_fetch_<fetchop_name><mode>): Likewise.
14408 (atomic_nand_fetch<mode>): Likewise.
14409 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
14411 (ATOMIC): On power8, add QImode, HImode modes.
14412 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
14413 modes that promote to SImode.
14414 (load_lockedti): Convert TImode arguments to PTImode, so that we
14415 get a guaranteed even/odd register pair.
14416 (load_lockedpti): Likewise.
14417 (store_conditionalti): Likewise.
14418 (store_conditionalpti): Likewise.
14420 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
14421 atomic load/store instructions.
14424 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
14426 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
14428 (insn_count): New attribute, with most cases extracted from...
14429 (length): ...here. Redefine most cases in terms of insn_count.
14430 (single_insn): Delete.
14431 (can_delay): Use insn_count to check for single instructions.
14432 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
14433 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
14434 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
14435 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
14436 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
14437 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
14438 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
14439 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
14440 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
14441 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
14442 rather than "length".
14443 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
14444 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
14445 Use "insn_count" rather than "length".
14446 * config/mips/mips-dsp.md
14447 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
14448 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
14451 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
14453 PR tree-optimization/57361
14454 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
14456 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
14458 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
14460 (aarch64_simd_combine<mode>): New instruction expansion.
14461 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
14462 function prototype.
14463 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
14464 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
14466 2013-06-12 Jan Hubicka <jh@suse.cz>
14468 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
14469 decl has when in streaming stage.
14470 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
14471 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
14473 2013-06-12 Roland Stigge <stigge@antcom.de>
14476 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
14478 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14480 PR tree-optimization/57537
14481 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
14482 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
14484 2013-06-12 Richard Biener <rguenther@suse.de>
14486 * data-streamer.h (streamer_write_char_stream): CSE
14487 obs->current_pointer.
14488 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
14489 streamer_write_char_stream manually and optimize the resulting loop.
14490 (streamer_write_hwi_stream): Likewise.
14492 2013-06-12 Jan Hubicka <jh@suse.cz>
14494 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
14495 * cgraph.h (varpool_create_empty_node): Declare.
14496 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
14498 * symtab.c (symtab_unregister_node): Be lax about missin entries
14500 (symtab_get_node): Update comment.
14501 * varpool.c (varpool_create_empty_node): Break out from ...
14502 (varpool_node_for_decl): ... here.
14503 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
14505 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
14507 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
14508 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
14509 part. Use straight-line flow at the end.
14510 <COMPONENT_REF>: Remove superfluous else.
14511 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
14513 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14516 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
14517 target hook even for !TREE_PUBLIC decls. If no resolution info
14518 is available, return false for common and external decls.
14520 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14522 * config/rl78/constraints.md (U): New constraint.
14523 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
14526 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
14529 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
14530 to allow returning address to AT_PLATFORM name.
14532 2013-06-11 Jan Hubicka <jh@suse.cz>
14534 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
14535 * cgraph.h (symtab_node_base): Add weakref flag.
14536 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
14537 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
14538 (output_weakrefs): Use weakref flag.
14539 * fold-const.c (simple_operand_p): Handle WEAK.
14540 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
14541 * ipa.c (varpool_externally_visible_p): Drop weakref.
14542 (function_and_variable_visibility): Update comment; fix weakref
14543 sanity checks; do not clear DECL_WEAK on them.
14544 * lto-cgraph.c (lto_output_node): update.
14545 (lto_output_varpool_node): Update.
14546 (input_overwrite_node): Update.
14547 (input_node): Update.
14548 (input_varpool_node): Update.
14549 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
14550 (lto_symtab_merge_symbols): Add sanity check.
14551 (lto_symtab_prevailing_decl): Do not special case weakrefs.
14552 * passes.c (rest_of_decl_compilation): Set static flag, too.
14553 * symtab.c (dump_symtab_base): Dump weakref.
14554 (verify_symtab_base): Sanity check weakrefs.
14555 (symtab_make_decl_local): Remove duplicated code.
14556 (symtab_alias_ultimate_target): Simplify.
14557 * varpool.c (varpool_create_variable_alias): Set weakref flag.
14559 2013-06-11 Tom de Vries <tom@codesourcery.com>
14561 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
14562 sequence_vect == NULL.
14564 2013-06-11 DJ Delorie <dj@redhat.com>
14566 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
14567 (rl78_unwind_word_mode): New.
14569 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14571 * final.c (debug_prefix_maps): Make static.
14573 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14575 * function.c (initial_trampoline): Remove stray copy.
14577 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
14579 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
14581 2013-06-11 Martin Jambor <mjambor@suse.cz>
14583 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
14584 within bounds at the beginning of the function.
14586 2013-06-11 Alan Modra <amodra@gmail.com>
14588 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
14590 (get_named_section): Don't NULL !DECL_P decl.
14592 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
14594 * doc/invoke.texi (core-avx2): Document.
14596 (atom): Updated with MOVBE.
14598 2013-06-11 Richard Biener <rguenther@suse.de>
14600 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
14602 2013-06-11 Anton Blanchard <anton@samba.org>
14604 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
14605 correct shift value in little-endian mode.
14607 2013-06-11 Jakub Jelinek <jakub@redhat.com>
14610 * varasm.c (get_variable_align): Move #endif to the right place.
14612 2013-06-10 Cary Coutant <ccoutant@google.com>
14614 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
14615 for hash so that hash table traversal order is deterministic.
14617 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
14618 Pat Haugen <pthaugen@us.ibm.com>
14619 Peter Bergner <bergner@vnet.ibm.com>
14621 * config/rs6000/vector.md (GPR move splitter): Do not split moves
14622 of vectors in GPRS if they are direct moves or quad word load or
14625 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
14627 (direct_move_p): Likewise.
14628 (quad_load_store_p): Likewise.
14630 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
14631 classes into bins based on the physical register type.
14632 (reg_class_to_reg_type): Likewise.
14633 (IS_STD_REG_TYPE): Likewise.
14634 (IS_FP_VECT_REG_TYPE): Likewise.
14635 (reload_fpr_gpr): Arrays to determine what insn to use if we can
14636 use direct move instructions.
14637 (reload_gpr_vsx): Likewise.
14638 (reload_vsx_gpr): Likewise.
14639 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
14640 information that is a simplification of register classes. Also
14641 precalculate direct move reload helpers.
14642 (direct_move_p): New function to return true if the operation can
14643 be done as a direct move instruciton.
14644 (quad_load_store_p): New function to return true if the operation
14645 is a quad memory operation.
14646 (rs6000_legitimize_address): If quad memory, only allow register
14647 indirect for TImode addresses.
14648 (rs6000_legitimate_address_p): Likewise.
14649 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
14650 (rs6000_reload_register_type): Likewise.
14651 (register_to_reg_type): Return register type.
14652 (rs6000_secondary_reload_simple_move): New helper function for
14653 secondary reload and secondary memory needed to identify anything
14654 that is a simple move, and does not need reloading.
14655 (rs6000_secondary_reload_direct_move): New helper function for
14656 secondary reload to identify cases that can be done with several
14657 instructions via the direct move instructions.
14658 (rs6000_secondary_reload_move): New helper function for secondary
14659 reload to identify moves between register types that can be done.
14660 (rs6000_secondary_reload): Add support for quad memory operations
14661 and for direct move.
14662 (rs6000_secondary_memory_needed): Likewise.
14663 (rs6000_debug_secondary_memory_needed): Change argument names.
14664 (rs6000_output_move_128bit): New function to return the move to
14665 use for 128-bit moves, including knowing about the various
14666 limitations of quad memory operations.
14668 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
14669 memory operations. call rs6000_output_move_128bit for the actual
14670 instruciton(s) to generate.
14671 (vsx_movti_64bit): Likewise.
14673 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
14674 (UNSPEC_P8V_MTVSRWZ): Likewise.
14675 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
14676 (UNSPEC_P8V_MTVSRD): Likewise.
14677 (UNSPEC_P8V_XXPERMDI): Likewise.
14678 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
14679 (UNSPEC_FUSION_GPR): Likewise.
14680 (FMOVE128_GPR): New iterator for direct move.
14681 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
14682 (f32_sv): Likewise.
14683 (f32_dm): Likewise.
14684 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
14685 loads and direct move instructions.
14686 (zero_extendsidi2_lfiwzx): Likewise.
14687 (extendsidi2_lfiwax): Likewise.
14688 (extendsidi2_nocell): Likewise.
14689 (floatsi<mode>2_lfiwax): Likewise.
14690 (lfiwax): Likewise.
14691 (floatunssi<mode>2_lfiwzx): Likewise.
14692 (lfiwzx): Likewise.
14693 (fix_trunc<mode>_stfiwx): Likewise.
14694 (fixuns_trunc<mode>_stfiwx): Likewise.
14695 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
14696 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
14697 (parity<mode>2_cmpb): Set length/type attr.
14698 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
14699 for 'mr.' to fast_compare.
14700 (bpermd_<mode>): Change type attr to popcnt.
14701 (p8_fmrgow_<mode>): New insns for power8 direct move support.
14702 (p8_mtvsrwz_1): Likewise.
14703 (p8_mtvsrwz_2): Likewise.
14704 (reload_fpr_from_gpr<mode>): Likewise.
14705 (p8_mtvsrd_1): Likewise.
14706 (p8_mtvsrd_2): Likewise.
14707 (p8_xxpermdi_<mode>): Likewise.
14708 (reload_vsx_from_gpr<mode>): Likewise.
14709 (reload_vsx_from_gprsf): Likewise.
14710 (p8_mfvsrd_3_<mode>): LIkewise.
14711 (reload_gpr_from_vsx<mode>): Likewise.
14712 (reload_gpr_from_vsxsf): Likewise.
14713 (p8_mfvsrd_4_disf): Likewise.
14714 (multi-word GPR splits): Do not split direct moves or quad memory
14717 2013-06-10 David Malcolm <dmalcolm@redhat.com>
14719 * tree-into-ssa.c (interesting_blocks): Make static.
14721 2013-06-10 Jakub Jelinek <jakub@redhat.com>
14724 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
14725 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
14726 Use DATA_ABI_ALIGNMENT for that case instead if defined.
14727 (get_variable_align): New function.
14728 (get_variable_section, emit_bss, emit_common,
14729 assemble_variable_contents, place_block_symbol): Use
14730 get_variable_align instead of DECL_ALIGN.
14731 (assemble_noswitch_variable): Add align argument, use it
14732 instead of DECL_ALIGN.
14733 (assemble_variable): Adjust caller. Use get_variable_align
14734 instead of DECL_ALIGN.
14735 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
14737 (DATA_ABI_ALIGNMENT): Define.
14738 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
14739 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
14740 opt is false, only return the psABI mandated alignment increase.
14741 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
14742 (DATA_ABI_ALIGNMENT): ... this.
14743 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
14744 (DATA_ABI_ALIGNMENT): ... this.
14745 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
14746 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
14747 (DATA_ABI_ALIGNMENT): ... this.
14748 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
14749 * doc/tm.texi: Regenerated.
14751 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
14753 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
14754 cmp_code to construct REG_EQUAL note.
14756 2013-06-09 Jakub Jelinek <jakub@redhat.com>
14759 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
14760 that operands[2] doesn't overlap with operands[0].
14762 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
14763 Jan Hubicka <jh@suse.cz>
14765 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
14766 hack to mark symbols as used.
14768 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
14770 PR rtl-optimization/57559
14771 * lra-constraints.c (process_alt_operands): Don't discourage
14772 memory with known offset for offsetable memory constraint.
14773 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
14775 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
14777 * varasm.c (struct oc_local_state): Reorder fields.
14778 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
14779 and adjust accordingly.
14780 (output_constructor): Reorder initialization code and adjust call to
14781 output_constructor_bitfield.
14783 2013-06-07 Jan Hubicka <jh@suse.cz>
14785 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
14787 2013-06-07 David Malcolm <dmalcolm@redhat.com>
14789 * tree-object-size.c (unknown): Make const.
14791 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14793 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
14794 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
14795 for last alternative in the cpu_facility attribute.
14797 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14800 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
14801 (xordi3): Change operand 2 constraint to arm_xordi_operand.
14802 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
14803 * config/arm/constraints.md (Dg): New constraint.
14804 * config/arm/neon.md (xordi3_neon): Remove.
14805 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
14806 * config/arm/predicates.md (arm_xordi_operand): New predicate.
14808 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14810 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
14811 Clean up alternatives.
14813 2013-06-07 Alan Modra <amodra@gmail.com>
14815 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
14818 2013-06-07 Alan Modra <amodra@gmail.com>
14820 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
14822 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14824 * config/arm/constraints.md (Df): New constraint.
14825 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
14826 Correct length attribute for last two alternatives.
14828 2013-06-07 Alan Modra <amodra@gmail.com>
14830 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
14831 override user -mfp-in-toc.
14832 (offsettable_ok_by_alignment): Consider just the current access
14833 rather than the whole object, unless BLKmode. Handle
14834 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
14835 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
14836 for -mcmodel=medium.
14837 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
14838 override user -mfp-in-toc or -msum-in-toc. Default to
14839 -mno-fp-in-toc for -mcmodel=medium.
14841 2013-06-06 DJ Delorie <dj@redhat.com>
14843 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
14844 TARGET_VALID_POINTER_MODE.
14846 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
14847 Pat Haugen <pthaugen@us.ibm.com>
14848 Peter Bergner <bergner@vnet.ibm.com>
14850 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
14851 Document new power8 builtins.
14853 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
14854 condition code register, to allow 128-bit logical operations to be
14855 done in the VSX or GPR registers.
14856 (nor<mode>3): Use the canonical form for nor.
14857 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
14858 vclz*, and vpopcnt* vector instructions.
14859 (nand<mode>3): Likewise.
14860 (orc<mode>3): Likewise.
14861 (clz<mode>2): LIkewise.
14862 (popcount<mode>2): Likewise.
14864 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
14865 that only the GPRs are recognized.
14867 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
14868 support for new power8 builtins.
14870 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
14872 (xscvdpspn): Likewise.
14878 (vpopcnt): Likewise.
14879 (vpopcntb): Likewise.
14880 (vpopcnth): Likewise.
14881 (vpopcntw): Likewise.
14882 (vpopcntd): Likewise.
14884 (vmrgew): Likewise.
14885 (vmrgow): Likewise.
14887 (eqv_v16qi3): Likewise.
14888 (eqv_v8hi3): Likewise.
14889 (eqv_v4si3): Likewise.
14890 (eqv_v2di3): Likewise.
14891 (eqv_v4sf3): Likewise.
14892 (eqv_v2df3): Likewise.
14894 (nand_v16qi3): Likewise.
14895 (nand_v8hi3): Likewise.
14896 (nand_v4si3): Likewise.
14897 (nand_v2di3): Likewise.
14898 (nand_v4sf3): Likewise.
14899 (nand_v2df3): Likewise.
14901 (orc_v16qi3): Likewise.
14902 (orc_v8hi3): Likewise.
14903 (orc_v4si3): Likewise.
14904 (orc_v2di3): Likewise.
14905 (orc_v4sf3): Likewise.
14906 (orc_v2df3): Likewise.
14908 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
14909 allow power8 quad mode in 64-bit.
14910 (rs6000_builtin_vectorized_function): Add support to vectorize
14911 ISA 2.07 count leading zeros, population count builtins.
14912 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
14913 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
14914 (builtin_function_type): Add vgbbd builtin function which takes an
14916 (altivec_expand_vec_perm_const): Add support for new power8 merge
14919 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
14920 that does not include TImdoe for use with 32-bit.
14921 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
14923 (UNSPEC_VSX_CVDPSPN): Likewise.
14924 (vsx_xscvdpspn): Likewise.
14925 (vsx_xscvspdpn): Likewise.
14926 (vsx_xscvdpspn_scalar): Likewise.
14927 (vsx_xscvspdpn_directmove): Likewise.
14928 (vsx_and<mode>3): Split logical operations into 32-bit and
14929 64-bit. Add support to do logical operations on TImode as well as
14930 VSX vector types. Allow logical operations to be done in either
14931 VSX registers or in general purpose registers in 64-bit mode. Add
14932 splitters if GPRs were used. For AND, add clobber of CCmode to
14933 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
14935 (vsx_and<mode>3_32bit): Likewise.
14936 (vsx_and<mode>3_64bit): Likewise.
14937 (vsx_ior<mode>3): Likewise.
14938 (vsx_ior<mode>3_32bit): Likewise.
14939 (vsx_ior<mode>3_64bit): Likewise.
14940 (vsx_xor<mode>3): Likewise.
14941 (vsx_xor<mode>3_32bit): Likewise.
14942 (vsx_xor<mode>3_64bit): Likewise.
14943 (vsx_one_cmpl<mode>2): Likewise.
14944 (vsx_one_cmpl<mode>2_32bit): Likewise.
14945 (vsx_one_cmpl<mode>2_64bit): Likewise.
14946 (vsx_nor<mode>3): Likewise.
14947 (vsx_nor<mode>3_32bit): Likewise.
14948 (vsx_nor<mode>3_64bit): Likewise.
14949 (vsx_andc<mode>3): Likewise.
14950 (vsx_andc<mode>3_32bit): Likewise.
14951 (vsx_andc<mode>3_64bit): Likewise.
14952 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
14953 and xxlorc instructions.
14954 (vsx_eqv<mode>3_64bit): Likewise.
14955 (vsx_nand<mode>3_32bit): Likewise.
14956 (vsx_nand<mode>3_64bit): Likewise.
14957 (vsx_orc<mode>3_32bit): Likewise.
14958 (vsx_orc<mode>3_64bit): Likewise.
14960 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
14962 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
14964 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
14965 (p8_vmrgow): Likewise.
14966 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
14967 GPRs to be split under VSX.
14968 (p8v_clz<mode>2): Add power8 count leading zero support.
14969 (p8v_popcount<mode>2): Add power8 population count support.
14970 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
14973 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
14976 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
14978 (vec_nand): Likewise.
14979 (vec_vclz): Likewise.
14980 (vec_vclzb): Likewise.
14981 (vec_vclzd): Likewise.
14982 (vec_vclzh): Likewise.
14983 (vec_vclzw): Likewise.
14984 (vec_vgbbd): Likewise.
14985 (vec_vmrgew): Likewise.
14986 (vec_vmrgow): Likewise.
14987 (vec_vpopcnt): Likewise.
14988 (vec_vpopcntb): Likewise.
14989 (vec_vpopcntd): Likewise.
14990 (vec_vpopcnth): Likewise.
14991 (vec_vpopcntw): Likewise.
14993 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14995 PR rtl-optimization/57468
14996 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
14999 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15001 PR rtl-optimization/57459
15002 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
15003 type when setting live regs.
15005 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15007 * config/s390/s390.opt (mlra): New option.
15008 * config/s390/s390.c (s390_decompose_address): Check displacement
15009 for all registers for LRA.
15010 (s390_secondary_reload): Don't used secondary reloads for LRA.
15011 (s390_lra_p): New function.
15012 (TARGET_LRA_P): Define.
15013 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
15014 of attribute cpu_facility to zarch for the last alternative.
15015 (*cmpmem_short): Ditto.
15017 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
15019 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
15020 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
15021 (arm_expand_prologue): Likewise.
15023 2013-06-06 Teresa Johnson <tejohnson@google.com>
15026 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
15027 as this is now done by redirect_edge_and_branch_force.
15028 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
15029 barriers, and fix interaction with splitting.
15030 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
15031 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
15032 reflect changes made in the routine.
15033 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
15034 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
15035 since this is called in cfglayout mode, and replace partition fixup
15036 with assert as that is now done by force_nonfallthru_and_redirect.
15037 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
15038 already be marked with region crossing note.
15039 (insert_section_boundary_note): Make non-static, gate on flag
15040 has_bb_partition, rewrite to also check for multiple partitions.
15041 (rest_of_handle_reorder_blocks): Remove call to
15042 insert_section_boundary_note, now done later during free_cfg.
15043 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
15044 * bb-reorder.h (insert_section_boundary_note): Declare.
15045 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
15046 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
15047 invoke insert_section_boundary_note.
15048 (try_redirect_by_replacing_jump): Remove unnecessary
15049 check for region crossing note.
15050 (fixup_partition_crossing): New function.
15051 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
15052 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
15053 in non-cfglayout mode.
15054 (force_nonfallthru_and_redirect): Fixup partition boundaries,
15055 remove old code that tried to do this. Emit barrier correctly
15056 when we are in cfglayout mode.
15057 (last_bb_in_partition): New function.
15058 (rtl_split_edge): Correctly fixup partition boundaries.
15059 (commit_one_edge_insertion): Remove old code that tried to
15060 fixup region crossing edge since this is now handled in
15061 split_block, and set up insertion point correctly since
15062 block may now end in a jump.
15063 (verify_hot_cold_block_grouping): Guard against checking when not in
15064 linearized RTL mode.
15065 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
15067 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
15068 rtl_verify_flow_info, so not called in cfglayout mode.
15069 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
15070 (fixup_reorder_chain): Remove old code that attempted to fixup region
15071 crossing note as this is now handled in force_nonfallthru_and_redirect.
15072 (duplicate_insn_chain): Don't duplicate switch section notes.
15073 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
15075 * basic-block.h (emit_barrier_after_bb): Declare.
15077 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15079 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
15080 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
15081 arm_usatsihi): Adjust alternatives for arm_restrict_it.
15083 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15085 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
15087 * config/arm/ldmstm.md: Regenerate.
15089 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15091 * config/arm/sync.md (atomic_loaddi_1):
15092 Disable predication for arm_restrict_it.
15093 (arm_load_exclusive<mode>): Likewise.
15094 (arm_load_exclusivesi): Likewise.
15095 (arm_load_exclusivedi): Likewise.
15096 (arm_load_acquire_exclusive<mode>): Likewise.
15097 (arm_load_acquire_exclusivesi): Likewise.
15098 (arm_load_acquire_exclusivedi): Likewise.
15099 (arm_store_exclusive<mode>): Likewise.
15100 (arm_store_exclusive<mode>): Likewise.
15101 (arm_store_release_exclusivedi): Likewise.
15102 (arm_store_release_exclusive<mode>): Likewise.
15104 2013-06-06 Richard Biener <rguenther@suse.de>
15106 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
15108 (lto_tag_is_tree_code_p): Adjust.
15109 (lto_tag_is_gimple_code_p): Likewise.
15110 (lto_gimple_code_to_tag): Likewise.
15111 (lto_tag_to_gimple_code): Likewise.
15112 (lto_tree_code_to_tag): Likewise.
15113 (lto_tag_to_tree_code): Likewise.
15114 * data-streamer.h (streamer_write_hwi_in_range): Use
15115 uhwi streaming to stream the normalized range.
15116 (streamer_read_hwi_in_range): Likewise.
15118 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15120 * config/arm/arm.md (enabled_for_depr_it): New attribute.
15121 (predicable_short_it): Likewise.
15122 (predicated): Likewise.
15123 (enabled): Handle above.
15124 (define_cond_exec): Set predicated attribute to yes.
15126 2013-06-05 Mike Stump <mikestump@comcast.net>
15128 * gdbinit.in (__FUNCTION__): Add.
15130 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
15132 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
15133 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
15135 2013-06-05 Jan Hubicka <jh@suse.cz>
15137 * varasm.c (mark_decl_referenced): Revert the removal until targets
15140 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
15142 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
15143 instead of mark_decl_referenced.
15145 2013-06-05 Jan Hubicka <jh@suse.cz>
15147 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
15148 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
15149 and symtab_used_from_object_file_p.
15150 (cgraph_make_node_local_1): Clear forced_by_abi.
15151 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
15152 * cgraph.h (symtab_node_base): Add forced_by_abi.
15153 (decide_is_variable_needed): Remove.
15154 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
15155 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
15156 (decide_is_symbol_needed): ... this one; handle symbols in general;
15157 always analyze virtuals; honnor forced_by_abi.
15158 (cgraph_finalize_function): Update.
15159 (varpool_finalize_decl): Update.
15160 (symbol_defined_and_needed): Remove.
15161 (analyze_functions): Update.
15162 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
15163 output_refs, input_overwrite_node): Handle forced_by_abi.
15164 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
15165 (address_taken_from_non_vtable_p): ... this one.
15166 (comdat_can_be_unshared_p_1): New function.
15167 (cgraph_comdat_can_be_unshared_p): Rename to ...
15168 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
15169 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
15170 (function_and_variable_visibility): Clear forced_by_abi as needed.
15171 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
15172 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
15173 * symtab.c (dump_symtab_base): Dump forced_by_abi.
15174 * varpool.c (decide_is_variable_needed): Remove.
15176 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15178 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
15179 (arm_option_override): Override arm_restrict_it where appropriate.
15180 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
15181 * config/arm/arm.opt (mrestrict-it): New command-line option.
15182 * doc/invoke.texi: Document -mrestrict-it.
15184 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15186 * tsan.c (tsan_atomic_table): Make const.
15188 2013-06-05 Richard Biener <rguenther@suse.de>
15190 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
15191 index associated with the tree we are supposed to replace.
15192 * tree-streamer-out.c (pack_ts_base_value_fields): Output
15193 TREE_ASM_WRITTEN as zero for everything but SSA names.
15195 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15197 * tree-ssa-structalias.c (call_stmt_vars): Make static.
15199 2013-06-04 Jan Hubicka <jh@suse.cz>
15201 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
15202 (input_node, input_varpool_node): Handle correctly external same
15204 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
15205 nodes at ltrans stage.
15207 2013-06-04 Jan Hubicka <jh@suse.cz>
15209 * ipa-inline.c (update_caller_keys): Fix availability test.
15210 (update_callee_keys): Likewise.
15211 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
15212 to follow ELF standard.
15214 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
15216 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
15217 (mips64r5900el-*-elf*): New configurations.
15218 * config/mips/mips-cpus.def (r5900): New processor.
15219 * config/mips/mips-tables.opt: Regenerate.
15220 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
15221 (mips_issue_rate): Handle PROCESSOR_R5900.
15222 (mips_reorg_process_insns): Force reorder mode for the R5900.
15223 * config/mips/mips.h (TARGET_MIPS5900): Define.
15224 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
15226 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
15228 * config/mips/mips.md (processor): Add r5900.
15229 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
15231 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15233 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
15234 into function to generate MOVI instruction.
15235 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
15236 (aarch64_preferred_simd_mode): Turn into wrapper.
15237 (aarch64_output_scalar_simd_mov_immediate): New function.
15238 * config/aarch64/aarch64-protos.h: Add prototype for above.
15240 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15242 * config/aarch64/aarch64.c (simd_immediate_info): Remove
15243 element_char member.
15244 (sizetochar): Return signed char.
15245 (aarch64_simd_valid_immediate): Remove elchar and other
15246 unnecessary variables.
15247 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
15248 Calculate element_char as required.
15249 * config/aarch64/aarch64-protos.h: Update and move prototype
15250 for aarch64_output_simd_mov_immediate.
15251 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
15254 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15256 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
15257 information completed by aarch64_simd_valid_immediate.
15258 (aarch64_legitimate_constant_p): Update arguments.
15259 (aarch64_simd_valid_immediate): Work with struct rather than many
15261 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
15262 (aarch64_simd_make_constant): Update arguments.
15263 (aarch64_output_simd_mov_immediate): Work with struct rather than
15264 many pointers. Output immediate directly rather than as operand.
15265 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15267 * config/aarch64/constraints.md (Dn): Update arguments.
15269 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15271 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
15273 (aarch64_simd_immediate_valid_for_move): Remove.
15274 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
15275 (aarch64_simd_make_constant): Update call.
15276 (aarch64_output_simd_mov_immediate): Update call.
15277 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15279 * config/aarch64/constraints.md (Dn): Update call.
15281 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15283 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
15284 return type to bool for prototype.
15285 (aarch64_legitimate_constant_p): Check for true instead of not -1.
15286 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
15287 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
15289 2013-06-04 Catherine Moore <clm@codesourcery.com>
15291 * config/mips/mips.opt (meva): New.
15292 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
15293 (ASM_SPEC): Handle -meva.
15294 * doc/invoke.texi (meva): Document.
15296 2013-06-04 Alan Modra <amodra@gmail.com>
15298 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
15301 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15303 * rtl.def: Add extra fourth optional field to define_cond_exec.
15304 * gensupport.c (process_one_cond_exec): Process attributes from
15306 * doc/md.texi: Document fourth field in define_cond_exec.
15308 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
15310 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
15311 out the processing order as in store_bit_field_1.
15313 2013-06-04 Jan Hubicka <jh@suse.cz>
15315 PR middle-end/57500
15316 * cgraphunit.c (cgraph_process_same_body_aliases): Create
15317 non-VAR_DECL node if it does not exist yet.
15319 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
15321 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
15322 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
15323 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
15324 target_cpu_default setting.
15326 2013-06-03 Teresa Johnson <tejohnson@google.com>
15328 * dumpfile.c (opt_info_switch_p): Change -fopt-info
15329 default to -fopt-info=optimized instead of all.
15330 * doc/invoke.texi: Ditto.
15331 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
15332 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
15333 (execute_vect_slp): Emit BB vectorization success under
15334 MSG_OPTIMIZED_LOCATIONS.
15335 * tree-vect-slp.c (vect_slp_transform_bb): Change
15336 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
15337 * tree-vect-loop.c (vect_transform_loop): Ditto.
15339 2013-06-03 Jason Merrill <jason@redhat.com>
15342 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
15343 Use TARGET_EXPR for C++.
15345 2013-06-03 Jakub Jelinek <jakub@redhat.com>
15347 PR rtl-optimization/57268
15348 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
15349 if DEBUG_INSN_P (insn).
15352 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15354 PR rtl-optimization/57268
15355 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15356 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15358 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15360 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
15361 (ix86_avoid_lea_for_addr): Likewise.
15362 (exact_dependency_1): Likewise.
15363 (ix86_adjust_cost): Likewise.
15364 (swap_top_of_ready_list): Fix formatting and !reload_completed check
15366 (do_reorder_for_imul): Fix typo, formatting and
15367 !reload_completed check removed.
15368 (ix86_sched_reorder): Fix typo and formatting.
15369 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
15372 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
15374 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
15376 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
15378 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
15379 <STRING_CST>: Likewise.
15380 <VECTOR_CST>: Likewise.
15382 2013-06-01 Janus Weil <janus@gcc.gnu.org>
15383 Mikael Morin <mikael@gcc.gnu.org>
15385 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
15386 * config.in: Regenerated.
15387 * configure: Regenerated.
15389 2013-06-01 Jan Hubicka <jh@suse.cz>
15391 PR middle-end/57366
15392 * cgraphunit.c (compile): When weakref is not supported,
15393 set up transparent aliases before final output pass.
15394 * varasm.c (assemble_alias): Do not try to do it here.
15396 2013-06-01 Jan Hubicka <jh@suse.cz>
15398 PR middle-end/57467
15399 * passes.c (for_per_function): Skip unanalyzed functions.
15401 2013-06-01 Jan Hubicka <jh@suse.cz>
15403 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
15404 (lto_symtab_merge_symbols_1): ... this one.
15405 (lto_symtab_merge_cgraph_nodes): Rename to ...
15406 (lto_symtab_merge_symbols): ... this one; simplify.
15407 * cgraph.c (same_body_aliases_done): Rename to ...
15408 (cpp_implicit_aliases_done): ... this one.
15409 (cgraph_create_function_alias): Update.
15410 (cgraph_same_body_alias): Update.
15411 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
15412 (verify_edge_corresponds_to_fndecl): Simplify.
15413 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
15414 (cgraph_node): Remove same_body_alias.
15415 (varpool_node): Remove alias_of and extra_name_alias.
15416 (same_body_aliases_done): Rename to ..
15417 (cpp_implicit_aliases_done): ... this one.
15418 (symtab_alias_ultimate_target): Add default parameter.
15419 (symtab_resolve_alias): New function.
15420 (fixup_same_cpp_alias_visibility): Declare.
15421 (cgraph_function_node): Add default parameter.
15422 (cgraph_node_asm_name): Likewise.
15423 (cgraph_function_or_thunk_node): Add default parameter; do
15424 not ICE when it is NULL.
15425 (varpool_variable_node): Likewise.
15426 * tree-emutls.c (create_emultls_var): Update.
15427 (ipa_lower_emutls): Update.
15428 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
15429 (cgraph_reset_node): Reset alias info.
15430 (cgraph_finalize_function): Update.
15431 (fixup_same_cpp_alias_visibility): Move to symtab.c.
15432 (analyze_function): Simplify.
15433 (cgraph_process_same_body_aliases): Simplify.
15434 (analyze_functions): Fixup same body aliases.
15435 (handle_alias_pairs): Simplify.
15436 (assemble_thunk): Update.
15437 (assemble_thunks_and_aliases): Update.
15438 (output_weakrefs): Rewrite.
15439 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
15440 (lto_output_varpool_node): Likewise.
15441 (compute_ltrans_boundary): Remve assert.
15442 (get_alias_symbol): New functoin.
15443 (input_node): Rewrite alias handling.
15444 (input_varpool_node): Likewise.
15445 * ipa-pure-const.c (propagate_pure_const): Fix formating.
15446 * ipa.c (process_references): Handle weakrefs correctly.
15447 (symtab_remove_unreachable_nodes): Likewise.
15448 * trans-mem.c (get_cg_data): Update.
15449 (ipa_tm_create_version_alias): Update.
15450 (ipa_tm_execute): Update.
15451 * symtab.c (dump_symtab_base): Dump aliases.
15452 (verify_symtab_base): Verify aliases.
15453 (symtab_node_availability): New function.
15454 (symtab_alias_ultimate_target): Simplify.
15455 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
15456 handle all the fixup cases.
15457 (symtab_resolve_alias): New function.
15458 * passes.c (ipa_write_summaries): Handle weakrefs.
15459 * varpool.c (varpool_analyze_node): Simplify.
15460 (assemble_aliases): Update.
15461 (varpool_create_variable_alias): Simplify.
15462 (varpool_extra_name_alias): Simplify.
15463 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
15464 (lto_symtab_merge_symbols): ... this one.
15466 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
15469 PR rtl-optimization/57268
15470 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15471 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15473 2013-06-01 Tobias Burnus <burnus@net-b.de>
15475 Partially reverted:
15476 2013-05-31 Tobias Burnus <burnus@net-b.de>
15478 PR middle-end/57073
15479 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
15482 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15484 PR rtl-optimization/57268
15485 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15486 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15488 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15490 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
15491 unordered comparison operators when -fno-trapping-math is in effect
15493 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
15494 and implement unordered comparison operators properly on the e500.
15496 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15498 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
15499 for constant scalar integers.
15500 (simplify_relational_operation_1): Likewise.
15502 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
15504 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
15505 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
15508 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15509 Igor Zamyatin <igor.zamyatin@intel.com>
15511 Silvermont (SLM) architecture performance tuning.
15512 * config/i386/i386.h (enum ix86_tune_indices): Add
15513 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
15514 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
15516 * config/i386/i386.c (initial_ix86_tune_features)
15517 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
15518 (ix86_lea_outperforms): Handle Silvermont tuning.
15519 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
15521 (ix86_use_lea_for_mov): Likewise.
15522 (ix86_avoid_lea_for_addr): Likewise.
15523 (ix86_lea_for_add_ok): Likewise.
15524 (exact_dependency_1): New function.
15525 (exact_store_load_dependency): Likewise.
15526 (ix86_adjust_cost): Handle Silvermont tuning.
15527 (do_reoder_for_imul): Likewise.
15528 (swap_top_of_ready_list): New function.
15529 (ix86_sched_reorder): Changed to handle Silvermont tuning.
15531 * config/i386/i386.md (peepholes that split memory operand in fp
15534 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15536 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
15537 Remove un-necessary braces.
15539 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15541 * config/aarch64/aarch64.c (aarch64_classify_symbol):
15542 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
15544 2013-05-31 Tobias Burnus <burnus@net-b.de>
15546 PR middle-end/57073
15547 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
15549 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15552 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
15553 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
15554 * config/arm/neon.md (iordi3_neon): Remove.
15555 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
15556 * config/arm/predicates.md (imm_for_neon_logic_operand):
15557 Move to earlier in the file.
15558 (neon_logic_op2): Likewise.
15559 (arm_iordi_operand_neon): New predicate.
15561 2013-05-31 Richard Biener <rguenther@suse.de>
15563 PR tree-optimization/57478
15564 PR tree-optimization/57453
15565 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
15568 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15570 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
15571 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
15573 2013-05-30 Tobias Burnus <burnus@net-b.de>
15574 Thomas Koenig <tkoenig@gcc.gnu.org>
15576 PR middle-end/57073
15577 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
15578 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
15580 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
15582 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
15584 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
15586 * target.def (register_usage_leveling_p): New hook.
15587 * targhooks.c (default_register_usage_leveling_p): New.
15588 * targhooks.h (default_register_usage_leveling_p): New prototype.
15589 * lra-assigns.c (register_usage_leveling_p): Use the hook.
15590 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
15591 * doc/tm.texi: Update.
15592 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
15594 2013-05-30 Ian Bolton <ian.bolton@arm.com>
15596 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
15597 (*insv_reg<mode>): New define_insn.
15599 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
15601 PR rtl-optimization/57439
15602 * postreload.c (move2add_valid_value_p): Check that we have
15603 a zero subreg_regno_offset when accessing the register in
15604 the requested mode.
15606 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15607 Igor Zamyatin <igor.zamyatin@intel.com>
15609 Silvermont (SLM) architecture pipeline model, tuning and
15611 * config.gcc: Add slm config options and target.
15613 * config/i386/slm.md: New.
15615 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
15617 * config/i386/i386-c.c (ix86_target_macros_internal): New case
15619 (ix86_target_macros_internal): Likewise.
15621 * config/i386/i386.c (slm_cost): New cost.
15622 (m_SLM): New macro flag.
15623 (initial_ix86_tune_features): Set m_SLM.
15624 (x86_accumulate_outgoing_args): Likewise.
15625 (x86_arch_always_fancy_math_387): Likewise.
15626 (processor_target_table): Add slm cost.
15627 (cpu_names): Add slm cpu name.
15628 (x86_option_override_internal): Set SLM ISA.
15629 (ix86_issue_rate): New case PROCESSOR_SLM.
15630 (ia32_multipass_dfa_lookahead): Likewise.
15631 (fold_builtin_cpu): Add slm.
15633 * config/i386/i386.h (TARGET_SLM): New target macro.
15634 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
15635 (processor_type): Add PROCESSOR_SLM.
15637 * config/i386/i386.md (cpu): Add new value "slm".
15638 (slm.md): Include slm.md.
15640 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
15641 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15643 * config/arm/arm-protos.h: Add and update function protos.
15644 * config/arm/arm.c (use_simple_return_p): New added.
15645 (thumb2_expand_return): Check simple_return flag.
15646 * config/arm/arm.md: Add simple_return and conditional simple_return.
15647 * config/arm/iterators.md: Add iterator for return and simple_return.
15649 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15651 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
15652 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
15653 (arm_emit_vfp_multi_reg_pop): Likewise.
15654 (thumb2_emit_ldrd_pop): Likewise.
15655 (arm_expand_epilogue): Add misc REG_CFA notes.
15656 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
15658 2013-05-29 Lawrence Crowl <crowl@google.com>
15660 * config/arm/t-arm: Update for below.
15662 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
15663 Change type to hash_table. Update dependent calls and types.
15665 * config/i386/t-cygming: Update for below.
15667 * config/i386/t-interix: Update for below.
15669 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
15670 Change type to hash_table. Update dependent calls and types.
15671 (i386_find_on_wrapper_list::wrappers): Likewise.
15673 * config/ia64/t-ia64: Update for below.
15675 * config/ia64/ia64.c (bundle_state_table):
15676 Change type to hash_table. Update dependent calls and types.
15678 * config/mips/mips.c (mips_reorg_process_insns::htab):
15679 Change type to hash_table. Update dependent calls and types.
15681 * config/sol2.c (solaris_comdat_htab):
15682 Change type to hash_table. Update dependent calls and types.
15684 * config/t-sol2: Update for above.
15686 2013-05-29 Teresa Johnson <tejohnson@google.com>
15688 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
15689 functions are not yet marked as defined.
15691 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
15692 Pat Haugen <pthaugen@us.ibm.com>
15693 Peter Bergner <bergner@vnet.ibm.com>
15695 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
15699 (vrotl<mode>3): Likewise.
15700 (vashl<mode>3): Likewise.
15701 (vlshr<mode>3): Likewise.
15702 (vashr<mode>3): Likewise.
15704 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15705 support for power8 V2DI builtins.
15707 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
15708 power8 V2DI builtins.
15709 (vupkhsw): Likewise.
15710 (vupklsw): Likewise.
15711 (vaddudm): Likewise.
15712 (vminsd): Likewise.
15713 (vmaxsd): Likewise.
15714 (vminud): Likewise.
15715 (vmaxud): Likewise.
15716 (vpkudum): Likewise.
15717 (vpksdss): Likewise.
15718 (vpkudus): Likewise.
15719 (vpksdus): Likewise.
15724 (vsubudm): Likewise.
15725 (vcmpequd): Likewise.
15726 (vcmpgtsd): Likewise.
15727 (vcmpgtud): Likewise.
15728 (vcmpequd_p): Likewise.
15729 (vcmpgtsd_p): Likewise.
15730 (vcmpgtud_p): Likewise.
15731 (vupkhsw): Likewise.
15732 (vupklsw): Likewise.
15733 (vaddudm): Likewise.
15734 (vmaxsd): Likewise.
15735 (vmaxud): Likewise.
15736 (vminsd): Likewise.
15737 (vminud): Likewise.
15738 (vpksdss): Likewise.
15739 (vpksdus): Likewise.
15740 (vpkudum): Likewise.
15741 (vpkudus): Likewise.
15746 (vsubudm): Likewise.
15748 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
15749 support for power8 V2DI instructions.
15751 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
15752 power8 V2DI instructions. Combine pack and unpack insns to use an
15753 iterator for each mode. Check whether a particular mode supports
15754 Altivec instructions instead of just checking TARGET_ALTIVEC.
15755 (UNSPEC_VPKUWUM): Likewise.
15756 (UNSPEC_VPKSHSS): Likewise.
15757 (UNSPEC_VPKSWSS): Likewise.
15758 (UNSPEC_VPKUHUS): Likewise.
15759 (UNSPEC_VPKSHUS): Likewise.
15760 (UNSPEC_VPKUWUS): Likewise.
15761 (UNSPEC_VPKSWUS): Likewise.
15762 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
15763 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
15764 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
15765 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
15766 (UNSPEC_VUPKHSB): Likewise.
15767 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
15768 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
15769 (UNSPEC_VUPKHSH): Likewise.
15770 (UNSPEC_VUPKLSB): Likewise.
15771 (UNSPEC_VUPKLSH): Likewise.
15773 (VI_char): Likewise.
15774 (VI_scalar): Likewise.
15775 (VI_unit): Likewise.
15777 (VP_small): Likewise.
15778 (VP_small_lc): Likewise.
15779 (VU_char): Likewise.
15780 (add<mode>3): Likewise.
15781 (altivec_vaddcuw): Likewise.
15782 (altivec_vaddu<VI_char>s): Likewise.
15783 (altivec_vadds<VI_char>s): Likewise.
15784 (sub<mode>3): Likewise.
15785 (altivec_vsubcuw): Likewise.
15786 (altivec_vsubu<VI_char>s): Likewise.
15787 (altivec_vsubs<VI_char>s): Likewise.
15788 (altivec_vavgs<VI_char>): Likewise.
15789 (altivec_vcmpbfp): Likewise.
15790 (altivec_eq<mode>): Likewise.
15791 (altivec_gt<mode>): Likewise.
15792 (altivec_gtu<mode>): Likewise.
15793 (umax<mode>3): Likewise.
15794 (smax<mode>3): Likewise.
15795 (umin<mode>3): Likewise.
15796 (smin<mode>3): Likewise.
15797 (altivec_vpkuhum): Likewise.
15798 (altivec_vpkuwum): Likewise.
15799 (altivec_vpkshss): Likewise.
15800 (altivec_vpkswss): Likewise.
15801 (altivec_vpkuhus): Likewise.
15802 (altivec_vpkshus): Likewise.
15803 (altivec_vpkuwus): Likewise.
15804 (altivec_vpkswus): Likewise.
15805 (altivec_vpks<VI_char>ss): Likewise.
15806 (altivec_vpks<VI_char>us): Likewise.
15807 (altivec_vpku<VI_char>us): Likewise.
15808 (altivec_vpku<VI_char>um): Likewise.
15809 (altivec_vrl<VI_char>): Likewise.
15810 (altivec_vsl<VI_char>): Likewise.
15811 (altivec_vsr<VI_char>): Likewise.
15812 (altivec_vsra<VI_char>): Likewise.
15813 (altivec_vsldoi_<mode>): Likewise.
15814 (altivec_vupkhsb): Likewise.
15815 (altivec_vupkhs<VU_char>): Likewise.
15816 (altivec_vupkls<VU_char>): Likewise.
15817 (altivec_vupkhsh): Likewise.
15818 (altivec_vupklsb): Likewise.
15819 (altivec_vupklsh): Likewise.
15820 (altivec_vcmpequ<VI_char>_p): Likewise.
15821 (altivec_vcmpgts<VI_char>_p): Likewise.
15822 (altivec_vcmpgtu<VI_char>_p): Likewise.
15823 (abs<mode>2): Likewise.
15824 (vec_unpacks_hi_v16qi): Likewise.
15825 (vec_unpacks_hi_v8hi): Likewise.
15826 (vec_unpacks_lo_v16qi): Likewise.
15827 (vec_unpacks_hi_<VP_small_lc>): Likewise.
15828 (vec_unpacks_lo_v8hi): Likewise.
15829 (vec_unpacks_lo_<VP_small_lc>): Likewise.
15830 (vec_pack_trunc_v8h): Likewise.
15831 (vec_pack_trunc_v4si): Likewise.
15832 (vec_pack_trunc_<mode>): Likewise.
15834 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
15836 (vec_vmaxsd): Likewise.
15837 (vec_vmaxud): Likewise.
15838 (vec_vminsd): Likewise.
15839 (vec_vminud): Likewise.
15840 (vec_vpksdss): Likewise.
15841 (vec_vpksdus): Likewise.
15842 (vec_vpkudum): Likewise.
15843 (vec_vpkudus): Likewise.
15844 (vec_vrld): Likewise.
15845 (vec_vsld): Likewise.
15846 (vec_vsrad): Likewise.
15847 (vec_vsrd): Likewise.
15848 (vec_vsubudm): Likewise.
15849 (vec_vupkhsw): Likewise.
15850 (vec_vupklsw): Likewise.
15852 2013-05-29 Jan Hubicka <jh@suse.cz>
15854 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
15855 flags; reorder rest of fields in more consistent way.
15856 (varpool_node): Remove analyzed, finalized and alias.
15857 (cgraph_ndoe): Likewise.
15858 (symtab_alias_ultimate_target): New function.
15859 (cgraph_function_node): Move offline.
15860 (cgraph_reset_node): Declare.
15861 (cgraph_comdat_can_be_unshared_p): Remove.
15862 (varpool_remove_initializer): Declare.
15863 (varpool_first_defined_variable, varpool_next_defined_variable
15864 cgraph_first_defined_function, cgraph_next_defined_function): Update.
15865 (cgraph_function_with_gimple_body_p): Update.
15866 (varpool_all_refs_explicit_p): Update.
15867 (symtab_alias_target): New function.
15868 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
15869 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
15870 (cgraph_function_or_thunk_node): Simplify using
15871 symtab_alias_ultimate_target.
15872 (varpool_variable_node): Likewise.
15873 * cgraph.c (cgraph_create_function_alias): Update.
15874 (cgraph_add_thunk): Update.
15875 (cgraph_remove_node): Update.
15876 (dump_cgraph_node): Do not dump removed flags.
15877 (cgraph_function_body_availability): Update.
15878 (cgraph_propagate_frequency): Update.
15879 (verify_cgraph_node): Check sanity of local flag.
15880 (cgraph_function_node): Move here from cgraph.h; revamp for
15881 cgraph_function_or_thunk_node.
15882 * lto-symtab.c (lto_varpool_replace_node): Update.
15883 (lto_symtab_resolve_can_prevail_p): Update.
15884 (lto_symtab_merge_cgraph_nodes): Update.
15885 * ipa-cp.c (determine_versionability, initialize_node_lattices,
15886 propagate_constants_accross_call, devirtualization_time_bonus,
15887 ipcp_propagate_stage): Update.
15888 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
15889 * ipa-inline-transform.c (clone_inlined_nodes,
15890 preserve_function_body_p): Update.
15891 * ipa-reference.c (propagate): Update.
15892 (write_node_summary_p): Update.
15893 * toplev.c (wrapup_global_declaration_2): Update.
15894 * cgraphunit.c (cgraph_analyze_function): Rename to ...
15895 (analyze_function) ... this one.
15896 (cgraph_process_new_functions): Update.
15897 (cgraph_reset_node): Export.
15898 (cgraph_finalize_function): Update.
15899 (cgraph_add_new_function): Update.
15900 (process_function_and_variable_attributes): Update.
15901 (varpool_finalize_decl): Update.
15902 (symbol_finalized): Remove.
15903 (symbol_finalized_and_needed): Rename to ...
15904 (symbol_defined_and_needed): ... update.
15905 (cgraph_analyze_functions): Update.
15906 (handle_alias_pairs): Update.
15907 (mark_functions_to_output): Update.
15908 (assemble_thunk): Update.
15909 (output_in_order): Update.
15910 (output_weakrefs): Update.
15911 (finalize_compilation_unit): Update.
15912 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
15913 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
15914 input_node, input_varpool_node): Update.
15915 * dbxout.c (dbxout_expand_expr): Update.
15916 * cgraphclones.c (cgraph_clone_node): Update.
15917 (cgraph_copy_node_for_versioning): Update.
15918 (cgraph_materialize_clone): Update.
15919 (cgraph_materialize_all_clones): Update.
15920 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
15921 propagate_pure_const, propagate_nothrow): Update.
15922 * lto-streamer-out.c (lto_output, write_symbol): Update.
15923 * ipa-utils.c (ipa_reverse_postorder): Update.
15924 * ipa-inline.c (can_inline_edge_p): Update.
15925 (update_caller_keys, ipa_inline): Update.
15926 * dwarf2out.c (reference_to_unused,
15927 premark_types_used_by_global_vars_helper): Update.
15928 * tree-eh.c (tree_could_trap_p): Update.
15929 * ipa-split.c (consider_split, execute_split_functions): Update.
15930 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
15931 has_addr_references_p): Update; move ahead in file for better
15933 (process_references): Simplify.
15934 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
15935 bodies are removed.
15936 (cgraph_comdat_can_be_unshared_p): Make static.
15937 (cgraph_externally_visible_p): Update.
15938 (varpool_externally_visible_p): Update.
15939 (function_and_variable_visibility): Update.
15940 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
15941 ipa_tm_mark_force_output_node): Update.
15942 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
15943 estimate_edge_devirt_benefit, inline_generate_summary,
15944 inline_write_summary): Update.
15945 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
15946 * ipa-prop.c (ipa_compute_jump_functions): Update.
15947 (ipa_print_node_params, ipa_prop_read_section,
15948 ipa_update_after_lto_read, read_replacements_section): Update.
15949 * varasm.c (mark_decl_referenced): Update.
15950 (assemble_alias, dump_tm_clone_pairs): Update.
15951 * tree-inline.c (copy_bb): Update.
15952 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
15954 * symtab.c (dump_symtab_base): Print new flags.
15955 (verify_symtab_base): Verify new flags.
15956 (symtab_alias_ultimate_target): New function.
15957 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
15958 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
15960 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
15962 * i386.c (ix86_get_function_versions_dispatcher,
15963 ix86_generate_version_dispatcher_body): Update.
15964 (fold_builtin_cpu): Use varpool_add_new_variable.
15965 * varpool.c (varpool_remove_initializer): Break out from ...
15966 (varpool_remove_node): ... this one.
15967 (dump_varpool_node, varpool_node_for_asm,
15968 cgraph_variable_initializer_availability, varpool_analyze_node,
15969 varpool_assemble_decl, varpool_remove_unreferenced_decls,
15970 varpool_finalize_named_section_flags, varpool_create_variable_alias):
15973 2013-05-29 Jan Hubicka <jh@suse.cz>
15975 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
15977 2013-05-29 Easwaran Raman <eraman@google.com>
15979 PR tree-optimization/57442
15980 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
15981 when control exits the main loop.
15983 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
15985 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
15987 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
15988 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
15989 * rx/t-rx: Add rx100 under multi library matches option for nofpu
15992 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15994 PR tree-optimization/57441
15995 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
15996 Don't limit size of incr_vec to number of candidates.
15998 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
16000 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
16001 and mips16 directories.
16002 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
16003 (MULTILIB_DIRNAMES): Ditto.
16004 (MULTILIB_EXCEPTIONS): Add new exceptions.
16005 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
16006 (MULTILIB_DIRNAMES): Ditto.
16007 (MULTILIB_EXCEPTIONS): Add new exceptions.
16009 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16010 Marcus Shawcroft <marcus.shawcroft@arm.com>
16012 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
16013 SYMBOL_TINY_ABSOLUTE.
16014 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
16015 SYMBOL_TINY_ABSOLUTE.
16016 (aarch64_expand_mov_immediate): Likewise.
16017 (aarch64_classify_symbol): Likewise.
16018 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
16019 Permit SYMBOL_TINY_ABSOLUTE.
16020 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
16022 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16023 Marcus Shawcroft <marcus.shawcroft@arm.com>
16025 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
16026 Refactor if/switch. Replace gcc_assert with if.
16028 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16030 * config/i386/i386.c (initial_ix86_tune_features): Enable
16031 FP Reassociation for AMD bdver1 and bdver2.
16033 2013-05-29 Martin Jambor <mjambor@suse.cz>
16035 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
16036 and IMAGPART_EXPR do not occur within other handled_components.
16038 2013-05-29 Richard Biener <rguenther@suse.de>
16040 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
16041 access on whether the use is in the BB we currently try to
16043 (vect_bb_vectorization_profitable_p): Pass the BB we currently
16044 vectorize to vect_bb_slp_scalar_cost.
16046 2013-05-29 Richard Biener <rguenther@suse.de>
16048 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
16049 computing scalar cost offsetted by stmts that are kept live
16051 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
16052 for computation of scalar cost.
16054 2013-05-28 Steve Ellcey <sellcey@mips.com>
16056 * config/mips/mips-cpus.def (mips32r2): Change processor type.
16058 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
16060 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
16061 array notation built-in reduction functions.
16062 * doc/passes.texi (Passes): Added documentation about changes done
16064 * doc/invoke.texi (C Dialect Options): Added documentation about
16065 the -fcilkplus flag.
16066 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
16067 (BUILTINS_DEF): Depend on cilkplus.def.
16068 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
16069 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
16070 * cilkplus.def: New file.
16072 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
16074 PR rtl-optimization/57439
16075 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
16077 2013-05-28 Easwaran Raman <eraman@google.com>
16079 PR tree-optimization/57337
16080 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
16081 (find_insert_point): Correctly identify the insertion point
16082 when two statements with the same UID is compared.
16084 2013-05-28 Richard Biener <rguenther@suse.de>
16086 PR tree-optimization/56787
16087 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
16088 from the list of data references.
16089 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
16091 (vect_analyze_loop_operations): Likewise.
16092 (vect_transform_loop): Remove clobbers.
16094 2013-05-28 Martin Jambor <mjambor@suse.cz>
16096 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
16097 and REALPART_EXPRs have scalar type.
16099 2013-05-28 Richard Biener <rguenther@suse.de>
16101 PR tree-optimization/57411
16102 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
16104 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
16105 virtual operand propagation.
16107 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16109 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
16110 destination register for bmasksi_vis.
16111 (vector_init_bshuffle): Likewise.
16112 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
16114 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16116 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
16117 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
16118 mode if the instruction isn't available in the original mode.
16119 * config/sparc/sparc.opt (mfix-ut699): New option.
16120 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
16121 (divdf3): Turn into expander.
16122 (divdf3_nofix): New insn.
16123 (divdf3_fix): Likewise.
16124 (divsf3): Disable if -mfix-ut699.
16125 (sqrtdf2): Turn into expander.
16126 (sqrtdf2_nofix): New insn.
16127 (sqrtdf2_fix): Likewise.
16128 (sqrtsf2): Disable if -mfix-ut699.
16130 2013-05-27 Richard Biener <rguenther@suse.de>
16132 PR middle-end/57412
16133 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
16134 block for the new loop.
16136 2013-05-27 Richard Biener <rguenther@suse.de>
16138 PR tree-optimization/57343
16139 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
16140 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
16141 (number_of_iterations_cond): Do not build the folded tree.
16143 2013-05-27 Richard Biener <rguenther@suse.de>
16146 PR middle-end/57381
16147 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16148 OEP_CONSTANT_ADDRESS_OF retained.
16150 PR tree-optimization/57417
16151 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
16152 for unchanged base.
16153 (set_ssa_val_to): Compare addresses using
16154 get_addr_base_and_unit_offset.
16156 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
16158 PR rtl-optimization/56833
16159 * postreload.c (move2add_record_mode): New function.
16160 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
16161 (move2add_use_add2_insn): Use move2add_record_sym_value.
16162 (move2add_use_add3_insn): Likewise.
16163 (reload_cse_move2add): Use move2add_valid_value_p and
16164 move2add_record_mode. Invalidate call-clobbered and REG_INC
16165 affected regs by setting reg_mode to VOIDmode.
16166 (move2add_note_store): Don't pretend the inside of a SUBREG is
16167 the actual destination. Invalidate single/leading registers by
16168 setting reg_mode to VOIDmode.
16169 Use move2add_record_sym_value, move2add_valid_value_p and
16170 move2add_record_mode.
16172 2013-05-27 Richard Biener <rguenther@suse.de>
16174 PR tree-optimization/57396
16175 * tree-affine.c (double_int_constant_multiple_p): Properly
16176 return false for val == 0 and div != 0.
16178 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16180 * config/mips/mips.h: Use #elif in preprocessor conditions.
16182 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16185 * config/mips/constraints.md (kl): New constraint.
16186 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
16187 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
16188 constraint for operand 0. Split after CSE for MIPS16. Emit a move
16189 from LO for MIPS16.
16190 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
16192 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16195 * config/mips/mips.c (mips_can_inline_p): New function.
16196 (TARGET_CAN_INLINE_P): Define.
16198 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
16200 * sched-int.h (ds_t, dw_t): Make unsigned int.
16201 Fix documentation that describes how all the ds_t bits are used.
16202 Reserve the last bit for delayed-branch scheduling.
16203 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
16204 (BITS_PER_DEP_WEAK): Fix definition and documentation.
16205 (gen_dep_weak_1): Remove prototype.
16206 * sched-deps.c (get_dep_weak_1): Make static.
16207 * target.def (speculate_insn, needs_block_p, gen_spec_check,
16208 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
16209 * doc/tm.texi: Regenerate.
16210 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
16212 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
16215 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
16217 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
16218 Sandra Loosemore <sandra@codesourcery.com>
16220 * config.gcc (powerpc-*): Allow native for with-cpu.
16222 2013-05-24 Jeff Law <law@redhat.com>
16224 PR tree-optimization/57124
16225 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
16226 conversion feeding a condition if the range has an overflow
16227 if -fstrict-overflow. Add warnings for when we do make the
16230 2013-05-24 Dehao Chen <dehao@google.com>
16232 * tree-cfg.c (locus_discrim_map): Fix the typo.
16233 (locus_discrim_hasher): Likewise.
16234 (locus_discrim_hasher::hash): Likewise.
16235 (locus_discrim_hasher::equal): Likewise.
16237 2013-05-24 Martin Jambor <mjambor@suse.cz>
16239 PR tree-optimization/57294
16240 * cgraph.h (ipa_record_stmt_references): Declare.
16241 * cgraphbuild.c (ipa_record_stmt_references): New function.
16242 (build_cgraph_edges): Use ipa_record_stmt_references.
16243 (rebuild_cgraph_edges): Likewise.
16244 (cgraph_rebuild_references): Likewise.
16245 * ipa-prop.c (ipa_modify_call_arguments): Discard references
16246 associated with the old statement and build references from the
16247 newly built statements.
16248 * ipa-ref.c (ipa_remove_stmt_references): New function.
16249 * ipa-ref.h (ipa_remove_stmt_references): Declare.
16251 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
16253 * lra-constraints.c (emit_spill_move): Use smaller mode for
16255 (check_and_process_move): Consider mem-reg moves for secondary
16257 (curr_insn_transform): Don't lose insns emitted before for
16258 secondary memory moves.
16259 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
16260 reg set up in the current insn.
16262 2013-05-24 Dehao Chen <dehao@google.com>
16264 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
16266 (locus_descrim_hasher::equal): Likewise.
16267 (build_gimple_cfg): New discriminator assignment algorithm.
16268 (make_edges): Likewise.
16269 (next_discriminator_for_locus): Likewise.
16270 (same_line_p): Likewise.
16271 (assign_discriminators): Likewise.
16272 (make_cond_expr_edges): Likewise.
16273 (make_gimple_switch_edges): Likewise.
16274 (make_goto_expr_edges): Likewise.
16275 (make_gimple_asm_edges): Likewise.
16277 2013-05-24 Ian Bolton <ian.bolton@arm.com>
16279 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
16280 X format specifier to only display bottom 16 bits.
16281 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
16282 immediate to match for operand 2, since it will be masked.
16284 2013-05-24 Richard Biener <rguenther@suse.de>
16286 PR tree-optimization/57287
16287 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
16288 all SSA names that occur in abnormal PHIs.
16290 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
16293 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
16294 that index is not negative.
16296 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
16298 PR rtl-optimization/55177
16299 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
16300 (simplify_byte_swapping_operation): New.
16301 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
16302 (simplify_relational_operation_1): Deal with BSWAP.
16304 2013-05-23 Richard Henderson <rth@redhat.com>
16307 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
16308 (ix86_reorg): Call it.
16310 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
16313 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
16314 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
16315 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
16317 2013-05-23 Christian Bruel <christian.bruel@st.com>
16320 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
16322 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16323 Marcus Shawcroft <marcus.shawcroft@arm.com>
16325 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
16326 * config/aarch64/constraints.md (Usa): Remove.
16327 * doc/md.texi (AArch64 Usa): Remove.
16329 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16330 Marcus Shawcroft <marcus.shawcroft@arm.com>
16332 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
16333 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
16334 * config/aarch64/predicates.md (aarch64_const_address): Remove.
16335 (aarch64_mov_operand): Use aarch64_mov_operand_p.
16337 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
16339 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
16340 instruction (AdvSIMD).
16341 * config/aarch64/aarch64-builtins.c
16342 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
16343 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
16345 2013-05-23 Martin Jambor <mjambor@suse.cz>
16347 PR middle-end/57347
16348 * tree.h (contains_bitfld_component_ref_p): Declare.
16349 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
16350 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
16352 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
16353 not access a bit-field. Assert all final offsets are byte-aligned.
16355 2013-05-23 Richard Biener <rguenther@suse.de>
16357 PR tree-optimization/57380
16358 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
16359 least one invariant or re-used load.
16360 * passes.c (init_optimization_passes): Move pass_phiprop before
16363 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
16365 * config/aarch64/aarch64-simd.md
16366 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
16368 2013-05-23 Richard Biener <rguenther@suse.de>
16370 PR middle-end/57381
16371 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16372 OEP_CONSTANT_ADDRESS_OF retained.
16374 2013-05-23 Jakub Jelinek <jakub@redhat.com>
16376 PR middle-end/57344
16377 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
16378 don't lower unit. Handle unit not being always BITS_PER_WORD.
16380 2013-05-23 Richard Biener <rguenther@suse.de>
16382 PR rtl-optimization/57341
16383 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
16384 instead of true_dependence.
16386 2013-05-22 David Malcolm <dmalcolm@redhat.com>
16388 * bb-reorder.c (branch_threshold): Make const.
16389 (exec_threshold): Ditto.
16391 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16392 Pat Haugen <pthaugen@us.ibm.com>
16393 Peter Bergner <bergner@vnet.ibm.com>
16395 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
16396 documentation for the power8 crypto builtins.
16398 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
16400 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
16401 macros for defining power8 builtin functions.
16402 (BU_P8V_AV_2): Likewise.
16403 (BU_P8V_AV_P): Likewise.
16404 (BU_P8V_VSX_1): Likewise.
16405 (BU_P8V_OVERLOAD_1): Likewise.
16406 (BU_P8V_OVERLOAD_2): Likewise.
16407 (BU_CRYPTO_1): Likewise.
16408 (BU_CRYPTO_2): Likewise.
16409 (BU_CRYPTO_3): Likewise.
16410 (BU_CRYPTO_OVERLOAD_1): Likewise.
16411 (BU_CRYPTO_OVERLOAD_2): Likewise.
16412 (XSCVSPDP): Fix typo, point to the correct instruction.
16413 (VCIPHER): Add power8 crypto builtins.
16414 (VCIPHERLAST): Likewise.
16415 (VNCIPHER): Likewise.
16416 (VNCIPHERLAST): Likewise.
16417 (VPMSUMB): Likewise.
16418 (VPMSUMH): Likewise.
16419 (VPMSUMW): Likewise.
16420 (VPERMXOR_V2DI): Likewise.
16421 (VPERMXOR_V4SI: Likewise.
16422 (VPERMXOR_V8HI: Likewise.
16423 (VPERMXOR_V16QI: Likewise.
16424 (VSHASIGMAW): Likewise.
16425 (VSHASIGMAD): Likewise.
16426 (VPMSUM): Likewise.
16427 (VPERMXOR): Likewise.
16428 (VSHASIGMA): Likewise.
16430 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
16431 __CRYPTO__ if the crypto instructions are available.
16432 (altivec_overloaded_builtins): Add support for overloaded power8
16435 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
16436 support for power8 crypto builtins.
16437 (builtin_function_type): Likewise.
16438 (altivec_init_builtins): Add support for builtins that take vector
16439 long long (V2DI) arguments.
16441 * config/rs6000/crypto.md: New file, define power8 crypto
16444 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16445 Pat Haugen <pthaugen@us.ibm.com>
16446 Peter Bergner <bergner@vnet.ibm.com>
16448 * doc/invoke.texi (Option Summary): Add power8 options.
16449 (RS/6000 and PowerPC Options): Likewise.
16451 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
16452 constraints.md instead of rs6000.h. Reorder w* constraints. Add
16453 wm, wn, wr documentation.
16455 * config/rs6000/constraints.md (wm): New constraint for VSX
16456 registers if direct move instructions are enabled.
16457 (wn): New constraint for no registers.
16458 (wq): New constraint for quad word even GPR registers.
16459 (wr): New constraint if 64-bit instructions are enabled.
16460 (wv): New constraint if power8 vector instructions are enabled.
16461 (wQ): New constraint for quad word memory locations.
16463 * config/rs6000/predicates.md (const_0_to_15_operand): New
16464 constraint for 0..15 for crypto instructions.
16465 (gpc_reg_operand): If VSX allow registers in VSX registers as well
16466 as GPR and floating point registers.
16467 (int_reg_operand): New predicate to match only GPR registers.
16468 (base_reg_operand): New predicate to match base registers.
16469 (quad_int_reg_operand): New predicate to match even GPR registers
16470 for quad memory operations.
16471 (vsx_reg_or_cint_operand): New predicate to allow vector logical
16472 operations in both GPR and VSX registers.
16473 (quad_memory_operand): New predicate for quad memory operations.
16474 (reg_or_indexed_operand): New predicate for direct move support.
16476 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
16477 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
16478 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
16479 (POWERPC_MASKS): Add power8 options.
16480 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
16483 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
16484 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
16486 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
16487 (-mpower8-fusion): New power8 options.
16488 (-mpower8-fusion-sign): Likewise.
16489 (-mpower8-vector): Likewise.
16490 (-mcrypto): Likewise.
16491 (-mdirect-move): Likewise.
16492 (-mquad-memory): Likewise.
16494 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
16495 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
16497 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
16498 (rs6000_debug_vector_unit): Add p8_vector.
16499 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
16500 definitions. Also print fusion state.
16501 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
16502 (rs6000_builtin_mask_calculate): Add power8 builtin support.
16503 (rs6000_option_override_internal): Add support for power8.
16504 (rs6000_common_init_builtins): Add debugging for skipped builtins
16505 if -mdebug=builtin.
16506 (rs6000_adjust_cost): Add power8 support.
16507 (rs6000_issue_rate): Likewise.
16508 (insn_must_be_first_in_group): Likewise.
16509 (insn_must_be_last_in_group): Likewise.
16510 (force_new_group): Likewise.
16511 (rs6000_register_move_cost): Likewise.
16512 (rs6000_opt_masks): Likewise.
16514 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
16515 power8 capable assembler, default to power7 options.
16516 (TARGET_DIRECT_MOVE): Likewise.
16517 (TARGET_CRYPTO): Likewise.
16518 (TARGET_P8_VECTOR): Likewise.
16519 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
16520 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
16521 (VECTOR_MEM_P8_VECTOR_P): Likewise.
16522 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
16523 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
16524 (TARGET_XSCVDPSPN): Likewise.
16525 (TARGET_XSCVSPDPN): Likewsie.
16526 (TARGET_SYNC_HI_QI): Likewise.
16527 (TARGET_SYNC_TI): Likewise.
16528 (MASK_CRYPTO): Likewise.
16529 (MASK_DIRECT_MOVE): Likewise.
16530 (MASK_P8_FUSION): Likewise.
16531 (MASK_P8_VECTOR): Likewise.
16532 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
16533 temporary used by some of the direct move instructions to get two FP
16534 temporary registers does not force creation of a stack frame.
16535 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
16536 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
16537 that any VSX registers are tieable, even if they are also an
16538 Altivec vector mode.
16539 (r6000_reg_class_enum): Add wm, wr, wv constraints.
16540 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
16541 (RS6000_BTM_CRYPTO): Likewise.
16542 (RS6000_BTM_COMMON): Likewise.
16544 * config/rs6000/rs6000.md (cpu attribute): Add power8.
16545 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
16546 (enum rs6000_vector): Add power8 vector support.
16548 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16552 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
16553 (any_sibcall_could_use_r3): this and handle indirect calls.
16554 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
16556 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16558 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
16560 2013-05-22 Richard Biener <rguenther@suse.de>
16562 PR middle-end/57349
16563 * profile.c (branch_prob): Do not split blocks that are
16564 abnormally receiving from ECF_RETURNS_TWICE functions.
16566 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
16568 * recog.c (offsettable_address_addr_space_p): Fix calculation of
16569 address mode. Move pointer mode initialization to the same place.
16571 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
16573 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
16574 while it has any effect.
16576 2013-05-21 Easwaran Raman <eraman@google.com>
16578 PR tree-optimization/57322
16579 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
16580 UID of the statement added to the BB to be 1.
16582 2013-05-21 Jakub Jelinek <jakub@redhat.com>
16584 PR tree-optimization/57331
16585 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
16586 of conversion from pointer type to integral type with integer.
16588 2013-05-21 Martin Jambor <mjambor@suse.cz>
16591 * ipa-prop.c (ipa_read_node_info): Process param_used and
16592 controlled_uses in the same order as when writing.
16594 2013-05-21 Magnus Granberg <baldrick@free.fr>
16597 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
16599 2013-05-21 Richard Biener <rguenther@suse.de>
16601 PR tree-optimization/57318
16602 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
16603 estimate stmts with side-effects as likely eliminated.
16605 2013-05-21 Richard Biener <rguenther@suse.de>
16607 PR tree-optimization/57330
16608 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
16609 preserve the call stmts fntype.
16611 2013-05-21 Richard Biener <rguenther@suse.de>
16613 PR tree-optimization/57303
16614 * tree-ssa-sink.c (statement_sink_location): Improve killing
16615 stmt detection and properly handle self-assignments.
16617 2013-05-21 Christian Bruel <christian.bruel@st.com>
16619 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
16620 spanning registers. LEAF_REG_REMAP is supported only for contiguous
16621 registers. Set register size out of the PARALLEL loop.
16623 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
16626 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
16627 (*fmasf4, *fmasf4_media): New insns.
16629 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16631 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
16632 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
16633 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
16634 (mips_idiv_insns): Update the comments to say that the returned
16635 instruction counts are in units of BASE_INSN_LENGTH.
16636 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
16637 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
16638 using 2 rather than 4 as the length of indirect MIPS16 and
16639 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
16640 length of a NOP. Don't divide MIPS16 lengths by 2.
16641 (mips16_split_long_branches): Assume a branch is long if the
16642 length is greater than 4 rather than 8.
16643 * config/mips/mips.md (length): Give MIPS16 lengths directly,
16644 rather than multiplying them by 2. Multiply instruction counts
16645 by BASE_INSN_LENGTH rather than 4.
16646 (*jump_mips16, tls_get_tp_mips16_<mode>)
16647 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
16649 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16651 * config/mips/mips.md (extended_mips16): Remove branch case.
16652 (length): Remove duplicated extended_mips16 test.
16654 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16656 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
16658 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
16660 * recog.h (Recog_data): Rename to...
16661 (recog_data_d): ...this.
16662 (recog_data): Update accordingly.
16663 * recog.c (recog_data): Likewise.
16664 * reload.c (save_recog_data): Likewise.
16665 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
16666 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
16668 2013-05-17 Julian Brown <julian@codesourcery.com>
16670 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
16671 found in a REG_EQUAL note, invalidate it.
16673 2013-05-17 Easwaran Raman <eraman@google.com>
16675 * tree-ssa-reassoc.c (find_insert_point): New function.
16676 (insert_stmt_after): Likewise.
16677 (get_def_stmt): Likewise.
16678 (ensure_ops_are_available): Likewise.
16679 (not_dominated_by): Likewise.
16680 (rewrite_expr_tree): Do not move statements beyond what is
16681 necessary. Remove call to swap_ops_for_binary_stmt...
16682 (reassociate_bb): ... and move it here.
16683 (build_and_add_sum): Assign UIDs for new statements.
16684 (linearize_expr): Likewise.
16685 (do_reassoc): Renumber gimple statement UIDs.
16687 2013-05-17 Jan Hubicka <jh@suse.cz>
16689 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
16691 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
16692 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
16694 (output_weakrefs): Update.
16696 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
16697 Martin Jambor <mjambor@suse.cz>
16699 PR middle-end/57276
16700 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
16701 value that corresponds to the given aggval is found in values vector.
16703 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
16705 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
16706 sse, sse2, sse3, ssse3 and sse4a flags to options.
16708 2013-05-17 David Malcolm <dmalcolm@redhat.com>
16710 * gengtype-state.c: (s_expr_writer): New class, to handle
16711 prettifying of output layout of s-expressions.
16712 (state_writer): New class, to write out gtype.state.
16713 (state_written_type_count): Move this variable into member data of
16715 (s_expr_writer::s_expr_writer): New code: constructor for new class
16716 (state_writer::state_writer(): ditto
16717 (s_expr_writer::write_new_line): New function
16718 (s_expr_writer::write_any_indent): ditto
16719 (s_expr_writer::begin_s_expr): ditto
16720 (s_expr_writer::end_s_expr): ditto
16721 (write_state_fileloc): convert to method of state_writer...
16722 (state_writer:: write_state_fileloc): ...and use methods of
16723 s_expr_writer to write indentation into the gtype.state output file
16724 to visually represent the hierarchical structure of the list
16726 (write_state_fields): ditto, renaming to...
16727 (state_writer::write_state_fields)
16728 (write_state_a_string): ditto, renaming to...
16729 (state_writer::write_state_a_string)
16730 (write_state_string_option): ditto, renaming to...
16731 (state_writer::write_state_string_option)
16732 (write_state_type_option): ditto, renaming to...
16733 (state_writer::write_state_type_option)
16734 (write_state_nested_option): ditto, renaming to...
16735 (state_writer::write_state_nested_option)
16736 (write_state_option): ditto, renaming to...
16737 (state_writer::write_state_option)
16738 (write_state_options): ditto, renaming to...
16739 (state_writer::write_state_options)
16740 (write_state_lang_bitmap): ditto, renaming to...
16741 (state_writer::write_state_lang_bitmap)
16742 (write_state_version): ditto, renaming to...
16743 (state_writer::write_state_version)
16744 (write_state_scalar_type): ditto, renaming to...
16745 (state_writer::write_state_scalar_type)
16746 (write_state_string_type): ditto, renaming to...
16747 (state_writer::write_state_string_type)
16748 (write_state_undefined_type): ditto, renaming to...
16749 (state_writer::write_state_undefined_type)
16750 (write_state_struct_union_type): ditto, renaming to...
16751 (state_writer::write_state_struct_union_type)
16752 (write_state_struct_type): ditto, renaming to...
16753 (state_writer::write_state_struct_type)
16754 (write_state_user_struct_type): ditto, renaming to...
16755 (state_writer::write_state_user_struct_type)
16756 (write_state_lang_struct_type): ditto, renaming to...
16757 (state_writer::write_state_lang_struct_type)
16758 (write_state_param_struct_type): ditto, renaming to...
16759 (state_writer::write_state_param_struct_type)
16760 (write_state_pointer_type): ditto, renaming to...
16761 (state_writer::write_state_pointer_type)
16762 (write_state_array_type): ditto, renaming to...
16763 (state_writer::write_state_array_type)
16764 (write_state_gc_used): ditto, renaming to...
16765 (state_writer::write_state_gc_used)
16766 (write_state_common_type_content): ditto, renaming to...
16767 (state_writer::write_state_common_type_content)
16768 (write_state_type): ditto, renaming to...
16769 (state_writer::write_state_type)
16770 (write_state_pair_list): ditto, renaming to...
16771 (state_writer::write_state_pair_list)
16772 (write_state_pair): ditto, renaming to...
16773 (state_writer::write_state_pair)
16774 (write_state_typedefs): ditto, renaming to...
16775 (state_writer::write_state_typedefs)
16776 (write_state_structures): ditto, renaming to...
16777 (state_writer::write_state_structures)
16778 (write_state_param_structs): ditto, renaming to...
16779 (state_writer::write_state_param_structs)
16780 (write_state_variables): ditto, renaming to...
16781 (state_writer::write_state_variables)
16782 (write_state_srcdir): ditto, renaming to...
16783 (state_writer::write_state_srcdir)
16784 (write_state_files_list): ditto, renaming to...
16785 (state_writer::write_state_files_list)
16786 (write_state_languages): ditto, renaming to...
16787 (state_writer::write_state_languages)
16788 (write_state): create a state_writer instance and use it when
16789 writing out the state file
16791 2013-05-17 Mike Stump <mikestump@comcast.net>
16793 PR rtl-optimization/57304
16794 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
16795 accessing DF_REF_REAL_LOC.
16797 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16799 PR rtl-optimization/57281
16800 PR rtl-optimization/57300
16801 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
16802 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
16803 what the other splitter did if the registers are dead.
16805 2013-05-17 Richard Biener <rguenther@suse.de>
16807 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
16810 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16812 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
16815 2013-05-17 Marek Polacek <polacek@redhat.com>
16817 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
16818 length when doing non-zero store of storing '\0' to '\0'.
16820 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16822 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
16823 vect_external_def oprnd1 with loop_vinfo, try to emit
16824 optional cast, negation and and stmts on the loop preheader
16825 edge instead of into the pattern def seq.
16827 PR tree-optimization/57051
16828 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
16829 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
16831 2013-05-16 Nick Clifton <nickc@redhat.com>
16833 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
16834 (rl78_is_naked_func): New function.
16835 (rl78_expand_prologue): Skip prologue generation for naked functions.
16836 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
16837 * doc/extend.texi (naked): Add RL78 to the list of processors
16838 that supports this attribute.
16840 2013-05-16 Jeff Law <law@redhat.com>
16842 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
16844 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
16846 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
16847 cache parameters using detect_caches_amd also for CYRIX,
16848 NSC and TM2 signatures.
16850 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
16851 Dzianis Kahanovich <mahatma@eu.by>
16855 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
16856 VIA/Centaur processors and determine their cache parameters
16857 using detect_caches_amd.
16859 2013-05-16 Teresa Johnson <tejohnson@google.com>
16861 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
16862 (rtl_verify_edges): New function.
16863 (rtl_verify_bb_insns): Ditto.
16864 (rtl_verify_bb_pointers): Ditto.
16865 (rtl_verify_bb_insn_chain): Ditto.
16866 (rtl_verify_fallthru): Ditto.
16867 (rtl_verify_bb_layout): Ditto.
16868 (rtl_verify_flow_info_1): Outline checks into new functions.
16869 (rtl_verify_flow_info): Ditto.
16871 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
16873 * cfghooks.c (copy_bbs): Add update_dominance argument.
16874 * cfghooks.h (copy_bbs): Update prototype.
16875 * tree-cfg.c (gimple_duplicate_sese_region):
16876 Add update_dominance argument.
16877 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
16878 * tree-ssa-loop-ch.c (copy_loop_headers): Update
16879 gimple_duplicate_sese_region call.
16880 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
16881 Update copy_bbs call.
16882 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
16883 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
16885 2013-05-16 Jakub Jelinek <jakub@redhat.com>
16887 * tree-vectorizer.h (NUM_PATTERNS): Increment.
16888 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
16889 vect_recog_rotate_pattern.
16890 (vect_recog_rotate_pattern): New function.
16892 2013-05-16 Jason Merrill <jason@redhat.com>
16894 * Makefile.in (LLINKER): New variable.
16895 (mostlyclean): Remove link mutex.
16896 * configure.ac: Handle --enable-link-mutex.
16897 * lock-and-run.sh: New script.
16899 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16902 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
16905 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16907 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
16909 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
16911 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
16912 * config/arm/arm.c (next_consecutive_mem): New function.
16913 (gen_movmem_ldrd_strd): Likewise.
16914 * config/arm/arm.md (movmemqi): Update condition and code.
16915 (unaligned_loaddi, unaligned_storedi): New patterns.
16917 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16919 * config.gcc: Obsolete *-*-solaris2.9*.
16920 * doc/install.texi (Specific, *-*-solaris2*): Document it.
16922 2013-05-16 Richard Biener <rguenther@suse.de>
16924 * passes.c (init_optimization_passes): Move pass_parallelize_loops
16925 earlier, after GRAPHITE transforms and IV canonicalization.
16927 2013-05-16 Jakub Jelinek <jakub@redhat.com>
16929 * omp-low.c (extract_omp_for_data): For collapsed loops,
16930 if at least one of the loops is known at compile time to
16931 iterate zero times, set count to 0.
16932 (expand_omp_regimplify_p): New function.
16933 (expand_omp_for_generic): For collapsed loops, if at least
16934 one of the loops isn't known to iterate at least once,
16935 add runtime check with setting count to 0.
16936 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
16937 For unsigned types if it isn't known at compile time that
16938 the loop will iterate at least once, add runtime check to bypass
16939 the whole loop if initial condition isn't true.
16941 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
16943 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
16945 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
16947 PR middle-end/57286
16948 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
16949 transformations to avoid an infinite loop.
16951 2013-05-16 Marek Polacek <polacek@redhat.com>
16953 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
16955 2013-05-15 Leif Ekblad <leif@rdos.net>
16957 * config/i386/i386.c (ix86_decompose_address): Use
16958 DEFAULT_TLS_SEG_REG to access TLS segment register.
16959 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
16960 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
16961 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
16963 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
16966 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
16967 sibling calls to functions that would normally be lazily bound,
16968 unless $gp is call-clobbered.
16970 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
16972 * config/i386/i386.c (ix86_option_override_internal): Update
16973 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
16974 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
16975 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
16976 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
16978 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
16980 2013-05-15 Andreas Schwab <schwab@suse.de>
16982 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
16983 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
16986 2013-05-15 Teresa Johnson <tejohnson@google.com>
16988 * loop-unroll.c (report_unroll_peel): Check decision before
16989 emitting unroll/peel message.
16991 2013-05-15 Teresa Johnson <tejohnson@google.com>
16993 * function.h (has_bb_partition): New rtl_data flag.
16994 (bb_reorder_complete): Ditto.
16995 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
16996 instead of flag_reorder_blocks_and_partition.
16997 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
16998 with some enhancements.
16999 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
17000 * bb-reorder.c (connect_traces): Check for has_bb_partition
17001 instead of flag_reorder_blocks_and_partition.
17002 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
17003 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
17004 verify_hot_cold_block_grouping.
17005 (partition_hot_cold_basic_blocks): Set has_bb_partition.
17007 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17010 * config/arm/predicates.md (call_insn_operand): New predicate.
17011 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
17012 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
17013 if insn is not a tail call.
17014 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
17016 * config/arm/arm.h (enum reg_class): New caller save register class.
17017 (REG_CLASS_NAMES): Likewise.
17018 (REG_CLASS_CONTENTS): Likewise.
17019 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
17022 2013-05-15 Richard Biener <rguenther@suse.de>
17024 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
17025 of MSG_OPTIMIZED_LOCATIONS.
17026 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
17027 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
17029 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
17030 of MSG_OPTIMIZED_LOCATIONS.
17031 (execute_vect_slp): Likewise.
17032 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
17033 (vect_create_cond_for_alias_checks): Likewise.
17034 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
17035 (vect_recog_widen_mult_pattern): Likewise.
17036 (vect_recog_widen_sum_pattern): Likewise.
17037 (vect_recog_over_widening_pattern): Likewise.
17038 (vect_recog_widen_shift_pattern): Likewise.
17039 (vect_recog_vector_vector_shift_pattern): Likewise.
17040 (vect_recog_divmod_pattern): Likewise.
17041 (vect_recog_mixed_size_cond_pattern): Likewise.
17042 (vect_recog_bool_pattern): Likewise.
17043 (vect_pattern_recog_1): Likewise.
17045 2013-05-15 Martin Jambor <mjambor@suse.cz>
17047 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
17048 non-functions to builtin_unreachable.
17049 * ipa-inline-transform.c (inline_call): Do not assert estimates were
17050 correct when new direct edges were discovered.
17052 2013-05-15 Martin Jambor <mjambor@suse.cz>
17054 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
17055 header, print symbol order instead of node uid, print more information
17056 about indirect edge targets.
17057 (ipa_make_edge_direct_to_target): Print symbol order instead of node
17059 (ipa_make_edge_direct_to_target): Likewise.
17060 (remove_described_reference): Likewise.
17061 (propagate_controlled_uses): Likewise.
17062 (ipa_print_node_params): Also print symbol order.
17063 (ipcp_transform_function): Print symbol order instead of node uids.
17064 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
17065 (cgraph_get_create_real_symbol_node): Likewise.
17066 * ipa-cp.c (print_lattice): Likewise.
17067 (print_all_lattices): Likewise.
17068 (determine_versionability): Likewise.
17069 (initialize_node_lattices): Likewise.
17070 (estimate_local_effects): Likewise.
17071 (update_profiling_info): Likewise.
17072 (create_specialized_node): Likewise.
17073 (perhaps_add_new_callers): Likewise.
17074 (decide_about_value): Likewise.
17075 (decide_whether_version_node): Likewise.
17076 (identify_dead_nodes): Likewise.
17077 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
17078 (dump_inline_summary): Likewise.
17079 (estimate_node_size_and_time): Likewise.
17080 (inline_analyze_function): Likewise.
17081 * ipa-inline.c (report_inline_failed_reason): Likewise.
17082 (want_early_inline_function_p): Likewise.
17083 (edge_badness): Likewise.
17084 (update_edge_key): Likewise.
17085 (inline_small_functions): Likewise. Add dumping of order to two other
17087 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
17088 instead of node uids.
17089 (propagate_pure_const): Likewise.
17090 (propagate_pure_const): Likewise.
17091 * ipa-utils.c (dump_cgraph_node_set): Likewise.
17092 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
17093 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
17095 * tree-pretty-print.c (dump_function_header): Likewise.
17096 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
17097 Print symbol order instead of node uids.
17099 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17101 * config/s390/s390.c (s390_register_move_cost): Don't impose the
17102 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
17104 2013-05-15 Richard Biener <rguenther@suse.de>
17106 PR tree-optimization/57275
17107 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
17108 return value for fail to do runtime alias checks for gather loads.
17110 2013-05-15 Jan Hubicka <jh@suse.cz>
17114 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
17115 weakrefs are not external.
17116 (lto_symtab_merge_decls): Fix thinko when dealing with
17117 non-lto_symtab decls.
17118 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
17119 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
17120 * varpool.c (dump_varpool_node): Dump more flags.
17122 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17124 * config/i386/i386.c (processor_alias_table): Add instruction
17125 FSGSBASE for AMD bdver3 architecture.
17127 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17129 * tree.c (warn_deprecated_use): Print file:line using locus color.
17130 * diagnostic.c (diagnostic_report_current_module): Print file:line
17131 and file:line:column using locus color.
17133 2013-05-14 Mike Stump <mikestump@comcast.net>
17135 * gdbinit.in: Add __null.
17137 2013-05-14 Mike Stump <mikestump@comcast.net>
17139 * recog.h: Rename struct recog_data to Recog_data.
17140 * recog.c: Likewise.
17141 * reload.c (can_reload_into): Likewise.
17142 * config/picochip/picochip.c: Likewise.
17144 2013-05-14 Mike Stump <mikestump@comcast.net>
17146 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
17148 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
17150 * resource.h (struct resources): Remove unch_memory member.
17151 (CLEAR_RESOURCE): Don't clear unch_memory.
17152 * resource.c (mark_referenced_resources): Don't set it.
17153 (mark_set_resources): Likewise.
17154 (mark_target_live_regs): Don't clear it.
17155 (init_resource_info): Likewise.
17156 * reorg.c (resource_conflicts_p): Don't compare it.
17157 (redundant_insn): Don't set it.
17159 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
17161 * emit-rtl.c (next_label): Remove unused function.
17162 (skip_consecutive_labels, link_cc0_insns): Move to ...
17163 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
17164 only place where these functions are used, and make them static.
17166 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17168 * fold-const.c (fold_negate_expr): Handle vectors.
17169 (fold_truth_not_expr): Make it static.
17170 (fold_invert_truthvalue): New static function.
17171 (invert_truthvalue_loc): Handle vectors. Do not call
17172 fold_truth_not_expr directly.
17173 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
17174 <TRUTH_NOT_EXPR>: Do not cast to boolean.
17175 (fold_comparison): Handle vector constants.
17176 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
17177 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
17178 * tree.h (fold_truth_not_expr): Remove declaration.
17180 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
17182 * config/aarch64/aarch64-simd.md
17183 (aarch64_vcond_internal<mode>): Rename to...
17184 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
17185 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
17186 float modes. Clarify all iterator modes.
17187 (vcond<mode><mode>): Use new name for vcond expanders.
17188 (vcond<v_cmp_result><mode>): Likewise.
17189 (vcondu<mode><mode>: Likewise.
17190 * config/aarch64/iterators.md (VDQF_COND): New.
17192 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17195 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
17196 variable for the shift amount. Check that we shift by non-negative
17199 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
17202 * config/arm/arm.h (EPILOGUE_USES): Only return true
17203 for LR_REGNUM after epilogue_completed.
17205 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
17207 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
17208 is error_mark_node.
17210 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17213 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
17214 and Solaris 11+/x86 with gld.
17215 * configure: Regenerate.
17217 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17219 * expmed.c (expand_shift_1): Canonicalize rotates by
17220 constant bitsize / 2 to bitsize - 1.
17221 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
17222 case ROTATERT>: Likewise.
17225 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17227 * config/i386/i386.md (rotateinv): New code attr.
17228 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17229 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17230 roll $31, %eax, etc.
17232 2013-05-14 Richard Biener <rguenther@suse.de>
17234 PR middle-end/57235
17235 * tree-eh.c (sink_clobbers): Give up for successors with
17236 multiple predecessors and no virtual uses.
17238 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
17240 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
17241 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
17243 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17245 PR middle-end/57251
17246 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
17247 the case when both op0 and op1 have VOIDmode.
17249 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17251 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
17252 in multiply-accumulate mode.
17254 2013-05-13 Guozhi Wei <carrot@google.com>
17256 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
17258 2013-05-13 Kai Tietz <ktietz@redhat.com>
17261 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
17262 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
17263 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
17264 * config/i386/i386.c (ix86_option_override_internal): Likewise.
17265 (ix86_expand_prologue): Likewise.
17266 (ix86_expand_split_stack_prologue): Likewise.
17267 (legitimate_pic_address_disp_p): Likewise.
17268 (legitimize_pic_address): Likewise.
17269 (legitimize_tls_address): Likewise.
17270 (legitimize_pe_coff_symbol): Likewise.
17271 (output_pic_addr_const): Likewise.
17272 (construct_plt_address): Likewise.
17273 (ix86_expand_call): Likewise.
17274 (x86_output_mi_thunk): Likewise.
17275 (x86_function_profiler): Likewise.
17277 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
17279 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
17280 similar switch cases.
17281 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
17282 (aarch64_simd_mov_to_<mode>low): Delete.
17283 (aarch64_simd_mov_to_<mode>high): Delete.
17284 (move_lo_quad_<mode>): Add w<-r alternative.
17285 (aarch64_simd_move_hi_quad_<mode>): Likewise.
17286 (aarch64_simd_mov_from_*): Update type attribute.
17287 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
17290 2013-05-13 Jan Hubicka <jh@suse.cz>
17292 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
17293 * config/i386/i386.c (ix86_compute_frame_layout,
17294 ix86_expand_epilogue, emit_i387_cw_initialization,
17295 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
17296 ix86_local_alignment): Fix use of size/speed predicates.
17298 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17300 PR tree-optimization/45216
17301 PR tree-optimization/57157
17302 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
17303 the (-Y) & (B - 1) variant if OP is |.
17304 * expmed.c (expand_shift_1): For rotations by const0_rtx just
17305 return shifted. Use (-op1) & (prec - 1) as other_amount
17306 instead of prec - op1.
17308 2013-05-13 Martin Jambor <mjambor@suse.cz>
17310 PR middle-end/42371
17311 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
17312 (ipa_constant_data): New type.
17313 (ipa_jump_func): Use ipa_constant_data to hold information about
17314 constant jump functions.
17315 (ipa_get_jf_constant): Adjust to jump function type changes.
17316 (ipa_get_jf_constant_rdesc): New function.
17317 (ipa_param_descriptor): New field controlled_uses.
17318 (ipa_get_controlled_uses): New function.
17319 (ipa_set_controlled_uses): Likewise.
17320 * ipa-ref.h (ipa_find_reference): Declare.
17321 * ipa-prop.c (ipa_cst_ref_desc): New type.
17322 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
17324 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
17325 New parameter cs. Adjust all callers.
17326 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
17327 (remove_described_reference): New function.
17328 (jfunc_rdesc_usable): Likewise.
17329 (try_make_edge_direct_simple_call): Decrement controlled use count,
17330 attempt to remove reference if it hits zero.
17331 (combine_controlled_uses_counters): New function.
17332 (propagate_controlled_uses): Likewise.
17333 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
17334 (ipa_edge_duplication_hook): Duplicate reference descriptions.
17335 (ipa_print_node_params): Print described use counter.
17336 (ipa_write_jump_function): Adjust to jump function type changes.
17337 (ipa_read_jump_function): New parameter CS, pass it to
17338 ipa_set_jf_constant. Adjust caller.
17339 (ipa_write_node_info): Stream controlled use count
17340 (ipa_read_node_info): Likewise.
17341 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
17343 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
17344 count. Remove cloning-added reference if it reaches zero.
17345 * ipa-ref.c (ipa_find_reference): New function.
17347 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17349 * config/i386/i386.c (processor_target_table): Modified default
17350 alignment values for AMD BD and BT architectures.
17352 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
17354 * tree-vect-generic.c (uniform_vector_p): Move ...
17355 * tree.c (uniform_vector_p): ... here.
17356 * tree.h (uniform_vector_p): Declare it.
17357 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
17360 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17362 PR tree-optimization/57230
17363 * tree-ssa-strlen.c (handle_char_store): Record length for
17364 array store from STRING_CST.
17366 PR tree-optimization/57230
17367 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
17370 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
17372 * config/epiphany/epiphany.c (epiphany_init): Check size of
17373 NUM_MODES_FOR_MODE_SWITCHING.
17374 (epiphany_expand_prologue):
17375 Remove CONFIG_REGNUM initial value handling code.
17376 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
17377 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
17378 (emit_set_fp_mode, epiphany_mode_after): Likewise.
17379 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
17380 Don't return 1 for FP_MODE_NONE.
17381 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
17382 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
17383 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
17384 * config/epiphany/epiphany.md (save_config): New pattern.
17386 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
17388 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
17390 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
17392 * config/i386/i386.md (memory): Handle sseishft1.
17393 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
17394 (*vec_extractv2di_1): Ditto.
17396 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
17398 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
17401 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
17403 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
17404 Add mthumb/march=armv7-a multilib.
17405 Add mthumb/march=armv7-r multilib.
17406 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
17408 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
17410 * config/v850/t-rtems: Add more multilibs.
17412 2013-05-10 Richard Biener <rguenther@suse.de>
17414 PR tree-optimization/57214
17415 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
17416 not propagate from SSA names that occur in abnormal PHI nodes.
17418 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
17420 * stor-layout.c (element_precision): New function.
17421 * machmode.h (element_precision): Declare it.
17422 * tree.c (build_minus_one_cst): New function.
17423 (element_precision): Likewise.
17424 * tree.h (build_minus_one_cst): Declare new function.
17425 (element_precision): Likewise.
17426 * fold-const.c (operand_equal_p): Use element_precision.
17427 (fold_binary_loc): Handle vector types.
17428 * convert.c (convert_to_integer): Use element_precision.
17429 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
17432 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
17434 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
17435 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
17436 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
17437 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17438 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
17439 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
17440 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
17441 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17442 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
17443 (Uuw8): New constraints.
17444 (Usb4): Move into alphabetical order.
17445 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
17446 (sd8_operand, ub8_operand, uw8_operand): New predicates.
17447 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
17448 previously unnamed patterns.
17449 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
17450 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
17451 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
17452 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
17453 of set_attr_alternative/if_then_else. Use extended_mips16 instead
17454 of specific lengths.
17456 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17458 * config/i386/i386.md (rotateinv): New code attr.
17459 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17460 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17461 roll $31, %eax, etc.
17463 PR tree-optimization/45216
17464 PR tree-optimization/57157
17465 * tree-ssa-forwprop.c (simplify_rotate): New function.
17466 (ssa_forward_propagate_and_combine): Call it.
17468 2013-05-10 Richard Biener <rguenther@suse.de>
17470 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
17471 disable peeling when we version for aliasing.
17472 (vector_alignment_reachable_p): Honor explicit user alignment.
17473 (vect_supportable_dr_alignment): Likewise.
17474 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
17475 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
17476 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
17477 then peeling to arrange for the cost-model check to come first.
17479 2013-05-10 Alan Modra <amodra@gmail.com>
17481 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
17482 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
17483 * configure: Regenerate.
17485 2013-05-10 Alan Modra <amodra@gmail.com>
17488 * varasm.c (default_elf_select_section): Move !DECL_P check..
17489 (get_named_section): ..to here before calling get_section_name.
17491 (default_section_type_flags): Add DECL_P check.
17492 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
17493 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
17495 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
17497 * config/epiphany/epiphany.c (epiphany_expand_prologue):
17498 When using gen_stack_adjust_str with a register offset, add a
17499 REG_FRAME_RELATED_EXPR note.
17501 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
17503 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
17504 (*vec_extractv4si_zext_mem): Ditto.
17505 (*vec_extractv2di): Add 0->x and x->x alternatives.
17506 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
17507 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
17509 2013-05-09 Jason Merrill <jason@redhat.com>
17511 N3639 C++1y VLA support
17512 * gimplify.c (gimplify_vla_decl): Don't touch an existing
17515 * tree.c (build_constructor_va): New.
17516 * tree.h: Declare it.
17518 2013-05-09 Martin Jambor <mjambor@suse.cz>
17521 * gimple-fold.c (canonicalize_constructor_val): Call
17522 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
17524 2013-05-09 Jan Hubicka <jh@suse.cz>
17525 Richard Biener <rguenther@suse.de>
17528 * symtab.c (symtab_make_decl_local): Do not add private names.
17530 2013-05-09 Jan Hubicka <jh@suse.cz>
17533 * symtab.c (insert_to_assembler_name_hash): Handle clones.
17534 (unlink_from_assembler_name_hash): Likewise.
17535 (symtab_prevail_in_asm_name_hash, symtab_register_node,
17536 symtab_unregister_node, symtab_initialize_asm_name_hash,
17537 change_decl_assembler_name): Update.
17539 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
17541 * config/aarch64/aarch64.md: New movtf split.
17542 (*movtf_aarch64): Update.
17543 (aarch64_movdi_tilow): Handle TF modes and rename to
17544 aarch64_movdi_<mode>low.
17545 (aarch64_movdi_tihigh): Handle TF modes and rename to
17546 aarch64_movdi_<mode>high
17547 (aarch64_movtihigh_di): Handle TF modes and rename to
17548 aarch64_mov<mode>high_di
17549 (aarch64_movtilow_di): Handle TF modes and rename to
17550 aarch64_mov<mode>low_di
17551 (aarch64_movtilow_tilow): Remove spurious whitespace.
17552 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
17554 (aarch64_print_operand): Update.
17556 2013-05-09 Alan Modra <amodra@gmail.com>
17558 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
17560 * configure: Regenerate.
17562 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17564 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
17565 splitter preparation statements.
17566 * config/i386/sse.md (*vec_extract* splitters): Ditto.
17567 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
17570 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17572 * gimple-ssa-strength-reduction.c (count_candidates): Change
17573 return value to int.
17574 (analyze_candidates_and_replace): Change type of length to int.
17576 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17578 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
17579 (*vec_extract<mode>): Use VI12_128 mode iterator.
17580 (*vec_extract<mode>_mem): Ditto.
17581 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
17584 2013-05-08 Diego Novillo <dnovillo@google.com>
17589 2012-08-17 Diego Novillo <dnovillo@google.com>
17592 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
17593 * config.in: Regenerate.
17594 * configure: Regenerate.
17595 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
17597 2013-05-08 Jan Hubicka <jh@suse.cz>
17600 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
17601 * cgraph.h (symtab_node_base): Add unique_name.
17602 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17603 input_overwrite_node, input_varpool_node): Stream unique_name.
17604 * cgraphclones.c (cgraph_create_virtual_clone,
17605 cgraph_function_versioning): Set unique_name.
17606 * ipa.c (function_and_variable_visibility): Set unique_name.
17608 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17610 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
17611 (alloc_cand_and_find_basis): Restrict conditional candidate
17612 processing to CAND_MULTs.
17614 2013-05-08 Jan Hubicka <jh@suse.cz>
17617 lto-symtab.c (lto_symtab_symbol_p): New function.
17618 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
17619 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
17620 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
17621 Skip static symbols.
17623 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
17625 PR tree-optimization/57200
17626 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
17627 Only call inform if the preceding warning_at returns true.
17629 2013-05-07 Han Shen <shenhan@google.com>
17631 * cfgexpand.c (record_or_union_type_has_array_p): New function.
17632 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
17633 * common.opt (fstack-protector-strong): New option.
17634 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
17635 * doc/invoke.texi (Optimization Options): Document
17636 "-fstack-protector-strong".
17637 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
17639 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17641 * config/mips/mips.c (mips_machine_reorg2): Return 0.
17643 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
17645 * ira.c (update_equiv_regs): Add insn having equiv memory even if
17646 it is not lhs of the insn.
17647 (setup_reg_equiv): Remove insn having equiv memory which it is not
17649 * lra-constraints.c (process_address): Try to improve generation
17650 code for address base + disp.
17651 (lra_constraints): Make correct the code for checking insn setting
17652 up backward equivalence. Remove insn only if it is in the init
17654 * lra-eliminations.c (update_reg_eliminate): Change return value.
17655 (lra_eliminate): Use the result.
17657 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
17659 * config/i386/sse.md (ssescalarnummask): New mode attribute.
17660 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
17661 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
17662 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
17663 register target operands.
17664 (*vec_extractv8hi_sse2): New pattern.
17665 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
17666 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
17667 (*vec_extract<mode>_mem): New insn and split pattern.
17669 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
17671 * config/arm/arm.c (arm_asan_shadow_offset): New function.
17672 (TARGET_ASAN_SHADOW_OFFSET): Define.
17673 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
17674 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
17676 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17678 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
17679 (incr_vec_index): Return -1 if increment not found.
17680 (create_add_on_incoming_edge): Assert if increment not found.
17681 (record_increment): Limit number of increments recorded.
17682 (all_phi_incrs_profitable): Return false if an increment not found.
17683 (replace_profitable_candidates): Don't process increments that were
17685 (analyze_candidates_and_replace): Limit size of incr_vec.
17687 2013-05-07 Richard Biener <rguenther@suse.de>
17689 * calls.c (special_function_p): setjmp-like functions are leaf.
17690 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
17691 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
17693 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
17695 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
17696 (aarch64_simd_mov<mode>): New expander.
17697 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
17698 (aarch64_simd_mov_to_<mode>high): Likewise.
17699 (aarch64_simd_mov_from_<mode>low): Likewise.
17700 (aarch64_simd_mov_from_<mode>high): Likewise.
17701 (aarch64_dup_lane<mode>): Update.
17702 (aarch64_dup_lanedi): New instruction pattern.
17703 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
17704 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
17706 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17708 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
17709 (replace_mult_candidate): Remove unnecessary argument; remove
17710 unnecessary parameter from call to introduce_cast_before_cand.
17711 (replace_unconditional_candidate): Remove unnecessary parameter
17712 from call to replace_mult_candidate.
17713 (replace_conditional_candidate): Likewise.
17714 (insert_initializers): Use make_temp_ssa_name.
17715 (introduce_cast_before_cand): Remove unnecessary argument; use
17716 make_temp_ssa_name.
17717 (replace_one_candidate): Remove unnecessary argument; remove
17718 unnecessary parameter from calls to introduce_cast_before_cand.
17719 (replace_profitable_candidates): Remove unnecessary parameters
17720 from calls to replace_one_candidate.
17722 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17724 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
17725 phi def as possibly hiding a basis for a CAND_ADD whose operands
17726 have been commuted in the analysis.
17727 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
17729 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17731 * config/aarch64/aarch64.md
17732 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
17733 shift value between 0-4.
17735 2013-05-07 Richard Biener <rguenther@suse.de>
17737 * double-int.h (rshift): New overload.
17738 * double-int.c (rshift): New function.
17739 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
17740 (create_reference_ops_from_ref): Remove.
17741 (vn_reference_insert): Use shared ops for constructing the
17742 reference and copy it.
17744 2013-05-07 Richard Biener <rguenther@suse.de>
17746 PR middle-end/57190
17747 * tree-eh.c (sink_clobbers): Properly propagate
17748 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
17750 2013-05-07 Jakub Jelinek <jakub@redhat.com>
17752 PR tree-optimization/57149
17753 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
17754 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
17755 collect_phi_def_edges, execute_late_warn_uninitialized): Use
17756 uninit_undefined_value_p instead of ssa_undefined_value_p.
17759 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
17760 for modifier == EXPAND_INITIALIZER.
17762 2013-05-07 Anton Blanchard <anton@samba.org>
17764 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
17765 for powerpc64 little endian.
17766 * configure: Regenerate.
17768 2013-05-06 Graham Stott <grahams@btinternet.com>
17770 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
17771 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
17772 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
17773 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
17775 2013-05-06 Graham Stott <grahams@btinternet.com>
17777 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
17778 codes which allow non-lvalues.
17780 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
17782 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
17783 components are all 1s.
17784 (integer_minus_onep): New function.
17785 * tree.h (integer_minus_onep): Declare it.
17786 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
17787 integer_minus_onep instead of integer_all_onesp.
17789 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17792 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
17793 variations of these patterns.
17795 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
17797 * config/i386/i386.md (isa): Add x64_sse4 member.
17798 (enabled): Handle x64_sse4.
17799 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
17800 instruction for 64bit SSE4_1 targets. Update insn attributes.
17801 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
17802 instruction for SSE4_1 targets. Update insn attributes.
17803 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17804 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
17805 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
17807 (*vec_extractv4si): Rename from *sse4_1_pextrd.
17808 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
17809 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
17811 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17814 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
17816 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
17818 * final.c (do_assembler_dialects): Don't handle curly braces and
17819 vertical bar escaped by % as dialect delimiters.
17820 (output_asm_insn): Print curly braces and vertical bar if escaped
17821 by % and ASSEMBLER_DIALECT defined.
17822 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
17823 * doc/tm.texi: Regenerated.
17825 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17827 * config/mips/mips.c: Include tree-pass.h.
17828 (mips_reorg): Split in pre- and post-dbr_schedule parts.
17829 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
17830 (pass_mips_machine_reorg2): New machine specific pass.
17831 (insert_pass_mips_machine_reorg2): New pass plugin definition.
17832 (mips_option_override): Register the new pass.
17833 * rtl.h (cleanup_barriers): Remove prototype.
17834 (dbr_schedule): Likewise.
17835 * jump.c (cleanup_barriers): Make static.
17836 * reorg.c (dbr_schedule): Likewise.
17838 2013-05-06 Richard Biener <rguenther@suse.de>
17840 PR tree-optimization/57185
17841 * tree-parloops.c (add_field_for_reduction): Handle anonymous
17842 SSA names properly.
17844 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
17847 * config/i386/i386.c (add_parameter_dependencies): Add dependence
17848 between "first_arg" and "insn", not "last" and "insn".
17850 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17852 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
17853 (find_candidates_in_block): Re-enable slsr_process_phi.
17854 (create_phi_basis): Fix double counting of candidate adjustment.
17856 2013-05-06 Richard Biener <rguenther@suse.de>
17858 PR middle-end/57147
17859 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
17860 the edge is also fallthru, preserve it and just clear the
17862 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
17863 also complex, preserve that and just clear the fallthru flag.
17864 * tree-inline.c (update_ssa_across_abnormal_edges): Also
17865 update virtual operands.
17867 2013-05-06 Alan Modra <amodra@gmail.com>
17869 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
17870 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
17871 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
17872 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
17873 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17874 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
17876 2013-05-06 Alan Modra <amodra@gmail.com>
17878 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
17879 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
17880 (DEFAULT_ASM_ENDIAN): Define.
17881 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17882 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
17883 Update -K PIC clause from sysv4.h.
17884 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
17885 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
17887 2013-05-06 Alan Modra <amodra@gmail.com>
17889 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
17890 twice for little-endian.
17891 (ashrdi3_no_power, ashrdi3): Support little-endian.
17893 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17896 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
17897 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
17899 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
17900 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
17901 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
17904 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
17905 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17907 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
17908 * config/pa/pa.opt: Make mbig-switch a no-op.
17909 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
17910 (CASE_VECTOR_MODE): Always return SImode.
17911 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
17912 for the !TARGET_BIG_SWITCH case.
17913 * config/pa/pa-linux.h: Likewise.
17914 * config/pa/pa-openbsd.h: Likewise.
17915 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
17916 * config/pa/pa.md (short_jump): Remove define_insn.
17917 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
17918 (casesi0): Remove define_insn.
17919 (type): Remove btable_branch.
17920 (pa_combine_type): Likewise.
17921 (in_nullified_branch_delay): Likewise.
17922 (in_call_delay): Likewise.
17923 (define_delay): Likewise.
17924 (define_insn_reservation "Z3"): Likewise.
17925 (define_insn_reservation "Z4"): Likewise.
17926 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
17927 (pa_adjust_insn_length): Remove adjustment for btable branches.
17928 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
17931 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
17933 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17934 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
17935 Add m->r,x alternatives.
17936 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
17937 splitters using SWI48x mode iterator.
17938 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
17939 TARGET_64BIT. Add m->x alternative.
17940 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
17941 Add o->x alternative. Enable for TARGET_SSE.
17942 (sse_storeq): Remove expander.
17943 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
17944 with memory input operand.
17945 (*vec_extractv2di_1 splitter): New.
17946 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
17947 * config/i386/i386.md (ssevecmodelower): New mode attribute.
17949 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17951 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
17952 (INT_LOWPART): Delete.
17953 (extract_MB): Adjust.
17954 (extract_ME): Adjust.
17955 (print_operand): Adjust.
17957 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17959 * config/rs6000/predicates.md (reg_or_add_cint_operand,
17960 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17961 (reg_or_logical_cint_operand, easy_fp_constant,
17962 logical_const_operand): Delete "CONST_DOUBLE" case.
17963 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
17964 "HOST_BITS_PER_WIDE_INT == 64" test.
17965 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
17966 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
17967 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
17968 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
17970 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
17971 CONST_DOUBLE DImode/VOIDmode case.
17972 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
17973 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17974 CONST_DOUBLE VOIDmode case.
17975 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17976 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
17977 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
17978 Delete CONST_DOUBLE case.
17979 (splitters for mov FMOVE64 const_double): Delete
17980 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17981 "HOST_BITS_PER_WIDE_INT >= 64" test.
17982 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
17984 (mov DI const_double): Delete.
17986 2013-05-04 Jakub Jelinek <jakub@redhat.com>
17988 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
17989 on op shows all bits zero in mode of a lowpart subreg, return zero.
17991 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
17994 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
17995 to save TFmode registers and DImode to save TImode registers for
17996 caller save operations.
17997 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
17998 mark being partially clobbered since they only use the first
18001 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
18002 and TDmode only use the upper 64-bits of each VSX register.
18004 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18006 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
18007 (find_candidates_in_block): Disable slsr_process_phi.
18009 2013-05-03 Guozhi Wei <carrot@google.com>
18011 * coverage.c (coverage_obj_init): Move the construction of gcov
18013 (build_init_ctor): ... here.
18015 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18017 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
18018 (slsr_cand_d): Redefine def_phi.
18019 (stride_status, phi_adjust_status, count_phis_status): New enums.
18020 (find_phi_def): New.
18021 (find_basis_for_base_expr): New.
18022 (find_basis_for_candidate): Handle hidden bases.
18023 (alloc_cand_and_find_basis): Handle phi candidates.
18024 (slsr_process_phi): New.
18025 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
18026 (create_mul_imm_cand): Likewise.
18027 (create_add_ssa_cand): Exclude phi base candidates.
18028 (create_add_imm_cand): Likewise.
18029 (slsr_process_cast): Likewise.
18030 (slsr_process_copy): Likewise.
18031 (find_candidates_in_block): Handle phi candidates.
18032 (dump_candidate): Likewise.
18033 (unconditional_cands): Delete.
18034 (unconditional_cands_with_known_stride_p): Delete.
18035 (phi_dependent_cand_p): New.
18036 (cand_increment): Handle phi-dependent candidates.
18037 (replace_dependent): Delete.
18038 (replace_mult_candidate): New.
18039 (replace_unconditional_candidate): New.
18040 (incr_vec_index): Move to avoid forward reference.
18041 (create_add_on_incoming_edge): New.
18042 (create_phi_basis): New.
18043 (replace_dependents): Delete.
18044 (replace_conditional_candidate): New.
18045 (phi_add_costs): New.
18046 (replace_uncond_cands_and_profitable_phis): New.
18047 (record_increment): Handle phi adjustments.
18048 (record_phi_increments): New.
18049 (record_increments): Handle phi adjustments.
18050 (phi_incr_cost): New.
18051 (lowest_cost_path): Handle phis.
18052 (total_savings): Likewise.
18053 (analyze_increments): Likewise.
18054 (ncd_with_phi): New.
18055 (ncd_of_cand_and_phis): New.
18056 (nearest_common_dominator_for_cands): Handle phi increments.
18057 (all_phi_incrs_profitable): New.
18058 (replace_profitable_candidates): Handle phi-dependent candidates.
18059 (analyze_candidates_and_replace): Likewise.
18061 2013-05-03 Teresa Johnson <tejohnson@google.com>
18064 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
18065 do not exceed REG_BR_PROB_BASE.
18067 2013-05-03 Jeff Law <law@redhat.com>
18069 PR tree-optimization/57144
18070 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
18071 operand of the condition will bit into the new type when eliminating
18072 a cast feeding a condition.
18074 2013-05-03 Jakub Jelinek <jakub@redhat.com>
18076 PR rtl-optimization/57130
18077 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
18078 of COMPARE as in_code to the recursive call if needed.
18080 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
18082 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
18083 (enabled): Handle new members.
18084 * config/i386/sse.md (*vec_concatv2si): Merge from
18085 *vec_concatv2si_sse2 and vec_concatv2si_sse.
18086 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
18088 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
18090 PR tree-optimization/57027
18091 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
18092 for fnms opportunity, check we got the prerequisite kind
18093 of tree / gimple before using accessor functions.
18095 2013-05-03 Richard Biener <rguenther@suse.de>
18097 * double-int.h (lshift): New overload without precision
18098 and arith argument.
18099 (operator *=, operator +=, operator -=): Move ...
18100 * double-int.c (operator *=, operator +=, operator -=): ... here
18101 and implement more efficiently.
18102 (mul_double_with_sign): Remove.
18103 (lshift_double): Adjust to take unsinged shift argument, push
18104 dispatching code to callers.
18105 (mul_double_wide_with_sign): Add early out for callers that
18106 are not interested in high parts or overflow.
18107 (lshift): New function.
18108 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
18109 dispatch code here.
18110 (lrotate, rrotate): Use logical shifts.
18111 * expr.c (get_inner_reference): Use lshift.
18112 * fixed-value.c (do_fixed_divide): Likewise.
18113 * tree-dfa.c (get_ref_base_and_extent): Likewise.
18114 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
18115 (indirect_refs_may_alias_p): Likewise.
18116 (stmt_kills_ref_p_1): Likewise.
18118 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18120 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
18122 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18124 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
18125 scalar form of FABD instruction.
18127 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
18129 * lra-constraints.c (process_alt_operands): Add checking alt
18130 number to choose the best alternative.
18132 2013-05-02 Richard Biener <rguenther@suse.de>
18134 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
18135 bitmap and its handling.
18136 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
18138 2013-05-02 Richard Biener <rguenther@suse.de>
18140 PR middle-end/57140
18141 * tree-inline.c (copy_loops): Properly handle removed loops.
18142 (copy_cfg_body): Mark destination loops for fixup if source
18143 loops needed fixup.
18145 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
18148 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
18149 generating simple_return for naked functions.
18151 2013-05-02 Martin Jambor <mjambor@suse.cz>
18153 PR middle-end/56988
18154 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
18155 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
18157 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
18158 ipa_agg_replacement_value structures.
18159 (known_aggs_to_agg_replacement_list): Likewise.
18160 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
18161 (read_agg_replacement_chain): Likewise.
18162 (ipcp_transform_function): Also check that by_ref flags match.
18164 2013-05-02 Richard Biener <rguenther@suse.de>
18166 * graphds.h (struct graph): Add obstack member.
18167 * graphds.c (new_graph): Initialize obstack and allocate
18169 (add_edge): Allocate edge from the obstack.
18170 (free_graph): Free the obstack instead of all edges and vertices.
18172 2013-05-02 Teresa Johnson <tejohnson@google.com>
18174 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
18176 * cfg.c (update_bb_profile_for_threading): Ditto.
18177 * tree-inline.c (copy_bb): Ditto.
18178 (copy_edges_for_bb): Ditto.
18179 (initialize_cfun): Ditto.
18180 (copy_cfg_body): Ditto.
18181 (expand_call_inline): Ditto.
18182 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
18183 (estimate_node_size_and_time): Ditto.
18184 (inline_merge_summary): Ditto.
18185 * cgraphclones.c (cgraph_clone_edge): Ditto.
18186 (cgraph_clone_node): Ditto.
18187 * sched-rgn.c (compute_dom_prob_ps): Ditto.
18188 (compute_trg_info): Ditto.
18190 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18192 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
18193 S reg when fp attribute set.
18194 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
18196 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18198 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
18200 (*and_one_cmplsi3_compare0_uxtw): Likewise.
18201 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
18202 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
18204 2013-05-02 Richard Biener <rguenther@suse.de>
18206 * tree-scalar-evolution.c (scev_info_hasher): Remove.
18207 (struct instantiate_cache_entry): New type.
18208 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
18209 (struct instantiate_cache_type): New type.
18210 (set_instantiated_value, get_instantiated_value): Remove.
18211 (get_instantiated_value_entry): New function.
18212 (instantiate_scev_name): Use the new cache and adjust.
18213 (instantiate_scev_poly): Adjust.
18214 (instantiate_scev_binary): Likewise.
18215 (instantiate_array_ref): Likewise.
18216 (instantiate_scev_convert): Likewise.
18217 (instantiate_scev_not): Likewise.
18218 (instantiate_scev_3): Likewise.
18219 (instantiate_scev_2): Likewise.
18220 (instantiate_scev_r): Likewise.
18221 (instantiate_scev): Likewise.
18222 (resolve_mixers): Likewise.
18224 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
18227 * lra-constraints.c (best_small_class_operands_num): Remove.
18228 (process_alt_operands): Remove small_class_operands_num. Take
18229 small classes operands into losers and only if the operand is not
18230 matched. Modify debugging output.
18231 (curr_insn_transform): Remove best_small_class_operands_num.
18234 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18236 * config/aarch64/aarch64-builtins.c
18237 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
18238 * config/aarch64/aarch64-simd-builtins.def
18239 (reduc_splus_): Add new modes.
18240 (reduc_uplus_): New.
18241 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
18242 (reduc_uplus_v4sf): Likewise.
18243 (reduc_splus_v4sf): Likewise.
18244 (aarch64_addv<mode>): Likewise.
18245 (reduc_uplus_<mode>): Likewise.
18246 (reduc_splus_<mode>): Likewise.
18247 (aarch64_addvv2di): Likewise.
18248 (reduc_uplus_v2di): Likewise.
18249 (reduc_splus_v2di): Likewise.
18250 (aarch64_addvv2si): Likewise.
18251 (reduc_uplus_v2si): Likewise.
18252 (reduc_splus_v2si): Likewise.
18253 (reduc_<sur>plus_<mode>): New.
18254 (reduc_<sur>plus_v2di): Likewise.
18255 (reduc_<sur>plus_v2si): Likewise.
18256 (reduc_<sur>plus_v4sf): Likewise.
18257 (aarch64_addpv4sf): Likewise.
18258 * config/aarch64/arm_neon.h
18259 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
18260 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
18261 add UNSPEC_SADDV, UNSPEC_UADDV.
18263 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
18265 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18267 * config/aarch64/arm_neon.h
18268 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
18270 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18272 * config/aarch64/aarch64-builtins
18273 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
18275 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18277 * config/aarch64/aarch64-simd-builtins.def
18278 (reduc_smax_): New.
18279 (reduc_smin_): Likewise.
18280 (reduc_umax_): Likewise.
18281 (reduc_umin_): Likewise.
18282 (reduc_smax_nan_): Likewise.
18283 (reduc_smin_nan_): Likewise.
18286 (smax): Update for V2SF, V4SF and V2DF modes.
18289 (smin_nan): Likewise.
18290 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
18291 (<su><maxmin><mode>3): ...This, refactor.
18292 (s<maxmin><mode>3): New.
18293 (<maxmin_uns><mode>3): Likewise.
18294 (reduc_<maxmin_uns>_<mode>): Refactor.
18295 (reduc_<maxmin_uns>_v4sf): Likewise.
18296 (reduc_<maxmin_uns>_v2si): Likewise.
18297 (aarch64_<fmaxmin><mode>: Remove.
18298 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
18300 (vmin<q>_f<32,64>): Likewise.
18301 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18303 (su): Add mappings for smax, smin, umax, umin.
18305 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18306 (FMAXMIN): Rename as...
18307 (FMAXMIN_UNS): ...This.
18309 (fmaxminv): Likewise.
18310 (fmaxmin): Likewise.
18312 (maxmin_uns_op): Likewise.
18314 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18316 * config/aarch64/arm_neon.h
18317 (vac<ge, gt><sd>_f<32, 64>): Rename to...
18318 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
18319 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
18321 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18323 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
18324 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
18326 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18328 * config/aarch64/aarch64-simd.md
18329 (vcond<mode>_internal): Handle special cases for constant masks.
18330 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
18331 (vcondu<mode><mode>): Likewise.
18332 (vcond<v_cmp_result><mode>): New.
18334 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18336 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
18337 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
18338 * config/aarch64/aarch64-simd-builtins.def
18339 (cmeq): Update to BUILTIN_VALLDI.
18344 * config/aarch64/arm_neon.h
18345 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
18346 to builtins or C as appropriate.
18348 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18350 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
18352 (cmhi): Rename to...
18354 * config/aarch64/aarch64-simd.md
18355 (simd_mode): Add SF.
18356 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
18357 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
18358 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
18359 (cstore<mode>_neg): ...This.
18360 * config/aarch64/iterators.md
18362 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
18363 (COMPARISONS): New.
18364 (UCOMPARISONS): Likewise.
18365 (optab): Add missing comparisons.
18371 (VCMP_S): Likewise.
18372 (VCMP_U): Likewise.
18373 (V_cmp_result): Add DF, SF modes.
18374 (v_cmp_result): Likewise.
18376 (vmtype): Likewise.
18377 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
18379 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
18381 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
18382 define_insn to define_insn_and_split.
18383 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
18384 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
18385 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
18386 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
18387 (thumb2_negscc): Likewise.
18389 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18391 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
18393 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18395 * config/arm/thumb2.md: Remove trailing whitespaces.
18397 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
18399 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
18400 Use gen_int_mode rather than GEN_INT.
18402 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
18404 * value-prof.c (stream_in_histogram_value): Remove the strayed
18407 2013-04-30 Richard Biener <rguenther@suse.de>
18409 PR middle-end/57122
18410 * cfghooks.c (split_edge): Properly check for the loop latch edge.
18412 2013-04-30 Richard Biener <rguenther@suse.de>
18414 PR middle-end/57107
18415 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
18417 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
18419 PR rtl-optimization/56957
18420 PR rtl-optimization/57105
18421 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
18422 variable. Use just INSN_UID for determining whether an insn
18423 should be only disconnected from the insn stream.
18424 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
18426 2013-04-30 Jakub Jelinek <jakub@redhat.com>
18428 PR tree-optimization/57104
18429 * tsan.c (instrument_expr): Don't instrument accesses to
18430 DECL_HARD_REGISTER VAR_DECLs.
18432 2013-04-30 Richard Biener <rguenther@suse.de>
18434 * function.h (loops_for_fn): New inline function.
18435 (set_loops_for_fn): Likewise.
18436 * cfgloop.h (place_new_loop): Add struct function parameter.
18437 (get_loop): Likewise.
18438 (get_loops): Likewise.
18439 (number_of_loops): Likewise.
18440 (fel_next): Adjust.
18441 (fel_init): Likewise.
18442 * cfg.c (get_loop_copy): Adjust.
18443 * cfgloop.c (flow_loops_dump): Likewise.
18444 (record_loop_exits): Likewise.
18445 (verify_loop_structure): Likewise.
18446 * cfgloopanal.c (mark_irreducible_loops): Likewise.
18447 (estimate_reg_pressure_cost): Likewise.
18448 (mark_loop_exit_edges): Likewise.
18449 * cfgloopmanip.c (place_new_loop): Likewise.
18450 (add_loop): Likewise.
18451 (duplicate_loop): Likewise.
18452 * graph.c (draw_cfg_nodes): Likewise.
18453 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
18454 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
18455 (extract_affine_chrec): Likewise.
18456 (build_scop_iteration_domain): Likewise.
18457 * graphite.c (graphite_initialize): Likewise.
18458 * ira-build.c (create_loop_tree_nodes): Likewise.
18459 (more_one_region_p): Likewise.
18460 (rebuild_regno_allocno_maps): Likewise.
18461 (mark_loops_for_removal): Likewise.
18462 (mark_all_loops_for_removal): Likewise.
18463 (remove_unnecessary_regions): Likewise.
18464 (ira_build): Likewise.
18465 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
18466 * loop-init.c (fix_loop_structure): Likewise.
18467 (gate_rtl_move_loop_invariants): Likewise.
18468 (gate_rtl_unswitch): Likewise.
18469 (gate_rtl_unroll_and_peel_loops): Likewise.
18470 (rtl_doloop): Likewise.
18471 * lto-streamer-in.c (input_cfg): Likewise.
18472 * lto-streamer-out.c (output_cfg): Likewise.
18473 * modulo-sched.c (sms_schedule): Likewise.
18474 * predict.c (tree_estimate_probability): Likewise.
18475 (tree_estimate_probability_driver): Likewise.
18476 (estimate_loops): Likewise.
18477 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
18478 (move_sese_region_to_fn): Likewise.
18479 (debug_loop_num): Likewise.
18480 * tree-chrec.c (chrec_evaluate): Likewise.
18481 (hide_evolution_in_other_loops_than_loop): Likewise.
18482 (chrec_component_in_loop_num): Likewise.
18483 (reset_evolution_in_loop): Likewise.
18484 (evolution_function_is_invariant_rec_p): Likewise.
18485 * tree-if-conv.c (main_tree_if_conversion): Likewise.
18486 * tree-inline.c (copy_loops): Likewise.
18487 (copy_cfg_body): Likewise.
18488 (tree_function_versioning): Likewise.
18489 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
18490 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
18492 (add_to_evolution_1): Likewise.
18493 (scev_const_prop): Likewise.
18494 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
18495 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
18496 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
18497 (tree_ssa_lim_initialize): Likewise.
18498 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
18499 (verify_loop_closed_ssa): Likewise.
18500 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
18501 (tree_ssa_loop_im): Likewise.
18502 (tree_ssa_loop_unswitch): Likewise.
18503 (tree_vectorize): Likewise.
18504 (check_data_deps): Likewise.
18505 (tree_ssa_loop_ivcanon): Likewise.
18506 (tree_ssa_loop_bounds): Likewise.
18507 (tree_complete_unroll): Likewise.
18508 (tree_complete_unroll_inner): Likewise.
18509 (tree_parallelize_loops): Likewise.
18510 (tree_ssa_loop_prefetch): Likewise.
18511 (tree_ssa_loop_ivopts): Likewise.
18512 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
18513 * tree-vectorizer.c (vectorize_loops): Likewise.
18515 2013-04-29 Mike Frysinger <vapier@gentoo.org>
18517 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
18518 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
18519 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
18520 with EABI_LINK_SPEC.
18522 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18525 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
18528 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
18531 * lra-constraints.c (process_alt_operands): Discourage a bit more
18532 using memory for pseudos. Print cost dump for alternatives.
18533 Modify cost values for conflicts with early clobbers.
18534 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
18536 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18539 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
18541 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18543 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
18544 from/to S register.
18545 (movdi_aarch64): Support LDR/STR from/to D register.
18547 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18549 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
18550 or higher by default.
18552 2013-04-29 Richard Biener <rguenther@suse.de>
18554 PR middle-end/57075
18555 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
18556 even if not adding abnormal edges for calls that can make
18559 2013-04-29 Richard Biener <rguenther@suse.de>
18561 PR middle-end/57103
18562 * tree-cfg.c (move_stmt_op): Fix condition under which to update
18564 (move_stmt_r): Remove redundant checking.
18566 2013-04-29 Teresa Johnson <tejohnson@google.com>
18569 * basic-block.h (apply_scale): New function.
18570 (apply_probability): Use apply_scale.
18571 * gimple-streamer-in.c (input_bb): Ditto.
18572 * lto-streamer-in.c (input_cfg): Ditto.
18573 * lto-cgraph.c (merge_profile_summaries): Ditto.
18574 * tree-optimize.c (execute_fixup_cfg): Ditto.
18575 * tree-inline.c (copy_bb): Update comment to use apply_scale.
18576 (copy_edges_for_bb): Ditto.
18577 (copy_cfg_body): Ditto.
18579 2013-04-29 Tom de Vries <tom@codesourcery.com>
18581 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
18582 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
18583 (tail_merge_optimize): Handle current_loops == NULL.
18585 2013-04-26 Jeff Law <law@redhat.com>
18587 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
18588 (simplify_cond_using_ranges): Generalize code to simplify
18589 COND_EXPRs where one argument is a constant and the other
18590 is an SSA_NAME created by an integral type conversion.
18592 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18594 * config/arm/arm.md (store_minmaxsi): Use only when
18595 optimize_insn_for_size_p.
18597 2013-04-29 Christian Bruel <christian.bruel@st.com>
18600 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
18602 2013-04-29 Richard Biener <rguenther@suse.de>
18604 PR middle-end/57089
18605 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
18606 loop tree make sure to schedule a fixup for the child as well.
18607 (expand_omp_for_generic): Properly add loops.
18608 (expand_omp_for_static_nochunk): Likewise.
18609 (expand_omp_for_static_chunk): Likewise.
18610 (expand_omp_for): For the degenerate case fixup loops.
18611 (expand_omp_sections): Fix default bb placement in loops.
18612 (expand_omp_atomic_pipeline): Properly add loops.
18614 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18616 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
18618 2013-04-29 Tom de Vries <tom@codesourcery.com>
18620 * tree-ssa-tail-merge.c: Update header comment.
18622 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18624 * config/aarch64/arm_neon.h
18625 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
18626 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
18627 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
18628 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
18629 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
18630 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
18631 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
18632 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
18634 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18636 * config/aarch64/aarch64-simd.md
18637 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
18638 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
18639 fix_trunc, fixuns_trunc.
18640 (ftrunc<VDQF:mode>2): New.
18641 * config/aarch64/iterators.md (optab): Add fix, fixuns.
18642 (fix_trunc_optab): New.
18644 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18646 * config/aarch64/aarch64-builtins.c
18647 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
18648 iceilf, lround, iroundf.
18650 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18653 * config/i386/i386.h (enum ix86_tune_indices)
18654 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
18655 New, split from X86_TUNE_INTER_UNIT_MOVES.
18656 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
18657 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
18658 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
18659 (TARGET_INTER_UNIT_MOVES): Remove.
18660 * config/i386/i386.c (initial_ix86_tune_features): Update.
18661 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
18662 (ix86_expand_convert_uns_didf_sse): Use
18663 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18664 (ix86_expand_vector_init_one_nonzero): Ditto.
18665 (ix86_expand_vector_init_interleave): Ditto.
18666 (inline_secondary_memory_needed): Return true for moves from SSE class
18667 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
18668 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
18669 * config/i386/constraints.md (Yi, Ym): Depend on
18670 TARGET_INTER_UNIT_MOVES_TO_VEC.
18671 (Yj, Yn): New constraints.
18672 * config/i386/i386.md (*movdi_internal): Change constraints of
18673 operand 1 from Yi to Yj and from Ym to Yn.
18674 (*movsi_internal): Ditto.
18675 (*movdf_internal): Ditto.
18676 (*movsf_internal): Ditto.
18677 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
18678 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18679 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
18680 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
18681 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
18682 * config/i386/sse.md (movdi_to_sse): Ditto.
18683 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
18684 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
18685 TARGET_INTER_UNIT_MOVES.
18686 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
18687 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
18688 instead of TARGET_INTER_UNIT_MOVES.
18689 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
18690 operand 1 from Yi to Yj and from Ym to Yn.
18692 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18694 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
18695 (float_truncate_hi_): Likewise.
18696 (float_extend_lo_): Likewise.
18697 (float_truncate_lo_): Likewise.
18698 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
18699 (aarch64_float_extend_lo_v2df): Likewise.
18700 (vec_unpacks_hi_v4sf): Likewise.
18701 (aarch64_float_truncate_lo_v2sf): Likewise.
18702 (aarch64_float_truncate_hi_v4sf): Likewise.
18703 (vec_pack_trunc_v2df): Likewise.
18704 (vec_pack_trunc_df): Likewise.
18706 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18708 * config/aarch64/aarch64-builtins.c
18709 (aarch64_fold_builtin): Fold float conversions.
18710 * config/aarch64/aarch64-simd-builtins.def
18711 (floatv2si, floatv4si, floatv2di): New.
18712 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
18713 * config/aarch64/aarch64-simd.md
18714 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
18715 * config/aarch64/iterators.md (FLOATUORS): New.
18716 (optab): Add float, floatuns.
18717 (su_optab): Likewise.
18719 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18721 * config/aarch64/aarch64-builtins.c
18722 (aarch64_builtin_vectorized_function): Use new names for
18724 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
18725 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
18726 (fcvtzu): Split as...
18727 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
18728 (fcvtas): Split as...
18729 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
18730 (fcvtau): Split as...
18731 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
18732 (fcvtps): Split as...
18733 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
18734 (fcvtpu): Split as...
18735 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
18736 (fcvtms): Split as...
18737 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
18738 (fcvtmu): Split as...
18739 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
18740 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
18741 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
18742 (lfrintnusf, lfrintnudf): Likewise.
18743 * config/aarch64/aarch64-simd.md
18744 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
18746 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
18747 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
18748 (fcvt_pattern): Likewise.
18750 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18752 * config/aarch64/aarch64-simd.md
18753 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
18754 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
18756 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18758 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
18759 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
18760 (vrnd<a,m,n,p>_f32): Implement using builtins.
18761 (vrnd<i,x><q>_f<32, 64>): New.
18763 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18765 * config/aarch64/aarch64-builtins.c
18766 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
18767 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
18768 (frintz): Rename to...
18770 (frintp): Rename to...
18772 (frintm): Rename to...
18774 (frinti): Rename to...
18775 (nearbyint): ...this.
18776 (frintx): Rename to...
18778 (frinta): Rename to...
18780 * config/aarch64/aarch64-simd.md
18781 (aarch64_frint<frint_suffix><mode>): Delete.
18782 (<frint_pattern><mode>2): Convert to insn.
18783 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
18784 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
18785 (frint_pattern): Likewise.
18786 (frint_suffix): Likewise.
18788 2013-04-29 Richard Biener <rguenther@suse.de>
18790 PR tree-optimization/57081
18791 * loop-init.c: Include tree-flow.h.
18792 (loop_optimizer_finalize): Free number of iteration estimates.
18793 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
18795 2013-04-29 Jakub Jelinek <jakub@redhat.com>
18797 PR tree-optimization/57083
18798 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
18799 non-singleton shift count range, zero extend low_bound for uns case.
18801 * config/i386/predicates.md (general_vector_operand): New predicate.
18802 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
18803 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
18804 if they aren't nonimmediate operands. If their original values
18805 satisfy const_vector_equal_evenodd_p, don't shift them.
18806 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
18807 predicates. For the SSE4.1 case force operands[{1,2}] into registers
18808 if not nonimmediate_operand.
18809 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
18810 instead of register_operand.
18811 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
18813 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
18815 * stor-layout.c (finalize_size_functions): Allocate a structure and
18816 reset cfun before dumping the functions.
18818 2013-04-27 Jakub Jelinek <jakub@redhat.com>
18820 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
18823 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
18824 use xop_pmacsdqh if uns_p.
18825 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
18826 the immediate rotate count.
18828 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
18830 * rtl.h (struct rtx_def): Add comment for field jump.
18831 (LRA_SUBREG_P): New macro.
18832 * recog.c (register_operand): Check LRA_SUBREG_P.
18833 * lra.c (lra): Add note at the end of RTL code. Align non-empty
18835 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
18836 (lra_final_code_change): Skip subreg change for operators.
18837 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
18838 if there are no operand changes.
18839 * lra-constraints.c (curr_insn_set): New.
18840 (match_reload): Set LRA_SUBREG_P.
18841 (emit_spill_move): Ditto.
18842 (check_and_process_move): Use curr_insn_set. Process only single
18843 set insns. Don't initialize sec_mem_p and change_p.
18844 (simplify_operand_subreg): Use LRA_SUBREG_P.
18845 (reg_in_class_p): New function.
18846 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
18847 of #ifdef. Add code to remove cycling.
18848 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
18849 non-null disp. Reload inner instead of disp when base and index
18850 are null. Try to put lo_sum into register.
18851 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
18852 (check_and_process_move): Move code for move cost check to
18853 simple_move_p. Remove equiv_substitution.
18854 (simple_move_p): New function.
18855 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
18856 curr_insn_set. Call check_and_process_move only for single set
18857 insns. Use the new function. Move call of check_and_process_move
18858 after operand equiv substitution and address process.
18860 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18863 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
18864 with nonlocal goto receivers or returns twice calls, ignore
18865 unininitialized values from abnormal edges to nl goto receiver
18866 or returns twice call.
18868 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18870 PR tree-optimization/57051
18871 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
18872 and VEC_RSHIFT_EXPR if shift count is a multiple of element
18875 2013-04-26 Richard Biener <rguenther@suse.de>
18877 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
18878 (expand_omp_taskreg): Likewise. Mark loops for fixup.
18879 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
18880 (fixup_loop_arrays_after_move): New function.
18881 (move_sese_region_to_fn): Properly outline the loop tree parts
18882 of the SESE region.
18884 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
18886 * config/i386/i386.md (type, unit): Fix long lines.
18888 2013-04-26 Richard Biener <rguenther@suse.de>
18890 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
18891 (lto-streamer-out.o): Likewise.
18892 * cfgloop.c (init_loops_structure): Export, add struct function
18893 argument and adjust.
18894 (flow_loops_find): Adjust.
18895 * cfgloop.h (enum loop_estimation): Add EST_LAST.
18896 (init_loops_structure): Declare.
18897 * lto-streamer-in.c: Include cfgloop.h.
18898 (input_cfg): Input the loop tree.
18899 * lto-streamer-out.c: Include cfgloop.h.
18900 (output_cfg): Output the loop tree.
18901 (output_struct_function_base): Do not drop PROP_loops.
18903 2013-03-26 Richard Biener <rguenther@suse.de>
18905 * tree-cfg.c (execute_build_cfg): Build the loop tree.
18906 (pass_build_cfg): Provide PROP_loops.
18907 (move_sese_region_to_fn): Remove loops that are outlined into fn
18909 * tree-inline.c: Include cfgloop.h.
18910 (initialize_cfun): Do not drop PROP_loops.
18911 (copy_loops): New function.
18912 (copy_cfg_body): Copy loop structure.
18913 (tree_function_versioning): Initialize destination loop tree.
18914 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
18915 (pass_parallelize_loops): Do IL verification.
18916 * loop-init.c (loop_optimizer_init): Fixup loops if required.
18917 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
18918 the CFG make sure we fixup loops as well.
18919 * tree-ssa-tail-merge.c: Include cfgloop.h.
18920 (replace_block_by): When merging loop latches mark loops for fixup.
18921 * lto-streamer-out.c (output_struct_function_base): Drop
18922 PROP_loops for now.
18923 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
18924 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
18925 * ipa-split.c: Include cfgloop.h.
18926 (split_function): Add the new return block to the loop tree root.
18927 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
18928 whether we have removed the forwarder block.
18929 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
18930 * cfgloop.h (place_new_loop): Declare.
18931 * cfgloopmanip.c (place_new_loop): Export.
18932 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
18933 (tree-switch-conversion.o): Likewise.
18934 (tree-complex.o): Likewise.
18935 (tree-inline.o): Likewise.
18936 (tree-ssa-tailmerge.o): Likewise.
18937 (ipa-split.o): Likewise.
18938 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
18939 (tree-ssa-copy.o): Likewise.
18940 * tree-switch-conversion.c: Include cfgloop.h
18941 (process_switch): If we emit a bit-test cascade, schedule loops
18943 * tree-complex.c: Include cfgloop.h.
18944 (expand_complex_div_wide): Properly add new basic-blocks to loops.
18945 * asan.c: Include cfgloop.h.
18946 (create_cond_insert_point): Properly add new basic-blocks to
18947 loops, schedule loop fixup.
18948 * cfgloop.c (verify_loop_structure): Check that looks are not
18950 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
18952 (expand_omp_for_generic): Likewise.
18953 (expand_omp_sections): Likewise.
18954 (expand_omp_atomic_pipeline): Schedule loops for fixup.
18955 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
18956 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
18957 is initialized, not when loops are present.
18958 * tree-parloops.c (parallelize_loops): Remove checking here.
18959 * passes.c (init_optimization_passes): Schedule a copy-propagation
18960 pass before complete unrolling of inner loops.
18962 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18964 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
18965 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
18966 (colorize_init): Add argument to _WIN32 version.
18967 * toplev.c: Include diagnostic-color.h.
18968 (process_options): Default to -fdiagnostics-color=auto if
18969 GCC_COLORS env var is in the environment.
18970 * common.opt (fdiagnostics-color=): Add Var and Init.
18971 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
18972 env var is in the environment, the default is auto rather than never.
18974 * diagnostic.h (file_name_as_prefix): Add context argument.
18975 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
18976 the string as locus.
18977 * langhooks.c (lhd_print_error_function): Adjust caller.
18979 2013-04-25 Lawrence Crowl <crowl@google.com>
18981 * var-tracking.c (shared_hash_def::htab):
18982 Change type to hash_table. Update dependent calls and types.
18984 2013-04-25 Lawrence Crowl <crowl@google.com>
18986 * Makefile.in: Update as needed below.
18988 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
18989 Move declaration to after the type's method definitons.
18991 * attribs.c (htab_t scoped_attributes::attribute_hash):
18992 Change type to hash_table. Update dependent calls and types.
18994 * bitmap.c (htab_t bitmap_desc_hash):
18995 Change type to hash_table. Update dependent calls and types.
18997 * cselib.c (htab_t cselib_hash_table):
18998 Change type to hash_table. Update dependent calls and types.
19000 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
19001 (hash_string_slot_node): Move implementation into lto-streamer.h
19002 struct string_slot_hasher.
19003 (eq_string_slot_node): Likewise.
19005 * data-streamer-out.c: Update output_block::string_hash_table
19006 dependent calls and types.
19008 * dwarf2cfi.c (htab_t trace_index):
19009 Change type to hash_table. Update dependent calls and types.
19011 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
19012 Change type to hash_table. Update dependent calls and types.
19013 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
19014 (htab_t optimize_external_refs::map): Likewise.
19015 (htab_t output_comp_unit::extern_map): Likewise.
19016 (htab_t output_comdat_type_unit::extern_map): Likewise.
19017 (htab_t output_macinfo::macinfo_htab): Likewise.
19018 (htab_t optimize_location_lists::htab): Likewise.
19019 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
19021 * except.c (htab_t ehspec_hash_type):
19022 Change type to hash_table. Update dependent calls and types.
19023 (assign_filter_values::ttypes): Likewise.
19024 (assign_filter_values::ehspec): Likewise.
19025 (sjlj_assign_call_site_values::ar_hash): Likewise.
19026 (convert_to_eh_region_ranges::ar_hash): Likewise.
19028 * gcse.c (htab_t pre_ldst_table):
19029 Change type to hash_table. Update dependent calls and types.
19031 * ggc-common.c (htab_t saving_htab):
19032 Change type to hash_table. Update dependent calls and types.
19033 (htab_t loc_hash): Likewise.
19034 (htab_t ptr_hash): Likewise.
19035 (call_count): Rename ggc_call_count.
19036 (call_alloc): Rename ggc_call_alloc.
19037 (loc_descriptor): Rename make_loc_descriptor.
19038 (add_statistics): Rename ggc_add_statistics.
19040 * ggc-common.c (saving_htab):
19041 Change type to hash_table. Update dependent calls and types.
19043 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
19044 (push_gimplify_context): Likewise.
19045 (pop_gimplify_context): Likewise.
19046 (struct gimple_temp_hash_elt): Added.
19047 (struct gimplify_hasher): Likewise.
19048 (struct gimplify_ctx.temp_htab):
19049 Change type to hash_table. Update dependent calls and types.
19051 * gimple-fold.c: Include gimplify-ctx.h.
19053 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
19054 Change type to hash_table. Update dependent calls and types.
19055 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
19056 avoid potential global name collision.
19058 * gimplify.c: Include gimplify-ctx.h.
19059 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
19060 (htab_t gimplify_ctx::temp_htab):
19061 Update dependent calls and types for new type hash_table.
19062 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
19063 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
19065 * gimplify-ctx.h: New.
19066 (struct gimple_temp_hash_elt): Move from gimplify.c.
19067 (class gimplify_hasher): New.
19068 (struct gimplify_ctx): Move from gimple.h.
19069 (htab_t gimplify_ctx::temp_htab):
19070 Change type to hash_table. Update dependent calls and types.
19072 * graphite-clast-to-gimple.c: Include graphite-htab.h.
19073 (htab_t ivs_params::newivs_index):
19074 Change type to hash_table. Update dependent calls and types.
19075 (htab_t ivs_params::params_index): Likewise.
19076 (htab_t print_generated_program::params_index): Likewise.
19077 (htab_t gloog::newivs_index): Likewise.
19078 (htab_t gloog::params_index): Likewise.
19080 * graphite.c: Include graphite-htab.h.
19081 4htab_t graphite_transform_loops::bb_pbb_mapping):
19082 Change type to hash_table. Update dependent calls and types.
19084 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
19085 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
19086 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
19088 * graphite-dependences.c: Include graphite-htab.h.
19089 (loop_is_parallel_p): Change hash table type of parameter.
19091 * graphite-htab.h: New.
19092 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
19093 (extern find_pbb_via_hash): Move from graphite-poly.h.
19094 (extern loop_is_parallel_p): Move from graphite-poly.h.
19095 (extern get_loop_body_pbbs): Move from graphite-poly.h.
19097 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
19098 (extern loop_is_parallel_p): Move to graphite-htab.h.
19099 (extern get_loop_body_pbbs): Move to graphite-htab.h.
19101 * haifa-sched.c (htab_t delay_htab):
19102 Change type to hash_table. Update dependent calls and types.
19103 (htab_t delay_htab_i2): Likewise.
19105 * ira-color.c (htab_t allocno_hard_regs_htab):
19106 Change type to hash_table. Update dependent calls and types.
19108 * ira-costs.c (htab_t cost_classes_htab):
19109 Change type to hash_table. Update dependent calls and types.
19111 * loop-invariant.c (htab_t merge_identical_invariants::eq):
19112 Change type to hash_table. Update dependent calls and types.
19114 * loop-iv.c (htab_t bivs):
19115 Change type to hash_table. Update dependent calls and types.
19117 * loop-unroll.c (htab_t opt_info::insns_to_split):
19118 Change type to hash_table. Update dependent calls and types.
19119 (htab_t opt_info::insns_with_var_to_expand): Likewise.
19121 * lto-streamer.h (struct string_slot): Move from data-streamer.h
19122 (struct string_slot_hasher): New.
19123 (htab_t output_block::string_hash_table):
19124 Change type to hash_table. Update dependent calls and types.
19126 * lto-streamer-in.c (freeing_string_slot_hasher): New.
19127 (htab_t file_name_hash_table):
19128 Change type to hash_table. Update dependent calls and types.
19130 * lto-streamer-out.c: Update output_block::string_hash_table dependent
19133 * lto-streamer.c (htab_t tree_htab):
19134 Change type to hash_table. Update dependent calls and types.
19136 * omp-low.c: Include gimplify-ctx.h.
19138 * passes.c (htab_t name_to_pass_map):
19139 Change type to hash_table. Update dependent calls and types.
19140 (pass_traverse): Rename to passes_pass_traverse.
19142 * plugin.c (htab_t event_tab):
19143 Change type to hash_table. Update dependent calls and types.
19145 * postreload-gcse.c (htab_t expr_table):
19146 Change type to hash_table. Update dependent calls and types.
19147 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
19149 * sese.c (debug_rename_map_1): Make extern.
19150 (htab_t copy_bb_and_scalar_dependences::rename_map):
19151 Change type to hash_table. Update dependent calls and types.
19153 * sese.h (extern debug_rename_map): Move to .c file.
19155 * store-motion.c (htab_t store_motion_mems_table):
19156 Change type to hash_table. Update dependent calls and types.
19158 * trans-mem.c (htab_t tm_new_mem_hash):
19159 Change type to hash_table. Update dependent calls and types.
19161 * tree-browser.c (htab_t TB_up_ht):
19162 Change type to hash_table. Update dependent calls and types.
19164 * tree-cfg.c (htab_t discriminator_per_locus):
19165 Change type to hash_table. Update dependent calls and types.
19167 * tree-complex.c: Include tree-hasher.h
19168 (htab_t complex_variable_components):
19169 Change type to hash_table. Update dependent calls and types.
19171 * tree-eh.c (htab_t finally_tree):
19172 Change type to hash_table. Update dependent calls and types.
19174 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
19175 struct int_tree_hasher.
19176 (extern int_tree_map_eq): Likewise.
19177 (uid_decl_map_hash): Removed.
19178 (extern decl_tree_map_eq): Likewise.
19180 * tree-hasher.h: New.
19181 (struct int_tree_hasher): New.
19182 (typedef int_tree_htab_type): New.
19184 * tree-inline.c: Include gimplify-ctx.h.
19186 * tree-mudflap.c: Include gimplify-ctx.h.
19188 * tree-parloops.c: Include tree-hasher.h.
19189 (htab_t eliminate_local_variables_stmt::decl_address):
19190 Change type to hash_table. Update dependent calls and types.
19191 (htab_t separate_decls_in_region::decl_copies): Likewise.
19193 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
19194 Change type to hash_table. Update dependent calls and types.
19196 * tree-sra.c (candidates):
19197 Change type to hash_table. Update dependent calls and types.
19199 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
19201 (int_tree_map_hash): Likewise.
19203 * tree-ssa-dom.c (htab_t avail_exprs):
19204 Change type to hash_table. Update dependent calls and types.
19206 * tree-ssa-live.c (var_map_base_init::tree_to_index):
19207 Change type to hash_table. Update dependent calls and types.
19209 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
19210 Change type to hash_table. Update dependent calls and types.
19212 * tree-ssa-phiopt.c (seen_ssa_names):
19213 Change type to hash_table. Update dependent calls and types.
19215 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
19216 Change type to hash_table. Update dependent calls and types.
19218 * tree-ssa-uncprop.c (equiv):
19219 Change type to hash_table. Update dependent calls and types.
19221 2013-04-25 Jakub Jelinek <jakub@redhat.com>
19223 PR rtl-optimization/57003
19224 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
19225 call note_stores with kill_clobbered_value callback again after
19226 killing regs_invalidated_by_call.
19228 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19230 * config/aarch64/aarch64-simd.md
19231 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
19232 (aarch64_simd_bsl<mode>): Likewise.
19233 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
19235 2013-04-25 Marek Polacek <polacek@redhat.com>
19237 PR tree-optimization/57066
19238 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
19240 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
19242 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
19244 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19246 * config/aarch64/aarch64-builtins.c
19247 (aarch64_fold_builtin): New.
19248 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
19249 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
19250 * config/aarch64/aarch64-simd-builtins.def (abs): New.
19251 * config/aarch64/arm_neon.h
19252 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
19254 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19255 Tejas Belagod <tejas.belagod@arm.com>
19257 * config/aarch64/aarch64-builtins.c
19258 (aarch64_gimple_fold_builtin): New.
19259 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
19260 * config/aarch64/aarch64-simd-builtins.def (addv): New.
19261 * config/aarch64/aarch64-simd.md (addpv4sf): New.
19262 (addvv4sf): Update.
19263 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
19265 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19267 * config/aarch64/aarch64.md
19268 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
19270 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19272 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
19273 (*ngcsi_uxtw): New pattern.
19275 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19276 Julian Brown <julian@codesourcery.com>
19278 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
19279 (TB_DREG): Add T_V4HF.
19280 (v4hf_UP): New macro.
19281 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19282 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19283 Handle initialisation of V4HF. Adjust initialisation of reinterpret
19285 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19286 (arm_vector_mode_supported_p): Handle V4HF.
19287 (arm_mangle_map): Handle V4HFmode.
19288 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
19289 * config/arm/arm_neon_builtins.def: Add entries for
19290 vcvtv4hfv4sf, vcvtv4sfv4hf.
19291 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
19292 (neon_vcvtv4hfv4sf): Likewise.
19293 * config/arm/neon-gen.ml: Handle half-precision floating point
19295 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
19296 * config/arm/arm_neon.h: Regenerate.
19297 * config/arm/neon.ml (type elts): Add F16.
19298 (type vectype): Add T_float16x4, T_floatHF.
19299 (type vecmode): Add V4HF.
19300 (type features): Add Requires_FP_bit feature.
19301 (elt_width): Handle F16.
19302 (elt_class): Likewise.
19303 (elt_of_class_width): Likewise.
19304 (mode_of_elt): Refactor.
19305 (type_for_elt): Handle F16, fix error messages.
19306 (vectype_size): Handle T_float16x4.
19307 (vcvt_sh): New function.
19308 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
19309 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
19310 (string_of_mode): Handle V4HF.
19311 * doc/arm-neon-intrinsics.texi: Regenerate.
19313 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19315 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
19316 format specifier in 'X' case.
19318 2013-04-25 Alan Modra <amodra@gmail.com>
19321 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
19322 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
19323 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
19324 Repeat for many other rotate/shift and mask patterns using subregs.
19325 Name lshiftrt insns.
19326 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
19327 on WORDS_BIG_ENDIAN.
19329 2013-04-25 Alan Modra <amodra@gmail.com>
19331 * config.gcc: Support little-endian powerpc-linux targets.
19332 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
19333 (LINK_OS_LINUX_SPEC): Define.
19334 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
19335 Preserve MASK_LITTLE_ENDIAN.
19336 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
19337 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
19338 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
19339 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
19340 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
19341 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
19342 Correct fp word order for little-endian. Don't shift toc entries
19343 smaller than a word for little-endian.
19344 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
19345 (bswapdi2 splits): Correct low-part subreg for little-endian.
19346 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
19347 low/high where such is correct only for be.
19348 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
19349 little-endian for -mcall-aixdesc.
19351 2013-04-25 Alan Modra <amodra@gmail.com>
19353 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
19354 replace_equiv_address_nv.
19356 2013-04-25 Alan Modra <amodra@gmail.com>
19358 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
19360 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19363 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19364 * rtl.h (struct rtx_def): ...
19366 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19368 PR rtl-optimizations/57046
19369 * lra-constraints (split_reg): Set up lra_risky_transformations_p
19370 for multi-reg splits.
19372 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
19374 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
19376 2013-04-24 Sterling Augustine <saugustine@google.com>
19378 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
19379 (comp_dir_string, debug_str_dwo_section): New.
19380 (DEBUG_STR_DWO_SECTION): Rename to ...
19381 (DEBUG_DWO_STR_SECTION): ... this.
19382 (DEBUG_NORM_STR_SECTION): Delete.
19383 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
19384 (DEBUG_STR_DWO_SECTION_FLAGS): New.
19385 (find_AT_string): Move most logic to ...
19386 (find_AT_string_in_table): ... here. New.
19387 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
19388 add_skeleton_AT_string. Delete logic.
19389 (output_skeleton_debug_sections): Remove call to
19390 add_top_level_skeleton_die_attrs.
19391 (add_comp_dir_attribute): Move logic to comp_dir_string.
19392 (dwarf2out_init): Initialize debug_str_dwo_section.
19393 (output_indirect_string): Call find_string_form.
19394 (output_indirect_strings): Rewrite.
19395 (prune_unused_types): Empty skeleton_debug_str_hash.
19396 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
19397 (dwarf2out_finish): Call output_indirect_strings.
19399 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19401 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
19403 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19405 * rtl.h (struct rtx_def): Add comment for field jump.
19406 (LRA_SUBREG_P): New macro.
19407 * recog.c (register_operand): Check LRA_SUBREG_P.
19408 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19410 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19411 (lra_final_code_change): Skip subreg change for operators.
19412 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19413 if there are no operand changes.
19414 * lra-constraints.c (curr_insn_set): New.
19415 (match_reload): Set LRA_SUBREG_P.
19416 (emit_spill_move): Ditto.
19417 (check_and_process_move): Use curr_insn_set. Process only single
19418 set insns. Don't initialize sec_mem_p and change_p.
19419 (simplify_operand_subreg): Use LRA_SUBREG_P.
19420 (reg_in_class_p): New function.
19421 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19422 of #ifdef. Add code to remove cycling.
19423 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19424 non-null disp. Reload inner instead of disp when base and index
19425 are null. Try to put lo_sum into register.
19426 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19427 (check_and_process_move): Move code for move cost check to
19428 simple_move_p. Remove equiv_substitution.
19429 (simple_move_p): New function.
19430 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19431 curr_insn_set. Call check_and_process_move only for single set
19432 insns. Use the new function. Move call of check_and_process_move
19433 after operand equiv substitution and address process.
19435 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
19437 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
19438 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
19439 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
19441 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19443 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
19445 2013-04-24 Marek Polacek <polacek@redhat.com>
19447 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
19448 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
19449 (select_loops_exit_conditions): Likewise.
19450 (number_of_iterations_for_all_loops): Likewise.
19451 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
19452 (scev_analysis): Likewise.
19454 2013-04-02 Catherine Moore <clm@codesourcery.com>
19455 Chao-ying Fu <fu@mips.com>
19457 * config/mips/micromips.md (jraddiusp): New pattern.
19458 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
19459 instruction if possible.
19461 2013-04-24 Alan Modra <amodra@gmail.com>
19463 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
19465 2013-04-24 Julian Brown <julian@codesourcery.com>
19466 Chung-Lin Tang <cltang@codesourcery.com>
19468 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
19469 dependency behavior in enumeration type DIE generation. Add TODO note
19470 to comments about future DW_FORM_sdata/udata re-work of related code.
19472 2013-04-23 Lawrence Crowl <crowl@google.com>
19474 * Makefile.in: Update as needed below.
19476 * hash-table.h (class hash_table):
19477 Correct many methods with parameter types compare_type to the correct
19478 value_type. (Correct code was unlikely to notice the change.)
19479 (hash_table::elements_with_deleted) New.
19480 (class hashtable::iterator): New.
19481 (hashtable::begin()): New.
19482 (hashtable::end()): New.
19483 (FOR_EACH_HASH_TABLE_ELEMENT): New.
19485 * statistics.c (statistics_hashes):
19486 Change type to hash_table. Update dependent calls and types.
19488 * tree-into-ssa.c (var_infos):
19489 Change type to hash_table. Update dependent calls and types.
19491 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
19492 Change type to hash_table. Update dependent calls and types.
19494 * tree-ssa-loop-im.c (struct mem_ref.refs):
19495 Change type to hash_table. Update dependent calls and types.
19497 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
19498 Change type to hash_table. Update dependent calls and types.
19500 * tree-ssa-sccvn.c (vn_tables_s::nary):
19501 Change type to hash_table. Update dependent calls and types.
19502 (vn_tables_s::phis): Likewise.
19503 (vn_tables_s::references): Likewise.
19505 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
19506 (vn_reference_eq): Update parameter and return types.
19508 * tree-ssa-structalias.c (pointer_equiv_class_table):
19509 Change type to hash_table. Update dependent calls and types.
19510 (location_equiv_class_table): Likewise.
19512 * tree-vect-data-refs.c: Consequential changes for making
19513 peeling a hash_table.
19515 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
19516 (destroy_loop_vec_info): Dependent hash_table update.
19518 * tree-vectorizer.h (peeling_htab):
19519 Change type to hash_table. Update dependent calls and types.
19521 2013-04-23 Shiva Chen <shiva0217@gmail.com>
19523 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
19524 to check the register content is equal or not.
19525 * lra-constraints.c (match_reload): Use lra_assign_reg_val
19526 to assign register content record.
19527 * lra-eliminations.c (update_reg_eliminate): Use
19528 lra_update_reg_val_offset to update register content offset.
19529 * lra-int.h (struct lra_reg): Add offset member.
19530 (lra_reg_val_equal_p): New static inline function.
19531 (lra_update_reg_val_offset): New static inline function.
19532 (lra_assign_reg_val): New static inline function.
19533 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
19534 to assign register content record.
19535 (initialize_lra_reg_info_element): Initial offset to zero.
19537 2013-04-23 Catherine Moore <clm@codesourcery.com>
19539 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
19540 operands. Record compression.
19542 2013-04-23 Xinliang David Li <davidxl@google.com>
19544 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
19546 2013-04-23 Richard Biener <rguenther@suse.de>
19548 PR middle-end/57036
19549 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
19550 parameter, only add abnormal goto edges from the copied body
19551 if the call could perform abnormal gotos.
19552 (copy_cfg_body): Adjust.
19554 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
19556 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
19558 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
19560 * coretypes.h (gimple_stmt_iterator): Add struct to make
19563 2013-04-23 Richard Biener <rguenther@suse.de>
19565 PR tree-optimization/57026
19566 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
19567 from SSA names occuring in abnormal PHI nodes.
19569 2013-04-22 Andi Kleen <ak@linux.intel.com>
19571 * lto/lto.c (print_lto_report_1): Fix LTO report names.
19573 2013-04-22 Andi Kleen <ak@linux.intel.com>
19575 * lto/lto.c (print_lto_report_1): Declare early.
19576 (read_cgraph_and_symbols): Call print_lto_report_1 early.
19578 2013-04-22 Andi Kleen <ak@linux.intel.com>
19580 * common.opt (-flto-report-wpa): Add.
19581 * doc/invoke.texi (-flto-report-wpa): Add.
19582 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
19585 2013-04-22 Xinliang David Li <davidxl@google.com>
19587 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
19588 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
19589 * Makefile.in: New dependency
19591 David Daney <ddaney.cavm@gmail.com>
19593 * configure.ac (gcc_cv_as_micromips_support): Use the
19594 --fatal-warnings option.
19595 * configure: Regenerate.
19597 2013-04-22 Marek Polacek <polacek@redhat.com>
19600 * tsan.c (instrument_expr): Don't instrument expression
19601 in case its size is zero.
19603 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
19607 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19609 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19611 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19613 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
19614 (gimple_stmt_iterator): New typedef.
19615 * gimple.h (gimple_stmt_iterator): Rename to...
19616 (gimple_stmt_iterator_d): ... This.
19617 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
19618 trees be valid for GIMPLE and GENERIC.
19619 (TARGET_GIMPLE_FOLD_BUILTIN): New.
19620 * gimple-fold.c (gimple_fold_call): Call target hook
19621 gimple_fold_builtin.
19622 * hooks.c (hook_bool_gsiptr_false): New.
19623 * hooks.h (hook_bool_gsiptr_false): New.
19624 * target.def (fold_stmt): New.
19625 * doc/tm.texi: Regenerate.
19627 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
19630 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
19631 a set sp if no stack realignment.
19633 2013-04-22 Nick Clifton <nickc@redhat.com>
19635 * config.gcc (tilegx-linux): Extend extra_objs rather than
19637 (tilepro-linux): Likewise.
19639 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19641 * config/aarch64/aarch64-builtins.c
19643 (CF0, CF1, CF2, CF3, CF4, CF10): New.
19644 (VAR<1-12>): Add MAP parameter.
19645 (BUILTIN_*): Likewise.
19646 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
19647 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
19648 (aarch64_ushl_n<mode>): Likewise.
19649 (aarch64_sshr_n<mode>): Likewise.
19650 (aarch64_ushr_n<mode>): Likewise.
19651 (aarch64_<maxmin><mode>): Likewise.
19652 (aarch64_sqrt<mode>): Likewise.
19653 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
19654 (vshr<q>_n_*): Likewise.
19656 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19658 * config/aarch64/aarch64-builtins.c
19659 (aarch64_simd_builtin_type_mode): Handle SF types.
19661 (BUILTIN_GPF): Define.
19662 (aarch64_init_simd_builtins): Handle SF types.
19663 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
19664 (frecps): Likewise.
19665 (frecpx): Likewise.
19666 * config/aarch64/aarch64-simd.md
19667 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
19668 (aarch64_frecpe<mode>): New.
19669 (aarch64_frecps<mode>): Likewise.
19670 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
19671 (v8type): Add frecp<esx>.
19672 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
19673 (aarch64_frecps<mode>): Likewise.
19674 * config/aarch64/iterators.md (FRECP): New.
19675 (frecp_suffix): Likewise.
19676 * config/aarch64/arm_neon.h
19677 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
19679 2013-04-22 Christian Bruel <christian.bruel@st.com>
19682 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
19683 (REG_CLASS_NAMES): Idem.
19684 (REG_CLASS_CONTENTS): Idem.
19685 (REGCLASS_HAS_FP_REG): Idem.
19686 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
19687 (sh_conditional_register_usage): Idem.
19689 2013-04-21 Jeff Law <law@redhat.com>
19691 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
19692 (ssa_forward_propagate_and_combine): Use it.
19694 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19696 * lra.c: Update the flow chart diagram.
19698 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19700 PR rtl-optimization/56847
19701 * lra-constraints.c (process_alt_operands): Discourage alternative
19702 with non-matche doffsettable memory constraint fro memory with
19705 2013-04-19 Richard Biener <rguenther@suse.de>
19707 PR tree-optimization/56982
19708 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
19710 * gimplify.c (gimplify_call_expr): Notice special calls.
19711 (gimplify_modify_expr): Likewise.
19712 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
19713 abnormal control flow receivers.
19714 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
19715 in the same way as cfun->has_nonlocal_labels.
19716 (gimple_purge_dead_abnormal_call_edges): Likewise.
19717 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
19718 receivers start a basic-block.
19720 2013-04-19 Richard Biener <rguenther@suse.de>
19722 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
19724 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
19725 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
19726 (SLP_TREE_LOAD_PERMUTATION): Add.
19727 (vect_transform_slp_perm_load): Adjust prototype.
19728 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
19729 (vect_free_slp_instance): Likewise.
19730 (vect_create_new_slp_node): Likewise.
19731 (vect_supported_slp_permutation_p): Remove.
19732 (vect_slp_rearrange_stmts): Adjust.
19733 (vect_supported_load_permutation_p): Likewise. Inline
19734 vect_supported_slp_permutation_p here.
19735 (vect_analyze_slp_instance): Compute load permutations per
19736 slp node instead of per instance.
19737 (vect_get_slp_defs): Adjust.
19738 (vect_transform_slp_perm_load): Likewise.
19739 (vect_schedule_slp_instance): Remove redundant code.
19740 (vect_schedule_slp): Remove hack for PR56270, add it ...
19741 * tree-vect-stmts.c (vectorizable_load): ... here, do not
19742 CSE loads for SLP. Adjust.
19744 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19746 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
19747 spelling in two comments.
19749 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19752 * config/arm/arm.c (load_multiple_sequence): Require SP
19753 as base register for loads if SP is in the register list.
19755 2013-04-19 Martin Jambor <mjambor@suse.cz>
19757 PR tree-optimization/56718
19758 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
19759 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
19760 and made public. Adjusted all callers.
19761 (ipa_intraprocedural_devirtualization): New function.
19762 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
19763 (ipa_intraprocedural_devirtualization): Likewise.
19764 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
19766 2013-04-19 Richard Biener <rguenther@suse.de>
19768 PR tree-optimization/57000
19769 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
19771 2013-04-19 Terry Guo <terry.guo@arm.com>
19773 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
19775 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
19776 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
19777 (cortex_m4_fmacs): Use new reservations.
19778 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
19780 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
19782 PR rtl-optimization/56999
19783 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
19785 (lra_coalesce): Remove split_origin_bitmap and related code.
19786 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
19787 ranges if necessary.
19789 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
19791 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
19793 (ix86_expand_call): Remove clobbered_registers array and use
19794 x86_64_ms_sysv_extra_clobbered_registers instead.
19795 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
19797 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
19799 * config/i386/i386.md (*call_rex64_ms_sysv): Use
19800 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
19801 (*call_value_rex64_ms_sysv): Ditto.
19803 2013-04-18 Cary Coutant <ccoutant@google.com>
19805 * dwarf2out.c (output_pubnames): Check die_perennial_p of
19806 parent instead of die_mark.
19808 2013-04-18 Diego Novillo <dnovillo@google.com>
19810 * gimple.c (create_gimple_tmp): New.
19811 (get_expr_type): New.
19812 (build_assign): New.
19813 (build_type_cast): New.
19814 * gimple.h (enum ssa_mode): Define.
19815 (gimple_seq_set_location): New.
19816 * asan.c (build_check_stmt): Change some gimple_build_* calls
19817 to use build_assign and build_type_cast.
19819 2013-04-18 Richard Biener <rguenther@suse.de>
19821 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
19822 handle negative step. Remove redundant checks.
19823 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
19824 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
19825 for negative step and grouped loads fail to vectorize.
19827 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19829 * emit-rtl.c (reset_insn_used_flags): New function.
19830 (reset_all_used_flags): Use it.
19831 (verify_insn_sharing): New function.
19832 (verify_rtl_sharing): Fix verification for SEQUENCEs.
19834 2013-04-18 Jakub Jelinek <jakub@redhat.com>
19836 PR tree-optimization/56984
19837 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
19838 and (x >> M) >= N don't register any assertion if N << M is the
19841 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19843 * lower-subreg.c (resolve_simple_move): If called self-recursive,
19844 do not delete_insn insns that have not yet been emitted, only
19845 unlink them with remove_insn.
19846 * df-scan.c (df_insn_delete): Revert r197492.
19848 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
19850 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
19851 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
19853 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
19855 * config/arm/arm.md (movsicc_insn): Convert define_insn into
19856 define_insn_and_split.
19857 (and_scc,ior_scc,negscc): Likewise.
19858 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
19860 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
19862 * config/arm/arm.c (use_return_insn): Return 0 for targets that
19863 can benefit from using a sequence of LDRD instructions in epilogue
19864 instead of a single LDM instruction.
19866 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
19869 * doc/extend.texi: Fix typo.
19871 2013-04-17 Richard Biener <rguenther@suse.de>
19873 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
19874 (vect_build_slp_tree): ... here.
19875 (vect_build_slp_tree_1): Compute which stmts of the SLP group
19876 match. Remove special-casing of mismatched complex loads.
19877 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
19878 re-try the match with swapped commutative operands.
19879 (vect_supported_load_permutation_p): Remove special-casing of
19880 mismatched complex loads.
19881 (vect_analyze_slp_instance): Adjust.
19883 2013-04-17 Richard Biener <rguenther@suse.de>
19885 PR rtl-optimization/56921
19886 * cfgloop.h (struct loop): Add simple_loop_desc member.
19887 (struct niter_desc): Mark with GTY(()).
19888 (simple_loop_desc): Do not use aux field but simple_loop_desc.
19889 * loop-iv.c (get_simple_loop_desc): Likewise.
19890 (free_simple_loop_desc): Likewise.
19893 2013-04-16 Richard Biener <rguenther@suse.de>
19895 PR rtl-optimization/56921
19896 * loop-init.c (pass_rtl_move_loop_invariants): Add
19897 TODO_do_not_ggc_collect to todo_flags_finish.
19898 (pass_rtl_unswitch): Same.
19899 (pass_rtl_unroll_and_peel_loops): Same.
19900 (pass_rtl_doloop): Same.
19902 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
19904 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
19905 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
19906 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
19908 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
19909 * tree-streamer.c (record_common_node): Adjust reference in comment.
19911 2013-04-17 Terry Guo <terry.guo@arm.com>
19913 * config/arm/cortex-m4.md: Add a new bypass.
19915 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19917 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
19919 (*subs_<optab><mode>_multp2): New pattern.
19920 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19921 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19923 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19925 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
19926 (*subs_mul_imm_<mode>): New pattern.
19928 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
19931 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
19932 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
19933 (vsx_movti_32bit): Change j->wa to O->wa.
19935 2013-04-16 Richard Biener <rguenther@suse.de>
19937 PR rtl-optimization/56921
19938 * loop-init.c (pass_rtl_move_loop_invariants): Add
19939 TODO_do_not_ggc_collect to todo_flags_finish.
19940 (pass_rtl_unswitch): Same.
19941 (pass_rtl_unroll_and_peel_loops): Same.
19942 (pass_rtl_doloop): Same.
19944 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
19946 * config/arm/arm.c (emit_multi_reg_push): New declaration
19947 for an existing function.
19948 (arm_emit_strd_push): New function.
19949 (arm_expand_prologue): Used here.
19950 (arm_emit_ldrd_pop): New function.
19951 (arm_expand_epilogue): Used here.
19952 (arm_get_frame_offsets): Update condition.
19953 (arm_emit_multi_reg_pop): Add a special case for load of a single
19954 register with writeback.
19956 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
19958 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
19961 2013-04-16 Richard Biener <rguenther@suse.de>
19963 PR tree-optimization/56756
19964 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
19965 (first_mem_ref_loc): New.
19966 (execute_sm): Place the load temporarily before a previous
19967 access instead of in the latch edge to ensure its SSA dependencies
19968 are defined at points dominating the load.
19970 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
19972 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
19973 correct fix by moving header and footer insn to the footer of
19974 the merged basic block. Clear BB_END of the merged-away block.
19976 PR middle-end/43631
19977 * emit-rtl.c (make_note_raw): New function.
19978 (link_insn_into_chain): New static inline function.
19979 (add_insn): Use it.
19980 (add_insn_before, add_insn_after): Factor insn chain linking code...
19981 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
19982 using link_insn_into_chain.
19983 (note_outside_basic_block_p): New helper function for emit_note_after
19984 and emit_note_before.
19985 (emit_note_after): Use nobb variant of add_insn_after if the note
19986 should not be contained in a basic block.
19987 (emit_note_before): Use nobb variant of add_insn_before if the note
19988 should not be contained in a basic block.
19989 (emit_note_copy): Use make_note_raw.
19990 (emit_note): Likewise.
19991 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
19992 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
19993 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
19994 the moved barrier the tail of the basic block it follows.
19995 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
19997 2013-04-15 Jakub Jelinek <jakub@redhat.com>
19999 PR tree-optimization/56962
20000 * gimple-ssa-strength-reduction.c (record_increment): Only set
20001 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
20002 either rhs1 or rhs2 is equal to c->base_expr.
20004 2013-04-15 Richard Biener <rguenther@suse.de>
20006 PR tree-optimization/56933
20007 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
20009 (GROUP_READ_WRITE_DEPENDENCE): Remove.
20010 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
20011 * tree-vect-data-refs.c (vect_analyze_group_access): Move
20012 dependence check ...
20013 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
20015 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
20016 GROUP_READ_WRITE_DEPENDENCE.
20018 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20020 * emit-rtl.c (reset_all_used_flags): New function.
20021 (verify_rtl_sharing): Call reset_all_used_flags before and after
20022 performing the checks.
20024 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20026 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
20027 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
20028 * config/arm/constraints.md (De): New constraint.
20029 * config/arm/neon.md (anddi3_neon): Delete.
20030 (neon_vand<mode>): Expand to standard anddi3 pattern.
20031 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
20032 Move earlier in the file.
20033 (neon_inv_logic_op2): Likewise.
20034 (arm_anddi_operand_neon): New predicate.
20036 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20038 * configure.ac (gcc_cv_ld_as_needed): Set
20039 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
20040 Use -z ignore, -z record on *-*-solaris2*.
20041 (HAVE_LD_AS_NEEDED): Update comment.
20042 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
20043 * configure: Regenerate.
20044 * config.in: Regenerate.
20045 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
20046 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
20047 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
20048 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
20049 equivalents. Fix markup.
20050 * doc/tm.texi: Regenerate.
20052 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
20054 * config/i386/i386.opt: New option mstack-protector-guard=.
20055 * config/i386/i386-opts.h: Add enum stack_protector_guard.
20056 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
20057 TARGET_SSP_TLS_GUARD.
20058 * config/i386/i386.c (ix86_option_override_internal): Set
20059 ix86_stack_protector_guard.
20060 * config/i386/i386.md (stack_protect_set): Enable for
20061 TARGET_SSP_TLS_GUARD only.
20062 (stack_protect_set_<mode>): Ditto.
20063 (stack_protect_test): Ditto.
20064 (stack_protect_test_<mode>): Ditto.
20065 * doc/invoke.texi (i386 Option): Document.
20067 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
20070 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
20071 (S_MODES): Set H_MODE bit.
20072 (SF_MODES): Set only S_MODE and SF_MODE bits.
20073 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
20074 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
20075 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
20076 <MODE_FLOAT>: Likewise.
20078 2013-04-15 Joey Ye <joey.ye@arm.com>
20080 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
20082 2013-04-15 Joey Ye <joey.ye@arm.com>
20084 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
20086 (thumb_far_jump_used_p): Count instruction size and set
20089 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
20091 * reorg.c (fill_simple_delay_slots): Reindent block of code.
20092 * resource.c (mark_target_live_regs): Reformat conditional block.
20094 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20096 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
20097 notes, they are emitted only just before final.
20098 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
20100 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20102 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
20103 * cfgrtl.c (delete_insn): Call it here instead.
20104 * lra-spills.c (lra_final_code_change): Use delete_insn.
20105 * haifa-sched.c (sched_remove_insn): Likewise.
20106 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
20107 returning to the nop pool.
20108 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
20109 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
20111 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20113 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
20114 * doc/tm.texi: Regenerated.
20116 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
20118 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
20121 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20123 * df-core.c (df_find_def): Compare register numbers.
20124 (df_find_use): Likewise.
20126 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
20129 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
20130 lra_in_progress for return.
20132 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20134 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
20135 define_insn into define_insn_and_split and emit movsicc patterns.
20137 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20139 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
20141 2013-04-12 Richard Biener <rguenther@suse.de>
20143 * tree-pass.h (TODO_do_not_ggc_collect): New.
20144 * passes.c (execute_one_ipa_transform_pass): Honor
20145 TODO_do_not_ggc_collect.
20146 (execute_one_pass): Likewise.
20149 2013-04-10 Richard Biener <rguenther@suse.de>
20151 * passes.c (init_optimization_passes): Remove reload pass.
20152 * ira.c (do_reload): Merge into ...
20154 (rest_of_handle_reload): Remove.
20155 (pass_reload): Likewise.
20156 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20157 of reload for vzeroupper pass placement.
20159 2013-04-12 Jakub Jelinek <jakub@redhat.com>
20161 PR tree-optimization/56918
20162 PR tree-optimization/56920
20163 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
20164 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
20165 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
20166 use wide_mul_with_sign method.
20168 2013-04-12 Richard Biener <rguenther@suse.de>
20170 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
20171 not be considered a gimple constant.
20173 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
20175 * fold-const.c (const_binop): Handle vector shifts by a scalar.
20176 (fold_binary_loc): Call const_binop also for mixed vector-scalar
20179 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
20180 Jakub Jelinek <jakub@redhat.com>
20182 * opts.c: Include diagnostic-color.h.
20183 (common_handle_option): Handle OPT_fdiagnostics_color_.
20184 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
20185 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
20186 (diagnostic-color.o): New.
20187 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
20188 (diagnostic_color_rule): New enum.
20189 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
20190 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
20191 the location string.
20192 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
20193 either NULL, or color kind.
20194 * diagnostic-color.c: New file.
20195 * diagnostic-color.h: New file.
20196 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
20198 * doc/invoke.texi (-fdiagnostics-color): Document.
20199 * pretty-print.h (pp_show_color): Define.
20200 (struct pretty_print_info): Add show_color field.
20201 * diagnostic.c: Include diagnostic-color.h.
20202 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
20203 macros. Colorize error:, warning: etc. strings and also the location
20205 (diagnostic_show_locus): Colorize the caret line.
20206 * pretty-print.c: Include diagnostic-color.h.
20207 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
20208 inside of %< %> quotes or quoted through q format modifier.
20210 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20212 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
20214 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20216 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
20217 code in CC_NZ mode.
20218 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
20221 2013-04-11 Marek Polacek <polacek@redhat.com>
20223 PR tree-optimization/48184
20224 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
20226 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
20228 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
20229 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
20230 (skip_simple_arithmetic): Tidy up.
20231 * tree.h (skip_simple_constant_arithmetic): Declare.
20233 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20235 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
20237 2013-04-11 Richard Biener <rguenther@suse.de>
20239 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20240 generate vector constants.
20242 2013-04-11 Richard Biener <rguenther@suse.de>
20244 PR tree-optimization/56878
20245 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
20246 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
20248 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
20249 Prefer to align the DR with the most invariant base address.
20251 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20253 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
20256 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
20258 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
20259 floating-point vector comparisons against 0.
20261 2013-04-11 Jakub Jelinek <jakub@redhat.com>
20263 PR tree-optimization/56899
20264 * fold-const.c (extract_muldiv_1): Apply distributive law
20265 only if TYPE_OVERFLOW_WRAPS (ctype).
20267 2013-04-11 Bin Cheng <bin.cheng@arm.com>
20270 * ira-costs.c (scan_one_insn): Check whether the source rtx of
20271 loading has side effect.
20273 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
20275 * config/sparc/sparc.c: Include tree-pass.h.
20276 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
20277 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
20278 head of file. Change return type. Split off gate function.
20279 (sparc_gate_work_around_errata): New function.
20280 (pass_work_around_errata): New pass definition.
20281 (insert_pass_work_around_errata) New pass insert definition to
20282 insert pass_work_around_errata just after delayed-branch scheduling.
20283 (sparc_option_override): Insert the pass.
20284 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
20286 2013-04-10 David S. Miller <davem@davemloft.net>
20288 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
20289 or -mcpu=hypersparc.
20291 * target.def (cstore_mode): New hook.
20292 * target.h: Include insn-codes.h
20293 * targhooks.c: Likewise.
20294 (default_cstore_mode): New function.
20295 * targhooks.h: Declare it.
20296 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
20297 * doc/tm.texi: Rebuild.
20298 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
20299 target hook, rather than inspecting the insn_data.
20300 * config/sparc/sparc.c (sparc_cstore_mode): New function.
20301 (TARGET_CSTORE_MODE): Redefine.
20302 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
20304 * config/sparc/predicates.md (cstore_result_operand): New special
20306 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
20307 Use it for operand 0.
20308 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
20309 (*snesi_special): Likewise.
20310 (*snesi_zero): Likewise.
20311 (*seqsi_zero): Likewise.
20312 (*sltu_insn): Likewise.
20313 (*sgeu_insn): Likewise.
20314 (*seqdi_special): Make operand 0 and comparison operation be of
20316 (*snedi_special): Likewise.
20317 (*snedi_special_vis3): Likewise.
20318 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
20319 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
20320 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
20321 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
20322 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
20323 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
20324 (*sltu_extend_sp64): Likewise.
20325 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
20326 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
20327 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
20328 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
20329 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
20331 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
20333 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
20334 (aarch64_start_file): Use the new function.
20336 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20337 Jason Merrill <jason@redhat.com>
20339 * common.opt: Add -gdwarf.
20340 * opts.c (common_handle_option): Handle it.
20341 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
20343 2013-04-10 Richard Biener <rguenther@suse.de>
20345 * passes.c (execute_todo): Do not call ggc_collect conditional here.
20346 (execute_one_ipa_transform_pass): But unconditionally here.
20347 (execute_one_pass): And here.
20348 (init_optimization_passes): Remove reload pass.
20349 * tree-pass.h (TODO_ggc_collect): Remove.
20350 (pass_reload): Likewise.
20351 * ira.c (do_reload): Merge into ...
20353 (rest_of_handle_reload): Remove.
20354 (pass_reload): Likewise.
20355 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20356 of reload for vzeroupper pass placement.
20357 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
20358 and todo_flags_finish of all passes.
20360 2013-04-10 Richard Biener <rguenther@suse.de>
20362 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
20363 first_const_oprnd field, rename first_def_type to first_op_type.
20364 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
20365 (vect_get_and_check_slp_defs): Always use the type of the
20366 operand. Allow mixed vect_external_def, vect_constant_def types.
20367 (vect_get_constant_vectors): Handle mixed vect_external_def,
20368 vect_constant_def types.
20370 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
20372 PR tree-optimization/55524
20373 * tree-ssa-math-opts.c
20374 (convert_mult_to_fma): Don't use an fms construct
20375 when we don't have an fms operation, but fnma, and it looks
20376 likely that we'll be able to use the latter.
20378 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
20380 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
20382 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
20383 inline fail caused by overwritable functions.
20385 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
20387 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
20388 unnecessary bits in the constant power of two case.
20390 2013-04-10 Richard Biener <rguenther@suse.de>
20392 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
20393 broken code swapping operands.
20394 (vect_build_slp_tree): Do not compute load permutations here.
20395 (vect_analyze_slp_instance): Compute load permutations here,
20396 after building the SLP tree.
20398 2013-04-09 Christian Bruel <christian.bruel@st.com>
20400 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
20401 of next/prev_real_insn.
20403 2013-04-09 Jan Hubicka <jh@suse.cz>
20405 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
20406 Drop aliased parameter.
20407 (function_and_variable_visibility): Do not handle alias pairs.
20408 * cgraph.c (varpool_externally_visible_p): Update prototype.
20409 * varpool.c (varpool_add_new_variable): Update.
20411 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20413 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
20415 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
20417 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
20419 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
20421 2013-04-09 Marek Polacek <polacek@redhat.com>
20423 PR tree-optimization/48762
20424 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
20426 2013-04-09 Richard Biener <rguenther@suse.de>
20428 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
20430 (vect_build_slp_tree): Likewise.
20431 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
20432 calculating the cost of a SLP instance.
20433 (vect_analyze_slp_instance): Use it from here, after building
20436 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20438 PR middle-end/56883
20439 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
20440 expand_omp_for_static_chunk): Use simple_p = true in
20441 force_gimple_operand_gsi calls when assigning to addressable decls.
20443 2013-04-09 Jeff Law <law@redhat.com>
20445 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
20446 when the boolean was created by converting a wider object which
20447 had a boolean range.
20449 2013-04-09 Richard Biener <rguenther@suse.de>
20451 * tree-vectorizer.h (slp_void_p): Remove.
20452 (slp_tree): Typedef before _slp_tree declaration.
20453 (struct _slp_tree): Use a vector of slp_tree as children.
20454 (vect_get_place_in_interleaving_chain): Remove.
20455 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
20457 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
20459 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
20460 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
20461 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
20462 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
20463 Use slp_node instead of slp_void_p and adjust.
20465 2013-04-09 Richard Biener <rguenther@suse.de>
20467 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
20468 work that is not necessary.
20470 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20472 PR tree-optimization/56854
20473 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
20474 forward into clobber stmts if it would change MEM_REF lhs into
20477 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
20479 * tree.c (type_hash_lookup, type_hash_add): Make static.
20480 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
20482 2013-04-09 Richard Biener <rguenther@suse.de>
20484 * tree.h (unsave_expr_now): Remove.
20485 * tree-inline.c (mark_local_for_remap_r): Remove.
20486 (unsave_expr_1): Likewise.
20487 (unsave_r): Likewise.
20488 (unsave_expr_now): Likewise.
20489 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
20490 (propagate_tree_value): Likewise.
20492 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
20494 * doc/rtl.texi (sequence): Rewrite documentation to match the
20495 current use of SEQUENCE rtl objects.
20496 * rtl.def (SEQUENCE): Likewise.
20498 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
20499 Update documentation.
20500 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
20501 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
20503 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
20505 2013-04-08 Teresa Johnson <tejohnson@google.com>
20507 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
20508 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
20510 (estimate_edge_size_and_time): Add comment to suggest using rounding
20512 (estimate_node_size_and_time): Ditto.
20513 (remap_edge_change_prob): Use helper rounding divide methods.
20514 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
20515 (gimple_mod_pow2_value_transform): Ditto.
20516 (gimple_mod_subtract_transform): Ditto.
20517 (gimple_ic_transform): Ditto.
20518 (gimple_stringops_transform): Ditto.
20519 * stmt.c (conditional_probability): Ditto.
20520 (emit_case_dispatch_table): Ditto.
20521 * lto-cgraph.c (merge_profile_summaries): Ditto.
20522 * tree-optimize.c (execute_fixup_cfg): Ditto.
20523 * cfgcleanup.c (try_forward_edges): Ditto.
20524 * cfgloopmanip.c (scale_loop_profile): Ditto.
20526 (duplicate_loop_to_header_edge): Ditto.
20527 (lv_adjust_loop_entry_edge): Ditto.
20528 * tree-vect-loop.c (vect_transform_loop): Ditto.
20529 * profile.c (compute_branch_probabilities): Ditto.
20530 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
20531 * lto-streamer-in.c (input_cfg): Ditto.
20532 * gimple-streamer-in.c (input_bb): Ditto.
20533 * ipa-cp.c (update_profiling_info): Ditto.
20534 (update_specialized_profile): Ditto.
20535 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
20536 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
20538 * sched-rgn.c (compute_dom_prob_ps): Ditto.
20539 (compute_trg_info): Ditto.
20540 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
20541 (purge_dead_edges): Ditto.
20542 * loop-unswitch.c (unswitch_loop): Ditto.
20543 * cgraphclones.c (cgraph_clone_edge): Ditto.
20544 (cgraph_clone_node): Ditto.
20545 * tree-inline.c (copy_bb): Ditto.
20546 (copy_edges_for_bb): Ditto.
20547 (initialize_cfun): Ditto.
20548 (copy_cfg_body): Ditto.
20549 (expand_call_inline): Ditto.
20551 2013-04-08 Kai Tietz <ktietz@redhat.com>
20553 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
20554 TARGET_CYGWIN64 by TARGET_64BIT.
20556 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20558 * config/epiphany/epiphany.md (GPR_1): New constant.
20559 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
20560 * config/epiphany/epiphany.c (gen_compare_reg):
20561 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
20562 is already in place.
20563 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
20564 Don't require being called during rtl expansion; If y operlaps r0,
20566 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
20567 (epiphany_expand_epilogue): Likewise.
20569 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
20570 Don't use CC_FPmode for ORDERED / UNORDERED.
20571 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
20573 * config/epiphany/constraints.md (CnL): New constraint.
20574 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
20575 * config/epiphany/predicates.md (add_operand): Allow 1024.
20577 * config/epiphany/epiphany.md (logical_op): New code iterator.
20578 (op_mnc): New code attribute.
20579 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
20580 (mov_f+1, mov_f+2): New peephole2 patterns.
20582 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
20583 (cstoresi4): Also allow re-use of zero result when doing a NE
20584 comparison to a non-zero operand.
20585 Use (clobber (scratch)) for first insn if the gpr output is not needed.
20587 * config/epiphany/epiphany.md (<insn_opname>v2si3):
20588 Use gen_addsi3_i / gen_subsi3_i.
20590 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20594 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
20595 contain anything but clobbers, at most one __builtin_stack_restore,
20596 optionally debug stmts and final resx, and if it has at least one
20597 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
20598 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
20599 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
20600 which isn't defaut definition, remove them.
20601 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
20602 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
20603 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
20604 with MEM_REF LHS with SSA_NAME address.
20606 2013-04-08 Jeff Law <law@redhat.com>
20608 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
20610 2013-04-08 Richard Biener <rguenther@suse.de>
20612 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
20614 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
20615 determined vector type.
20616 (vect_analyze_data_refs): Likewise.
20617 (vect_get_new_vect_var): Adjust.
20618 (vect_create_destination_var): Preserve SSA name versions.
20619 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
20620 not dump anything here.
20622 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20624 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
20625 Add member lr_slot_known.
20626 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
20628 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
20629 Remove code that sets lr_slot_offset according to what a previous
20630 version of epiphany_emit_save_restore used to do.
20631 (epiphany_emit_save_restore): When doing an lr save or restore,
20632 set/verify lr_slot_known and lr_slot_offset.
20634 2013-04-08 Xinyu Qi <xyqi@marvell.com>
20637 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
20640 2013-04-08 Richard Biener <rguenther@suse.de>
20642 * alias.c (find_base_term): Fix thinko in previous change.
20644 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20646 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
20647 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
20648 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
20649 if possible to compute val.
20650 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
20651 For QImode integers don't require anything about precision. Use
20652 const_with_all_bytes_same to find out if the constant doesn't have
20653 repeated bytes in it.
20655 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20657 * config/s390/s390.c (s390_expand_insv): Only accept insertions
20660 2013-04-08 Marek Polacek <polacek@redhat.com>
20662 PR rtl-optimization/48182
20663 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
20666 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20669 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
20670 nuses, make sure we have a label.
20672 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20675 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
20676 (rs6000_emit_swdiv_low_precision): Remove.
20677 (rs6000_emit_swdiv): Rewrite to handle between one and four
20678 iterations of Newton-Raphson generally; modify required number of
20679 iterations for some cases.
20680 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
20682 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
20684 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
20685 set-but-unused variable.
20687 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
20688 basic blocks of released function bodies garbage-collectable.
20690 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
20691 (struct rtl_opt_pass): Add TODO_df_finish.
20693 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
20695 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20697 * config/arm/constraints.md (q): New constraint.
20698 * config/arm/ldrdstrd.md: New file.
20699 * config/arm/arm.md (ldrdstrd.md) New include.
20700 (arm_movdi): Use "q" instead of "r" constraint
20701 for double-word memory access.
20702 (movdf_soft_insn): Likewise.
20703 * config/arm/vfp.md (movdi_vfp): Likewise.
20704 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
20705 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
20706 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
20707 (mem_ok_for_ldrd_strd): Likewise.
20708 (output_move_double): Update assertion.
20710 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20712 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
20714 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20716 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
20717 define_insn_and_split.
20718 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
20720 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20722 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
20723 define_insn_and_split.
20724 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
20725 (shiftsi3_compare): New pattern.
20726 (rrx): New pattern.
20727 * config/arm/unspecs.md (UNSPEC_RRX): New.
20729 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20731 * config/arm/arm.md (negdi_extendsidi): New pattern.
20732 (negdi_zero_extendsidi): Likewise.
20734 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20736 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
20737 define_insn_and_split.
20738 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
20739 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
20741 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20743 * config/arm/arm.md (arm_subdi3): Convert define_insn into
20744 define_insn_and_split.
20745 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
20746 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
20748 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20750 * config/arm/arm.md (subsi3_carryin): New pattern.
20751 (subsi3_carryin_const): Likewise.
20752 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
20753 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
20755 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20757 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
20759 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20761 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
20762 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
20764 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20766 * config/arm/arm.c (arm_expand_builtin): Change fcode
20767 type to unsigned int.
20769 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20771 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
20773 2013-04-04 Ian Lance Taylor <iant@google.com>
20775 * doc/standards.texi (Standards): The Go frontend supports the Go 1
20778 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
20780 PR middle-end/56729
20781 * df-scan.c (df_insn_delete): Disable failing assert.
20783 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20785 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
20786 New function prototype.
20787 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
20788 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
20789 (arm_builtin_vectorized_function): New function.
20791 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20793 * config/arm/arm_neon_builtins.def: New file.
20794 * config/arm/arm.c (neon_builtin_data): Move contents to
20795 arm_neon_builtins.def.
20796 (enum arm_builtins): Include neon builtin definitions.
20797 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
20798 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
20800 2013-04-04 Marek Polacek <polacek@redhat.com>
20802 PR tree-optimization/48186
20803 * predict.c (maybe_hot_frequency_p): Return false if
20804 HOT_BB_FREQUENCY_FRACTION is 0.
20805 (cgraph_maybe_hot_edge_p): Likewise.
20807 2013-04-04 Richard Biener <rguenther@suse.de>
20809 PR tree-optimization/56826
20810 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
20813 2013-04-04 Richard Biener <rguenther@suse.de>
20815 PR tree-optimization/56213
20816 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
20817 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
20819 2013-04-04 Richard Biener <rguenther@suse.de>
20821 PR tree-optimization/56837
20822 * tree-loop-distribution.c (classify_partition): For non-zero
20823 values require that the value has the same precision as its
20824 mode to be useful as memset value.
20826 2013-04-03 Nick Clifton <nickc@redhat.com>
20828 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
20829 (fmssf4): Use fmsf.s on E3V5 architectures.
20830 (fnmasf4): Use fnmaf.s on E3V5 architectures.
20831 (fnmssf4): Use fnmsf.s on E3V5 architectures.
20833 2013-04-03 Jeff Law <law@redhat.com>
20835 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
20836 (lra-eliminations.o): Likewise.
20838 2013-04-03 Teresa Johnson <tejohnson@google.com>
20840 * gcov-io.c (compute_working_sets): Moved most of body of old
20841 compute_working_sets here from profile.c.
20842 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
20843 (gcov_working_set_t): Moved typedef here from basic-block.h
20844 (compute_working_set): Declare.
20845 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
20846 (get_working_sets): Renamed from compute_working_set,
20847 replace most of body with call to new compute_working_sets.
20848 (get_exec_counts): Replace call to compute_working_sets
20849 to get_working_sets.
20850 * profile.h (get_working_sets): Renamed from compute_working_set.
20851 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
20852 to get_working_sets.
20853 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
20854 * gcov-dump.c (dump_working_sets): New function.
20856 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
20858 * hwint.c (sext_hwi, zext_hwi): New functions.
20859 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
20860 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
20861 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
20862 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
20863 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
20864 (sext_hwi, zext_hwi): New functions.
20866 2013-04-03 Jeff Law <law@redhat.com>
20868 PR tree-optimization/56799
20869 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
20870 back test for widening conversion erroneously dropped in prior change.
20872 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20875 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
20876 instead of next_real_insn.
20878 2013-04-03 Marek Polacek <polacek@redhat.com>
20881 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
20883 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20886 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
20888 (thumb1_output_casesi): Likewise.
20889 (thumb2_output_casesi): Likewise.
20891 2013-04-03 Richard Biener <rguenther@suse.de>
20893 PR tree-optimization/56817
20894 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
20896 (tree_unroll_loops_completely_1): ... new function to manually
20897 walk the loop tree, properly defering outer loops of unrolled
20898 loops to later iterations.
20900 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20902 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
20903 (vectorizable_load): Likewise.
20904 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
20905 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
20907 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20909 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
20912 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
20914 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
20916 2013-04-03 Bin Cheng <bin.cheng@arm.com>
20918 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
20920 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20922 PR tree-optimization/56790
20923 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
20926 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20928 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
20930 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
20931 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
20934 2013-04-03 Jakub Jelinek <jakub@redhat.com>
20937 * tree.h (force_folding_builtin_constant_p): New decl.
20938 * builtins.c (force_folding_builtin_constant_p): New variable.
20939 (fold_builtin_constant_p): Fold immediately also if
20940 force_folding_builtin_constant_p.
20942 2013-04-03 Richard Biener <rguenther@suse.de>
20944 PR tree-optimization/56812
20945 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
20946 DRs of the same interleaving chain are independent.
20948 2013-04-02 Jason Merrill <jason@redhat.com>
20950 * gdbinit.in (pbb): Use debug fn.
20952 2013-04-02 Lawrence Crowl <crowl@google.com>
20954 * sese.h (struct ivtype_map_elt_s): Remove unused.
20955 (extern debug_ivtype_map): Remove unused.
20956 (extern eq_ivtype_map_elts): Remove unused.
20957 * sese.c (debug_ivtype_map): Removed unused.
20958 (debug_ivtype_map_1): Removed unused.
20959 (debug_ivtype_elt): Remove unused.
20960 (eq_ivtype_map_elts): Remove unused.
20963 2013-04-02 Kai Tietz <ktietz@redhat.com>
20966 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
20967 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
20968 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
20970 (legitimize_pe_coff_symbol): Likewise.
20971 (is_imported_p): New helper-function.
20972 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
20973 for Windows x64 targets.
20974 (ix86_expand_prologue): Optimize for pe-coff targets.
20975 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
20976 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
20977 medium/large code-model.
20978 (legitimize_pic_address): Likewise.
20979 (legitimize_tls_address): Likewise.
20980 (ix86_expand_call): Likewise.
20981 (x86_output_mi_thunk): Likewise.
20982 (get_dllimport_decl): Add new beimport argument.
20983 (construct_plt_address): Don't assert for x64 pe-coff targets.
20984 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
20986 (SYMBOL_FLAG_STUBVAR): New macro.
20987 (SYMBOL_REF_STUBVAR_P): Likewise.
20988 * config/i386/winnt.c (stub_list): New structure.
20989 (stub_head): New local variable.
20990 (i386_pe_record_stub): New function.
20991 (i386_pe_file_end): Emit refptr-stubs.
20993 2013-04-02 Jakub Jelinek <jakub@redhat.com>
20995 PR rtl-optimization/56745
20996 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
20997 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
21000 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
21001 and both of them are MEM_REFs, just compare first argument for
21002 equality and attempt to deal even with differing offsets.
21005 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
21006 of gimple_clobber_p to be MEM_REF.
21007 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
21008 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
21009 after gimplification.
21010 * asan.c (get_mem_ref_of_assignment): Don't instrument
21011 gimple_clobber_p stmts.
21012 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
21013 gimple_clobber_p stmt if they have MEM_REF lhs and
21014 are dead because of another gimple_clobber_p stmt.
21015 * tree-ssa-live.c (clear_unused_block_pointer): Treat
21016 gimple_clobber_p stmts like debug stmts.
21017 (remove_unused_locals): Remove clobbers with MEM_REF lhs
21018 that refer to unused VAR_DECLs or uninitialized values.
21019 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
21020 gimple_clobber_p stmts if they refer to removed parameters.
21021 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
21024 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
21026 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
21027 using SWI48 mode attribute.
21029 2013-04-02 Wei Mi <wmi@google.com>
21031 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
21032 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
21033 *<rotate_insn><mode>3_mask in i386.md.
21035 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
21037 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
21039 2013-04-02 Richard Biener <rguenther@suse.de>
21041 PR tree-optimization/56778
21042 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
21043 Runtime alias tests are not supported for gather loads.
21044 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
21045 stmts referenced from SSA operands before updating SSA form.
21047 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
21048 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21050 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
21051 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
21052 * config/arm/cortex-a53.md: New file.
21053 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
21054 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
21055 * config/arm/arm.c (arm_issue_rate): Likewise.
21056 * config/arm/arm-tune.md: Regenerate
21057 * config/arm/arm-tables.opt: Regenerate.
21058 * config/arm/arm-cores.def: Add cortex-a53.
21060 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
21062 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
21065 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21067 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
21068 scalar load/store operations using B/H registers.
21069 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
21071 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21073 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
21075 * config/aarch64/aarch64.c
21076 (aarch64_simd_scalar_immediate_valid_for_move): New.
21077 * config/aarch64/aarch64-protos.h
21078 (aarch64_simd_scalar_immediate_valid_for_move): New.
21079 * config/aarch64/constraints.md (Dh, Dq): New.
21080 * config/aarch64/iterators.md (hq): New.
21082 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
21084 * reorg.c (get_branch_condition): Deal with conditional returns.
21085 (fill_simple_delay_slots): Remove dead code dealing with jumps.
21087 2013-04-01 Wei Mi <wmi@google.com>
21089 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
21090 Truncate operand 2 using %b asm operand modifier.
21091 (*<shift_insn><mode>3_mask): Ditto.
21092 (*<rotate_insn><mode>3_mask): Ditto.
21094 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
21096 PR middle-end/56798
21097 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
21099 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
21101 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
21103 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
21105 2013-03-30 Lawrence Crowl <crowl@google.com>
21107 * dse.c (clear_alias_sets): Remove never set.
21108 (disqualified_clear_alias_sets): Remove never set.
21109 (clear_alias_mode_pool): Remove never set.
21110 (dse_step0): Remove condition that is never true.
21111 (canon_address): Remove condition that is never true.
21112 (dse_step7): Remove condition that is never true.
21113 (rest_of_handle_dse): Remove condition that is never true.
21114 (rest_of_handle_dse::did_global): Remove never read from above.
21115 (dse_step2_spill): Remove never called from above.
21116 (dse_step5_spill): Remove never called from above.
21118 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
21120 * doc/md.texi (Standard Names) <casesi>: Update documentation for
21121 JUMP_TABLE_DATA changes.
21122 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
21123 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
21124 (Insns) <jump_table_data>: New entry.
21125 * doc/tm.texi: Regenerate.
21127 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
21129 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
21130 for table jump at the end of a basic block using tablejump_p.
21131 * targhooks.c (default_invalid_within_doloop): Likewise.
21132 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
21133 target hook implementation that is identical to the default hook.
21134 (rs6000_invalid_within_doloop): Remove.
21136 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
21137 unused variable from tablejump_p call.
21139 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
21140 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
21141 (INSN_DELETED_P): Likewise.
21142 (emit_jump_table_data): New prototype.
21143 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
21144 after 4th as unused.
21145 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
21146 * sched-vis.c (print_insn): Likewise.
21147 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
21148 insn for compatibility with back ends that use next_active_insn to
21149 identify jump table data.
21150 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
21151 (remove_insn): Likewise.
21152 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
21154 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
21155 (emit_jump_table_data): New function.
21157 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
21158 basic block, a JUMP_TABLE_DATA never is.
21159 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
21160 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
21161 off from code handling real insns.
21162 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
21163 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
21164 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
21165 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
21166 is not a NONDEBUG_INSN_P.
21167 * ira-costs.c (scan_one_insn): Likewise.
21168 * jump.c (mark_all_labels): Likewise.
21169 (mark_jump_label_1): Likewise.
21170 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21171 * lra.c (get_insn_freq): Expect all insns reaching here to be in
21173 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
21174 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
21175 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
21176 JUMP_TABLE_DATA_P insns.
21177 (calculate_elim_costs_all_insns): Likewise.
21178 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
21179 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
21180 (delete_output_reload): Code style fixups.
21181 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
21182 insn flags on this non-insn.
21183 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
21184 as scheduling barriers, for pre-change compatibility.
21185 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
21186 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
21188 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
21189 redundant JUMP_TABLE_DATA_P test.
21190 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
21191 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21192 (frv_for_each_packet): Likewise.
21193 * config/i386/i386.c (min_insn_size): Likewise.
21194 (ix86_avoid_jump_mispredicts): Likewise.
21195 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21196 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
21197 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21198 (mips16_insn_length): Robustify.
21199 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
21200 (mips16_split_long_branches): Likewise.
21201 * config/pa/pa.c (pa_combine_instructions): Likewise.
21202 * config/rs6000/rs6000.c (get_next_active_insn): Treat
21203 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
21204 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
21205 as contributing to pool range lengths.
21206 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
21207 Remove redundant JUMP_TABLE_DATA_P test.
21208 (sh_loop_align): Likewise.
21209 (split_branches): Likewise.
21210 (sh_insn_length_adjustment): Likewise.
21211 * config/spu/spu.c (get_branch_target): Likewise.
21213 2013-03-29 Jan Hubicka <jh@suse.cz>
21215 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
21216 gcov streaming; stream hot bb threshold to ltrans.
21217 * predict.c (get_hot_bb_threshold): Break out from ....
21218 (maybe_hot_count_p): ... here.
21219 (set_hot_bb_threshold): New function.
21220 * lto-section-in.c (lto_section_name): Add profile.
21221 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
21222 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
21223 and data-streamer.h
21224 (histogram_entry): New structure.
21225 (histogram, histogram_pool): New global vars.
21226 (histogram_hash): New structure.
21227 (histogram_hash::hash): New method.
21228 (histogram_hash::equal): Likewise.
21229 (account_time_size): New function.
21230 (cmp_counts): New function.
21231 (dump_histogram): New function.
21232 (ipa_profile_generate_summary): New function.
21233 (ipa_profile_write_summary): New function.
21234 (ipa_profile_read_summary): New function.
21235 (ipa_profile): Decide on threshold.
21236 (pass_ipa_profile): Add ipa_profile_write_summary and
21237 ipa_profile_read_summary.
21238 * Makefile.in (ipa.o): Update dependencies.
21239 * lto-streamer.h (LTO_section_ipa_profile): New section.
21241 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
21243 * tree.h (VAR_P): New.
21245 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
21248 * doc/invoke.texi ([-fwhole-program]): Fix typo.
21250 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
21252 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
21253 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21254 (control_flow_insn_p): Likewise.
21255 * cfgrtl.c (duplicate_insn_chain): Likewise.
21256 * final.c (get_attr_length_1): Likewise.
21257 (shorten_branches): Likewise.
21258 (final_scan_insn): Likewise.
21259 * function.c (instantiate_virtual_regs): Likewise.
21260 * gcse.c (insert_insn_end_basic_block): Likewise.
21261 * ira-costs.c (scan_one_insn): Likewise.
21262 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21263 * lra.c (check_rtl): Likewise.
21264 * reload1.c (elimination_costs_in_insn): Likewise.
21265 * reorg.c (follow_jumps): Likewise.
21267 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
21268 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21269 (thumb_far_jump_used_p): Likewise.
21270 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
21271 (workaround_speculation): Likewise.
21272 (add_sched_insns_for_speculation): Likewise.
21273 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
21274 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21275 (frv_for_each_packet): Likewise.
21276 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
21277 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
21278 (final_emit_insn_group_barriers): Likewise.
21279 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21280 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21281 (mips16_insn_length): Likewise.
21282 * config/pa/pa.c (pa_reorg): Likewise.
21283 (pa_combine_instructions): Likewise.
21284 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
21285 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
21286 (sh_reorg): Likewise.
21287 (split_branches): Likewise.
21288 * config/spu/spu.c (get_branch_target): Likewise.
21290 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
21293 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
21295 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
21296 Fix declaration name.
21298 2013-03-28 Lawrence Crowl <crowl@google.com>
21300 * graphds.h (struct graph.indicies): Remove unused.
21301 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
21302 (SCOP_ORIGINAL_PDDRS): Remove unused.
21303 * sese.h (extern insert_loop_close_phis): Removed unused.
21304 (extern insert_guard_phis): Removed unused.
21305 (extern ivtype_map_elt_info): Removed unused.
21306 (new_ivtype_map_elt): Removed unused.
21307 * sese.c (ivtype_map_elt_info): Removed unused.
21309 2013-03-28 Lawrence Crowl <crowl@google.com>
21311 * Makefile.in: Add several missing include dependences.
21313 (test-dump.o): New. This object is not added to any executable,
21314 but is present for ad-hoc testing.
21316 (debug (const bitmap_head_def &)): New.
21317 (debug (const bitmap_head_def *)): New.
21319 (extern debug (const bitmap_head_def &)): New.
21320 (extern debug (const bitmap_head_def *)): New.
21322 (debug (edge_def &)): New.
21323 (debug (edge_def *)): New.
21325 (debug (basic_block_def &)): New.
21326 (debug (basic_block_def *)): New.
21328 (dump_node (const_tree, int, FILE *)): Correct source file.
21330 (debug (die_struct &)): New.
21331 (debug (die_struct *)): New.
21333 (extern debug (die_struct &)): New.
21334 (extern debug (die_struct *)): New.
21335 * gimple-pretty-print.c
21336 (debug (gimple_statement_d &)): New.
21337 (debug (gimple_statement_d *)): New.
21338 * gimple-pretty-print.h
21339 (extern debug (gimple_statement_d &)): New.
21340 (extern debug (gimple_statement_d *)): New.
21342 (debug (ira_allocno_copy &)): New.
21343 (debug (ira_allocno_copy *)): New.
21344 (debug (ira_allocno &)): New.
21345 (debug (ira_allocno *)): New.
21347 (extern debug (ira_allocno_copy &)): New.
21348 (extern debug (ira_allocno_copy *)): New.
21349 (extern debug (ira_allocno &)): New.
21350 (extern debug (ira_allocno *)): New.
21352 (debug (live_range &)): New.
21353 (debug (live_range *)): New.
21355 (debug (lra_live_range &)): New.
21356 (debug (lra_live_range *)): New.
21358 (debug (lra_live_range &)): New.
21359 (debug (lra_live_range *)): New.
21361 (debug (omega_pb_d &)): New.
21362 (debug (omega_pb_d *)): New.
21364 (extern debug (omega_pb_d &)): New.
21365 (extern debug (omega_pb_d *)): New.
21367 (debug (const rtx_def &)): New.
21368 (debug (const rtx_def *)): New.
21370 (debug_tree (tree): Move within file.
21371 (debug_raw (const tree_node &)): New.
21372 (debug_raw (const tree_node *)): New.
21373 (dump_tree_via_hooks (const tree_node *, int)): New.
21374 (debug (const tree_node &)): New.
21375 (debug (const tree_node *)): New.
21376 (debug_verbose (const tree_node &)): New.
21377 (debug_verbose (const tree_node *)): New.
21378 (debug_head (const tree_node &)): New.
21379 (debug_head (const tree_node *)): New.
21380 (debug_body (const tree_node &)): New.
21381 (debug_body (const tree_node *)): New.
21382 (debug_vec_tree (tree): Move and reimplement in terms of dump.
21383 (debug (vec<tree, va_gc> &)): New.
21384 (debug (vec<tree, va_gc> *)): New.
21386 (extern debug (const rtx_def &)): New.
21387 (extern debug (const rtx_def *)): New.
21389 (debug_raw (simple_bitmap_def &)): New.
21390 (debug_raw (simple_bitmap_def *)): New.
21391 (debug (simple_bitmap_def &)): New.
21392 (debug (simple_bitmap_def *)): New.
21394 (extern debug (simple_bitmap_def &)): New.
21395 (extern debug (simple_bitmap_def *)): New.
21396 (extern debug_raw (simple_bitmap_def &)): New.
21397 (extern debug_raw (simple_bitmap_def *)): New.
21399 (debug (vinsn_def &)): New.
21400 (debug (vinsn_def *)): New.
21401 (debug_verbose (vinsn_def &)): New.
21402 (debug_verbose (vinsn_def *)): New.
21403 (debug (expr_def &)): New.
21404 (debug (expr_def *)): New.
21405 (debug_verbose (expr_def &)): New.
21406 (debug_verbose (expr_def *)): New.
21407 (debug (vec<rtx> &)): New.
21408 (debug (vec<rtx> *)): New.
21410 (extern debug (vinsn_def &)): New.
21411 (extern debug (vinsn_def *)): New.
21412 (extern debug_verbose (vinsn_def &)): New.
21413 (extern debug_verbose (vinsn_def *)): New.
21414 (extern debug (expr_def &)): New.
21415 (extern debug (expr_def *)): New.
21416 (extern debug_verbose (expr_def &)): New.
21417 (extern debug_verbose (expr_def *)): New.
21418 (extern debug (vec<rtx> &)): New.
21419 (extern debug (vec<rtx> *)): New.
21421 (_list_iter_cond_expr): Make inline instead of static.
21423 (debug (sreal &)): New.
21424 (debug (sreal *)): New.
21426 (extern debug (sreal &)): New.
21427 (extern debug (sreal *)): New.
21429 (extern debug_raw (const tree_node &)): New.
21430 (extern debug_raw (const tree_node *)): New.
21431 (extern debug (const tree_node &)): New.
21432 (extern debug (const tree_node *)): New.
21433 (extern debug_verbose (const tree_node &)): New.
21434 (extern debug_verbose (const tree_node *)): New.
21435 (extern debug_head (const tree_node &)): New.
21436 (extern debug_head (const tree_node *)): New.
21437 (extern debug_body (const tree_node &)): New.
21438 (extern debug_body (const tree_node *)): New.
21439 (extern debug (vec<tree, va_gc> &)): New.
21440 (extern debug (vec<tree, va_gc> *)): New.
21442 (debug (struct loop &)): New.
21443 (debug (struct loop *)): New.
21444 (debug_verbose (struct loop &)): New.
21445 (debug_verbose (struct loop *)): New.
21446 * tree-dump.c: Add header dependence.
21448 (extern debug (struct loop &)): New.
21449 (extern debug (struct loop *)): New.
21450 (extern debug_verbose (struct loop &)): New.
21451 (extern debug_verbose (struct loop *)): New.
21453 (debug (data_reference &)): New.
21454 (debug (data_reference *)): New.
21455 (debug (vec<data_reference_p> &)): New.
21456 (debug (vec<data_reference_p> *)): New.
21457 (debug (vec<ddr_p> &)): New.
21458 (debug (vec<ddr_p> *)): New.
21460 (extern debug (data_reference &)): New.
21461 (extern debug (data_reference *)): New.
21462 (extern debug (vec<data_reference_p> &)): New.
21463 (extern debug (vec<data_reference_p> *)): New.
21464 (extern debug (vec<ddr_p> &)): New.
21465 (extern debug (vec<ddr_p> *)): New.
21467 (debug (pt_solution &)): New.
21468 (debug (pt_solution *)): New.
21470 (extern debug (pt_solution &)): New.
21471 (extern debug (pt_solution *)): New.
21473 (debug (_var_map &)): New.
21474 (debug (_var_map *)): New.
21475 (debug (tree_live_info_d &)): New.
21476 (debug (tree_live_info_d *)): New.
21478 (extern debug (_var_map &)): New.
21479 (extern debug (_var_map *)): New.
21480 (extern debug (tree_live_info_d &)): New.
21481 (extern debug (tree_live_info_d *)): New.
21483 2013-03-28 Jan Hubicka <jh@suse.cz>
21485 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
21487 2013-03-28 Ian Bolton <ian.bolton@arm.com>
21489 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
21490 record only when desired or required.
21492 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
21494 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
21495 *vec_extractv2di_1_rex64. Use x64 isa attribute.
21497 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21499 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
21500 (*andsi3_compare0_uxtw): New pattern.
21501 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
21502 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
21504 2013-03-28 Jan Hubicka <jh@suse.cz>
21506 * data-streamer-in.c (streamer_read_gcov_count): New function.
21507 * gimple-streamer-out.c: Include value-prof.h.
21508 (output_gimple_stmt): Output histogram.
21509 (output_bb): Use streamer_write_gcov_count.
21510 * value-prof.c: Include data-streamer.h
21511 (dump_histogram_value): Add HIST_TYPE_MAX.
21512 (stream_out_histogram_value): New function.
21513 (stream_in_histogram_value): New function.
21514 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
21515 (stream_out_histogram_value, stream_in_histogram_value): Declare.
21516 * data-streamer-out.c (streamer_write_gcov_count): New function.
21517 (streamer_write_gcov_count_stream): New function.
21518 * lto-cgraph.c (lto_output_edge): Update counter streaming.
21519 (lto_output_node): Likewise.
21520 (input_node, input_edge): Likewise.
21521 * lto-streamer-out.c (output_cfg): Update streaming.
21522 * lto-streamer-in.c (input_cfg): Likewise.
21523 * data-streamer.h (streamer_write_gcov_count,
21524 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
21525 * gimple-streamer-in.c: Include value-prof.h
21526 (input_gimple_stmt): Input histograms.
21527 (input_bb): Update profile streaming.
21529 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21531 * genmodes.c (emit_max_int): New function.
21532 (emit_insn_modes_h): Added call to emit_max_function.
21533 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
21535 * machmode.def: Fixed comment.
21537 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21539 * combine.c (try_combine): Removed useless assert.
21540 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
21542 2013-03-28 Marek Polacek <polacek@redhat.com>
21543 Richard Biener <rguenther@suse.de>
21545 PR tree-optimization/56695
21546 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
21547 build signed result of a vector comparison.
21548 * tree-cfg.c (verify_gimple_comparison): Check that a result
21549 of a vector comparison has signed type.
21551 2013-03-28 Richard Biener <rguenther@suse.de>
21553 PR tree-optimization/37021
21554 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
21555 do not restrict gaps between groups.
21556 * tree-vect-stmts.c (vectorizable_load): Properly account for
21557 a gap between groups.
21559 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
21561 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
21562 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
21565 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
21567 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
21568 * doc/extend.texi (Named Address Spaces): Ditto.
21569 (Variable Attributes): Ditto.
21571 2013-03-27 Kai Tietz <ktietz@redhat.com>
21573 * config.build: Add support for cygwin x64 target.
21574 * config.gcc: Likewise.
21575 * config.host: Likewise.
21576 * configure.ac: Likewise
21577 * configure: Regenerated.
21579 2013-03-27 Kai Tietz <ktietz@redhat.com>
21581 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
21582 * config/i386/t-cygwin-w64: New file.
21583 * config/i386/cygwin-w64.h: New file.
21584 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
21585 and add support for x64-cygwin target.
21586 (CPP_SPEC): Likewise.
21587 (CXX_WRAP_SPEC_LIST): Undefine before define.
21588 (LIBGCJ_SONAME): Use 15 as version.
21590 2013-03-27 Richard Biener <rguenther@suse.de>
21592 PR tree-optimization/56716
21593 * tree-ssa-structalias.c (perform_var_substitution): Adjust
21594 dumping for ref nodes.
21596 2013-03-27 Martin Jambor <mjambor@suse.cz>
21598 PR tree-optimization/55334
21599 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
21600 restricted pointers to arrays.
21602 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
21604 * Makefile.in (.SUFFIXES): Add .cc.
21605 (.c.o): Apply same recipe for implicit rule .cc.o.
21607 2013-03-27 Richard Biener <rguenther@suse.de>
21609 PR tree-optimization/37021
21610 * tree-vect-data-refs.c (vect_check_strided_load): Allow
21611 REALPART/IMAGPART_EXPRs around the supported refs.
21612 * tree-ssa-structalias.c (find_func_aliases): Assume that
21613 floating-point values are not used to transfer pointers.
21615 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
21617 * target.def (TARGET_HAS_IFUNC_P): New target hook.
21618 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
21619 * doc/tm.texi: Regenerate.
21620 * targhooks.h (default_has_ifunc_p): New.
21621 * targhooks.c (default_has_ifunc_p): Ditto.
21622 * config/linux-protos.h: New file.
21623 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
21624 hook for linux which disables support of indirect functions in android.
21625 * config/linux-android.c: New file.
21626 * config/t-linux-android.c: Ditto.
21627 * config.gcc: Added new object file linux-android.o.
21628 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
21629 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
21630 * varasm.c (do_assemble_alias): Likewise.
21631 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
21632 doesn't support indirect functions.
21633 * configure: Regenerate.
21635 2013-03-27 Bin Cheng <bin.cheng@arm.com>
21638 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
21639 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
21642 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21644 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
21646 2013-03-27 Terry Guo <terry.guo@arm.com>
21648 * config/arm/arm-cores.def: Added core cortex-r7.
21649 * config/arm/arm-tune.md: Regenerated.
21650 * config/arm/arm-tables.opt: Regenerated.
21651 * doc/invoke.texi: Added entry for core cortex-r7.
21653 2013-03-27 Walter Lee <walt@tilera.com>
21655 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
21656 double-decrement of next_scratch_regno.
21658 2013-03-27 Walter Lee <walt@tilera.com>
21660 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
21662 (insn_v1mulus): Ditto.
21663 (insn_v2muls): Ditto.
21665 2013-03-27 Walter Lee <walt@tilera.com>
21667 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
21668 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
21670 2013-03-27 Walter Lee <walt@tilera.com>
21672 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
21673 (*sibcall_value): Ditto.
21675 2013-03-27 Walter Lee <walt@tilera.com>
21677 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
21678 (insn_mnz_v8qi): ... this ...
21679 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
21681 (insn_v<n>mnz): Replaced by ...
21682 (insn_v1mnz): ... this ...
21683 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
21685 (insn_mz_<mode>): Replaced by ...
21686 (insn_mz_v8qi): ... this ...
21687 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
21689 (insn_v<n>mz): Replaced by ...
21690 (insn_v1mz): ... this ...
21691 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
21694 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
21696 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
21698 2013-03-26 Roland McGrath <mcgrathr@google.com>
21700 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
21701 than fprintf with a non-constant, non-format string.
21703 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21705 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
21706 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
21707 operand 0 predicate.
21708 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
21709 attribute. Use general_x64nomem_operand as operand 1 predicate.
21710 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
21711 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21712 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
21713 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21714 (mov<mode>_insv_1): Remove expander. Merge insn with
21715 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
21716 Use general_x64nomem_operand as operand 1 predicate.
21717 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
21718 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
21719 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21720 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
21721 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
21722 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
21723 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21724 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
21725 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
21726 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
21727 (general_x64nomem_operand): Ditto.
21729 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
21731 * config/rtems.opt: Add -pthread option.
21733 2013-03-26 Richard Biener <rguenther@suse.de>
21735 * alias.c (find_base_term): Avoid redundant and not used recursion.
21736 (base_alias_check): Get the initial base term from the caller.
21737 (true_dependence_1): Compute and pass base terms to base_alias_check.
21738 (write_dependence_p): Likewise.
21739 (may_alias_p): Likewise.
21741 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
21743 * config/aarch64/aarch64.c (aarch64_classify_address): Support
21744 PC-relative load in SI modes and above only.
21746 2013-03-26 Xinyu Qi <xyqi@marvell.com>
21748 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
21749 * config/arm/iwmmxt.md (WCGR0): Update.
21750 (WCGR1, WCGR2, WCGR3): Likewise.
21752 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21754 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
21755 Use x64 and nox64 isa attributes.
21757 2013-03-26 Richard Biener <rguenther@suse.de>
21759 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
21760 alignment computations and rely on get_object_alignment_1
21761 for the !TYPE_P case.
21762 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
21764 2013-03-26 Walter Lee <walt@tilera.com>
21766 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
21767 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
21769 2013-03-25 Jeff Law <law@redhat.com>
21771 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
21772 check for INTEGRAL_TYPE_P that was missing due to checking in
21773 wrong version of prior patch.
21775 2013-03-25 Walter Lee <walt@tilera.com>
21777 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
21778 TILEGX_INSN_SHUFFLEBYTES1.
21779 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
21781 (tilegx_builtins): Ditto.
21782 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
21784 2013-03-25 Walter Lee <walt@tilera.com>
21786 * config/tilegx/tilegx.md (floatsisf2): New pattern.
21787 (floatunssisf2): New pattern.
21788 (floatsidf2): New pattern.
21789 (floatunssidf2): New pattern.
21791 2013-03-25 Walter Lee <walt@tilera.com>
21793 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
21794 tests for constraint J, K, N, P.
21796 2013-03-25 Walter Lee <walt@tilera.com>
21798 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
21799 Use indirect/pcrel encoding.
21800 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
21803 2013-03-25 Steve Ellcey <sellcey@mips.com>
21805 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
21806 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
21807 * config/mips/mips.c (mips_option_override): Set IMADD default.
21808 * config/mips/mips.h (PTF_AVOID_IMADD): New.
21809 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
21810 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
21811 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
21812 * doc/invoke.texi (-mimadd/-mno-imadd): New.
21814 2013-03-25 Jeff Law <law@redhat.com>
21816 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
21817 slightly to avoid creating and folding useless trees. Simplify
21818 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
21820 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
21822 * config/i386/i386.md (*zero_extendsidi2): Merge with
21823 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
21824 * config/i386/predicates.md (x86_64_zext_operand): Rename from
21825 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
21826 targets. Clarify comment.
21828 2013-03-25 Martin Jambor <mjambor@suse.cz>
21830 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
21831 pass-through jump functions differently.
21832 (ipa_read_jump_function): Likewise. Also use setter functions to set
21835 2013-03-25 Martin Jambor <mjambor@suse.cz>
21837 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
21838 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
21840 (ipa_get_indirect_edge_target): New function.
21841 (devirtualization_time_bonus): New parameter known_aggs, pass it to
21842 ipa_get_indirect_edge_target. Update all callers.
21843 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
21844 ipa_get_indirect_edge_target_1 instead of calling
21845 ipa_get_indirect_edge_target.
21846 (create_specialized_node): Pass aggvlas to
21847 ipcp_discover_new_direct_edges.
21849 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21851 * config/arm/arm.md (f_sels, f_seld): New types.
21852 (*cmov<mode>): New pattern.
21853 * config/arm/predicates.md (arm_vsel_comparison_operator): New
21856 2013-03-25 Kai Tietz <ktietz@redhat.com>
21858 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
21859 POSIX-printf for mingw-hosted builds.
21861 2013-03-25 Richard Biener <rguenther@suse.de>
21863 PR middle-end/56694
21864 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
21865 must-not-throw stmt location.
21867 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21869 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
21870 Emit load-acquire versions when acq is true.
21871 (arm_emit_store_exclusive): Add rel parameter.
21872 Emit store-release versions when rel is true.
21873 (arm_split_compare_and_swap): Use acquire-release instructions
21875 of barriers when appropriate.
21876 (arm_split_atomic_op): Likewise.
21877 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
21878 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
21879 (VUNSPEC_SLX): Likewise.
21880 (VUNSPEC_LDA): Likewise.
21881 (VUNSPEC_STL): Likewise.
21882 * config/arm/sync.md (atomic_load<mode>): New pattern.
21883 (atomic_store<mode>): Likewise.
21884 (arm_load_acquire_exclusive<mode>): Likewise.
21885 (arm_load_acquire_exclusivesi): Likewise.
21886 (arm_load_acquire_exclusivedi): Likewise.
21887 (arm_store_release_exclusive<mode>): Likewise.
21889 2013-03-25 Catherine Moore <clm@codesourcery.com>
21891 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
21892 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
21893 * config/mip/predicates.md (lwsp_swsp_operand,
21894 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
21895 sb16_operand, db4_operand, db7_operand, ib3_operand,
21896 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
21897 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
21898 andi16_operand): New predicates.
21899 * config/mips/mips.md (compression): New attribute.
21900 (enabled): New attribute.
21901 (length): Consider compression in computing length.
21902 (shift_compression): New code attribute.
21903 (*add<mode>3): New operands. Record compression.
21904 (sub<mode>3): Likewise.
21905 (one_cmpl<mode>2): Likewise.
21906 (*and<mode>3): Likewise.
21907 (*ior<mode>3): Likewise.
21908 (unnamed pattern for xor): Likewise.
21909 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
21910 (*<optab><mode>3): Likewise.
21911 (*mov<mode>_internal: Likewise.
21912 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
21913 (mips_unsigned_immediate_p): New.
21914 (umips_lwsp_swsp_address_p): New.
21915 (m16_based_address_p): New.
21916 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
21917 (mips_unsigned_immediate_p): New prototype.
21918 (lwsp_swsp_address_p): New prototype.
21919 (m16_based_address_p): New prototype.
21920 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
21921 (mips_signed_immediate_p): New function.
21922 (m16_based_address_p): New function.
21923 (lwsp_swsp_address_p): New function.
21924 (mips_print_operand_punctuation): Recognize short delay slot insns
21925 for microMIPS.add<mode>3"
21927 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21930 * config/arm/iterators.md (v_cmp_result): New mode attribute.
21931 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
21933 2013-03-25 Richard Biener <rguenther@suse.de>
21935 PR tree-optimization/56689
21936 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
21939 2013-03-25 Richard Biener <rguenther@suse.de>
21941 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
21943 (memory_references): Likewise.
21944 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
21945 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
21946 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
21947 (gather_mem_refs_in_loops): Fold into ...
21948 (analyze_memory_references): ... this. Move initialization
21949 to tree_ssa_lim_initialize.
21950 (fill_always_executed_in): Rename to ...
21951 (fill_always_executed_in_1): ... this.
21952 (fill_always_executed_in): Move contains_call computation to
21953 this new function from ...
21954 (tree_ssa_lim_initialize): ... here.
21955 (tree_ssa_lim): Call fill_always_executed_in.
21957 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21959 * postreload.c (reload_combine): Fix code detecting returns.
21961 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21963 * function.c (emit_use_return_register_into_block): On cc0 targets,
21964 do not emit the sequence between cc0 setter and user.
21966 2013-03-25 Kai Tietz <ktietz@redhat.com>
21968 * config/i386/predicates.md (local_symbolic_operand): Interpret
21969 dll-imported symbols as none-local.
21971 2013-03-25 Richard Biener <rguenther@suse.de>
21973 * tree-ssa-loop-im.c (struct depend): Remove.
21974 (struct lim_aux_data): Make depends a vec of gimples.
21975 (free_lim_aux_data): Adjust.
21976 (add_dependency): Likewise.
21977 (set_level): Likewise.
21979 2013-03-25 Richard Biener <rguenther@suse.de>
21981 PR middle-end/56434
21982 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
21983 the pointer returned by calls with ECF_MALLOC set.
21985 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
21987 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
21989 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
21991 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
21992 using MMXMODE mode iterator.
21993 (*move<mode>_internal): Merge with *movv2sf_internal and
21994 *movv2sf_internal_rex64 using MMXMODE mode iterator.
21996 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
21998 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
21999 (record_last_mem_set_info): Likewise.
22001 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
22002 of XNEWVEC followed by memset.
22003 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
22005 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22007 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
22008 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
22009 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
22010 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
22011 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
22012 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
22013 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
22014 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
22015 BARRIER_P instead of GET_CODE.
22017 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22019 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
22020 inaccuracy in the probing code.
22022 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
22023 (ctrapdi4): Likewise.
22025 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22027 * calls.c (expand_call): Add missing guard to code handling return
22028 of non-BLKmode structures in MSB.
22029 * function.c (expand_function_end): Likewise.
22031 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22033 * combine.c (try_combine): Adjust comment. Do not add the set of
22034 insn #0 if the destination indirectly is set or dies in insn #2.
22035 Tidy up code to distribute a new note.
22037 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22039 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
22040 also for alternatives 16 and 17.
22042 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22044 * config/i386/sse.md (*mov<mode>_internal): Merge with
22045 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
22046 Emit insn template depending on type attribute. Use
22047 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22048 movd instead of movq mnemonic for interunit moves. Rewrite mode
22049 attribute calculation. Remove unit attribute calculation.
22050 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
22051 Set prefix_data16 attribute for DImode ssemov types.
22052 Use Ym instead of y for SSE-MMX conversion alternatives.
22053 Reorder operand constraints.
22055 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
22057 * df.h (df_insn_delete): Adjust prototype.
22058 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
22059 and let it decide whether mark the basic block dirty.
22060 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
22061 * df-scan.c (df_insn_info_delete): New helper function, split
22062 off from df_insn_delete.
22063 (df_scan_free_bb_info): Use it.
22064 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
22066 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
22067 that the insn is actually an insn and it has a non-NULL basic block.
22068 Do not mark basic block dirty if only deleting a DEBUG_INSN.
22070 2013-03-22 Richard Biener <rguenther@suse.de>
22072 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
22074 (mem_ref_alloc): Do not allocate them.
22075 (refs_independent_p): Do not query or maintain a cache.
22077 2013-03-22 Richard Biener <rguenther@suse.de>
22079 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
22080 (gather_mem_refs_in_loops): Do not compute it.
22081 (analyze_memory_references): Do not allocate it.
22082 (tree_ssa_lim_finalize): Do not free it.
22083 (for_all_locs_in_loop): Do not query all_refs_in_loop.
22085 2013-03-22 Richard Biener <rguenther@suse.de>
22087 * is-a.h (as_a): Use gcc_checking_assert.
22089 2013-03-22 Ian Bolton <ian.bolton@arm.com>
22091 * config/aarch64/aarch64.c (aarch64_print_operand): New
22092 format specifier for printing a constant in hex.
22093 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
22094 format specifier for printing second operand.
22096 2013-03-22 Richard Biener <rguenther@suse.de>
22098 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
22100 (gather_mem_refs_in_loops): Perform store accumulation here.
22101 (create_vop_ref_mapping_loop): Remove.
22102 (create_vop_ref_mapping): Likewise.
22103 (analyze_memory_references): Initialize refs_stored_in_loop.
22104 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
22105 (record_indep_loop): Remove.
22106 (record_dep_loop): New function.
22107 (ref_indep_loop_p_1): Adjust to only walk over references
22108 in the loop, not its subloops.
22109 (ref_indep_loop_p): Rename to ...
22110 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
22111 maintaining a more fine-grained cache.
22112 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
22113 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
22115 2013-03-22 Richard Biener <rguenther@suse.de>
22117 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
22118 (struct mem_ref): Make accesses_in_loop a vec of a vec of
22119 aggregate mem_ref_loc.
22120 (free_mem_ref_locs): Inline into ...
22121 (memref_free): ... this and adjust.
22122 (mem_ref_alloc): Adjust.
22123 (mem_ref_locs_alloc): Remove.
22124 (record_mem_ref_loc): Adjust.
22125 (get_all_locs_in_loop): Rewrite into ...
22126 (for_all_locs_in_loop): ... this iterator.
22127 (rewrite_mem_ref_loc): New functor.
22128 (rewrite_mem_refs): Use for_all_locs_in_loop.
22129 (sm_set_flag_if_changed): New functor.
22130 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
22131 (ref_always_accessed): New functor.
22132 (ref_always_accessed_p): Use for_all_locs_in_loop.
22134 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22136 * tree-pass.h (PROP_gimple_lvec): New.
22137 * passes.c (dump_properties): Handle PROP_gimple_lvec.
22138 (init_optimization_passes): Move pass_lower_vector.
22139 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
22141 (pass_lower_vector): Provide PROP_gimple_lvec.
22142 (pass_lower_vector_ssa): Likewise.
22143 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
22145 2013-03-21 Mark Wielaard <mjw@redhat.com>
22147 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
22149 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22151 * config/i386/i386.md (*movdi_internal): Disparage slightly
22152 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
22153 conversion alternatives.
22155 2013-03-21 Jakub Jelinek <jakub@redhat.com>
22157 PR middle-end/48087
22158 * diagnostic.def (DK_WERROR): New kind.
22159 * diagnostic.h (werrorcount): Define.
22160 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
22161 promoted to DK_ERROR, increment DK_WERROR counter instead of
22163 * toplev.c (toplev_main): Call print_ignored_options even if
22164 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
22165 even if just werrorcount is non-zero.
22168 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
22170 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
22171 (string_cst_pool_decl): New function.
22172 (optimize_one_addr_into_implicit_ptr): New function.
22173 (resolve_addr_in_expr): Optimize DWARF location expression
22174 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
22175 which doesn't live in memory, but has DW_AT_location or
22176 DW_AT_const_value, or refers to a string literal, into
22177 DW_OP_GNU_implicit_pointer.
22178 (optimize_location_into_implicit_ptr): New function.
22179 (resolve_addr): If removing DW_AT_location of a variable because
22180 it was DW_OP_addr of address of the variable, but the variable doesn't
22181 live in memory, try to emit const value attribute for the initializer.
22183 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22185 * tree.h (VECTOR_TYPE_P): New macro.
22186 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
22187 TYPE_MODE): Use it.
22188 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
22189 VEC_COND_EXPR cannot be lvalues.
22190 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
22192 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22194 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
22195 Restrict the transformation to equal modes.
22197 2013-03-21 Richard Biener <rguenther@suse.de>
22199 PR tree-optimization/39326
22200 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
22201 (MEM_ANALYZABLE): Adjust.
22202 (record_mem_ref_loc): Move bitmap ops ...
22203 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
22204 unanalyzable refs, do not record locations for it.
22205 (analyze_memory_references): Allocate ref zero as shared
22207 (refs_independent_p): Do not test for unanalyzed mems here.
22208 (ref_indep_loop_p_1): Special-case disambiguation against
22209 the unanalyzed ref.
22210 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
22212 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
22214 * config/arm/arm-protos.h (tune_params): Add
22215 prefer_neon_for_64bits field.
22216 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
22217 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
22218 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
22219 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
22220 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
22221 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
22222 (arm_option_override): Handle -mneon-for-64bits new option.
22223 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
22224 (prefer_neon_for_64bits): Declare new variable.
22225 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
22226 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
22227 (arch_enabled): Handle new arch types. Remove support for onlya8
22229 (one_cmpldi2): Use new arch names.
22230 (zero_extend<mode>di2, extend<mode>di2): Ditto.
22231 * config/arm/arm.opt (mneon-for-64bits): Add option.
22232 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
22233 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
22234 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
22236 * doc/invoke.texi (-mneon-for-64bits): Document.
22238 2013-03-21 Richard Biener <rguenther@suse.de>
22240 PR tree-optimization/39326
22241 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
22242 (sort_bbs_in_loop_postorder_cmp): New function.
22243 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
22245 2013-03-21 Richard Biener <rguenther@suse.de>
22247 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
22248 (vect_insert_into_interleaving_chain): Likewise.
22249 (vect_drs_dependent_in_basic_block): Inline ...
22250 (vect_slp_analyze_data_ref_dependence): ... here. New function,
22252 (vect_analyze_data_ref_dependence): ... here. Simplify.
22253 (vect_check_interleaving): Simplify.
22254 (vect_analyze_data_ref_dependences): Likewise. Split out ...
22255 (vect_slp_analyze_data_ref_dependences): ... this new function.
22256 (dr_group_sort_cmp): New function.
22257 (vect_analyze_data_ref_accesses): Compute data-reference groups
22258 here instead of in vect_analyze_data_ref_dependence. Use
22259 a more efficient algorithm.
22260 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
22261 vect_slp_analyze_data_ref_dependences. Call
22262 vect_analyze_data_ref_accesses earlier.
22263 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
22264 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
22265 (vect_slp_analyze_data_ref_dependences): New prototype.
22267 2013-03-21 Richard Biener <rguenther@suse.de>
22269 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
22270 ref is stored in the loop.
22271 (find_refs_for_sm): Walk only over all stores.
22272 (store_motion_loop): Allocate from lim_bitmap_obstack.
22273 (store_motion): Likewise.
22275 2013-03-21 Richard Biener <rguenther@suse.de>
22277 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
22278 Update virtual SSA form.
22280 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22282 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
22283 * configure: Regenerate.
22284 * config.in: Regenerate.
22285 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
22286 if !HAVE_LD_EH_FRAME_CIEV3.
22288 2013-03-21 Richard Biener <rguenther@suse.de>
22290 * tree-cfg.c (verify_expr_no_block): New function.
22291 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
22292 nor DECL_VALUE_EXPR have locations with associated blocks.
22293 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
22294 (clear_unused_block_pointer): Remove code dealing with
22295 blocks in DECL_DEBUG_EXPR locations.
22297 2013-03-21 Richard Biener <rguenther@suse.de>
22299 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
22300 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
22301 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
22302 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
22303 instead of DECL_DEBUG_EXPR_IS_FROM.
22304 * gimplify.c (gimplify_modify_expr): Likewise.
22305 * tree-cfg.c (verify_expr_location_1): Likewise.
22306 * tree-complex.c (create_one_component_var): Likewise.
22307 * tree-sra.c (create_access_replacement): Likewise.
22308 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
22309 (clear_unused_block_pointer): Likewise.
22310 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
22311 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
22312 * var-tracking.c (var_debug_decl): Likewise.
22313 (track_expr_p): Likewise.
22314 * tree-inline.c (add_local_variables): Likewise. Set
22315 DECL_HAS_DEBUG_EXPR_P after copying it.
22316 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
22317 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
22319 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22322 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
22323 * configure: Regenerate.
22324 * config.in: Regenerate.
22325 * config/i386/i386.md (*movdf_internal): Use
22326 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22327 movd instead of movq mnemonic for interunit moves.
22328 (*movdi_internal): Ditto.
22330 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22332 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
22333 (abd<mode>_3): New pattern.
22334 (aba<mode>_3): New pattern.
22335 (fabd<mode>_3): New pattern.
22337 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22339 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
22340 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
22341 occurrence of REGISTER_PREFIX as its empty string.
22343 2013-03-20 Jeff Law <law@redhat.com>
22345 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
22346 addititional equivalences for equality comparisons between an SSA_NAME
22347 and a constant where the SSA_NAME was set from a widening conversion.
22349 2013-03-20 Walter Lee <walt@tilera.com>
22351 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
22353 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22355 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
22356 depending on type attribute.
22357 (*movti_internal): Ditto.
22358 (*movtf_internal): Ditto.
22359 (*movxf_internal): Ditto.
22360 (*movdf_internal): Ditto.
22361 (*movsf_internal): Ditto.
22363 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22365 * config/i386/i386.md (*movti_internal): Set prefix attribute to
22366 maybe_vex for sselog1 and ssemov types.
22367 (*movdi_internal): Reorder operand constraints.
22368 (*movsi_internal): Ditto. Set prefix attribute to
22369 maybe_vex for sselog1 and ssemov types.
22370 (*movtf_internal): Set prefix attribute to maybe_vex
22371 for sselog1 and ssemov types.
22372 (*movdf_internal): Ditto. Set prefix_data16 attribute for
22373 DImode ssemov types. Reorder operand constraints.
22374 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
22375 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
22376 attribute for SImode ssemov types. Reorder operand constraints.
22378 2013-03-20 Martin Jambor <mjambor@suse.cz>
22380 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
22381 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
22383 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
22385 * config/rs6000/predicates.md (indexed_address, update_address_mem
22386 update_indexed_address_mem): New predicates.
22387 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
22388 attribute for load/store instructions.
22389 * config/rs6000/dfp.md (movsd_store): Likewise.
22390 (movsd_load): Likewise.
22391 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
22392 (unnamed HI->DI extend define_insn): Likewise.
22393 (unnamed SI->DI extend define_insn): Likewise.
22394 (unnamed QI->SI extend define_insn): Likewise.
22395 (unnamed QI->HI extend define_insn): Likewise.
22396 (unnamed HI->SI extend define_insn): Likewise.
22397 (unnamed HI->SI extend define_insn): Likewise.
22398 (extendsfdf2_fpr): Likewise.
22399 (movsi_internal1): Likewise.
22400 (movsi_internal1_single): Likewise.
22401 (movhi_internal): Likewise.
22402 (movqi_internal): Likewise.
22403 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
22404 attribute for load/store instructions.
22405 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
22407 (mov<mode>_softfloat): Likewise.
22408 (mov<mode>_hardfloat32): Likewise.
22409 (mov<mode>_hardfloat64): Likewise.
22410 (mov<mode>_softfloat64): Likewise.
22411 (movdi_internal32): Likewise.
22412 (movdi_internal64): Likewise.
22413 (probe_stack_<mode>): Likewise.
22415 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
22417 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
22418 floating point, and decimal floating point to reload iterator.
22420 * config/rs6000/constraints.md (wl constraint): New constraints to
22421 return FLOAT_REGS if certain options are used to reduce the number
22422 of separate patterns that exist in the file.
22423 (wx constraint): Likewise.
22424 (wz constraint): Likewise.
22426 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22427 -mdebug=reg, print wg, wl, wx, and wz constraints.
22428 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22429 Initialize the reload functions for 64-bit binary/decimal floating
22431 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22432 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22433 create the buffer on the stack to overcome not having a 32-bit
22435 (rs6000_emit_move): Likewise.
22436 (rs6000_secondary_memory_needed_rtx): Likewise.
22437 (rs6000_alloc_sdmode_stack_slot): Likewise.
22438 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22439 via xxlxor, just like DFmode 0.0.
22441 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
22442 define as 1 if we are running on a power7 or newer.
22443 (enum r6000_reg_class_enum): Add new constraints.
22445 * config/rs6000/dfp.md (movsd): Delete, combine with binary
22446 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22447 with other moves by using conditional constraits (wg). Use LFIWZX
22448 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
22449 (movsd splitter): Likewise.
22450 (movsd_hardfloat): Likewise.
22451 (movsd_softfloat): Likewise.
22453 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
22454 binary and decimal floating point moves.
22455 (fmove_ok): New attributes to combine binary and decimal floating
22456 point moves, and to combine power6x (mfpgpr) moves along normal
22458 (real_value_to_target): Likewise.
22459 (f32_lr): Likewise.
22460 (f32_lm): Likewise.
22461 (f32_li): Likewise.
22462 (f32_sr): Likewise.
22463 (f32_sm): Likewise.
22464 (f32_si): Likewise.
22465 (movsf): Combine binary and decimal floating point moves. Combine
22466 power6x (mfpgpr) moves with other moves by using conditional
22467 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
22468 (mov<mode> for SFmode/SDmode); Likewise.
22469 (SFmode/SDmode splitters): Likewise.
22470 (movsf_hardfloat): Likewise.
22471 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
22472 (movsf_softfloat): Likewise.
22473 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
22475 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
22476 wx and wz constraints.
22478 * config/rs6000/constraints.md (wg constraint): New constraint to
22479 return FLOAT_REGS if -mmfpgpr (power6x) was used.
22481 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
22484 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22485 -mdebug=reg, print wg, wl, wx, and wz constraints.
22486 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22487 Initialize the reload functions for 64-bit binary/decimal floating
22489 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22490 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22491 create the buffer on the stack to overcome not having a 32-bit
22493 (rs6000_emit_move): Likewise.
22494 (rs6000_secondary_memory_needed_rtx): Likewise.
22495 (rs6000_alloc_sdmode_stack_slot): Likewise.
22496 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22497 via xxlxor, just like DFmode 0.0.
22499 * config/rs6000/dfp.md (movdd): Delete, combine with binary
22500 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22501 with other moves by using conditional constraits (wg). Use LFIWZX
22502 and STFIWX for loading SDmode on power7.
22503 (movdd splitters): Likewise.
22504 (movdd_hardfloat32): Likewise.
22505 (movdd_softfloat32): Likewise.
22506 (movdd_hardfloat64_mfpgpr): Likewise.
22507 (movdd_hardfloat64): Likewise.
22508 (movdd_softfloat64): Likewise.
22510 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
22511 64-bit binary and decimal floating point moves.
22512 (FMOVE64X): Likewise.
22513 (movdf): Combine 64-bit binary and decimal floating point moves.
22514 Combine power6x (mfpgpr) moves with other moves by using
22515 conditional constraits (wg).
22516 (mov<mode> for DFmode/DDmode): Likewise.
22517 (DFmode/DDmode splitters): Likewise.
22518 (movdf_hardfloat32): Likewise.
22519 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
22520 (movdf_softfloat32): Likewise.
22521 (movdf_hardfloat64_mfpgpr): Likewise.
22522 (movdf_hardfloat64): Likewise.
22523 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
22524 (movdf_softfloat64): Likewise.
22525 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
22526 (reload_<mode>_load): Move to later in the file so they aren't in
22527 the middle of the floating point move insns.
22528 (reload_<mode>_store): Likewise.
22530 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
22533 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
22534 constraint if -mdebug=reg.
22535 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
22536 Enable using dd reload support if needed.
22538 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
22539 binary and decimal floating point moves in rs6000.md.
22540 (movtd_internal): Likewise.
22542 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
22543 decimal floating point moves.
22545 (movtf_internal): Likewise.
22546 (mov<mode>_internal, TDmode/TFmode): Likewise.
22547 (movtf_softfloat): Likewise.
22548 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
22550 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
22551 movdi_internal64, using wg constraint for move direct operations.
22552 (movdi_internal64): Likewise.
22554 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
22555 MODES_TIEABLE_P for selected modes. Print the numerical value of
22556 the various virtual registers. Use GPR/FPR first/last values,
22557 instead of hard coding the register numbers. Print which modes
22558 have reload functions registered.
22559 (rs6000_option_override_internal): If -mdebug=reg, trace the options
22560 settings before/after setting cpu, target and subtarget settings.
22561 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
22562 and for secondary reload failures in rs6000_secondary_reload_inner.
22563 (rs6000_secondary_reload_fail): Likewise.
22564 (rs6000_secondary_reload_inner): Likewise.
22566 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
22567 macros for first/last GPR and FPR registers.
22568 (LAST_GPR_REGNO): Likewise.
22569 (FIRST_FPR_REGNO): Likewise.
22570 (LAST_FPR_REGNO): Likewise.
22572 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
22573 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
22574 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
22575 (vcond<mode><mode>): Likewise.
22576 (vcondu<mode><mode>): Likewise.
22577 (vector_gtu<mode>): Likewise.
22578 (vector_gte<mode>): Likewise.
22579 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
22580 to prevent the compiler from converting DImode operations to TImode.
22581 (ior<mode>3): Likewise.
22582 (and<mode>3): Likewise.
22583 (one_cmpl<mode>2): Likewise.
22584 (nor<mode>3): Likewise.
22585 (andc<mode>3): Likewise.
22587 * config/rs6000/constraints.md (wt constraint): New constraint
22588 that returns VSX_REGS if TImode is allowed in VSX registers.
22590 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
22591 constant under VSX.
22593 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
22594 similar to TImode, but it is restricted to being in the GPRs.
22596 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
22597 TImode to occupy a single VSX register.
22599 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
22600 -mvsx-timode for power7/power8.
22601 (power7 cpu): Likewise.
22602 (power8 cpu): Likewise.
22604 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
22605 sure that TFmode/TDmode take up two registers if they are ever
22606 allowed in the upper VSX registers.
22607 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
22609 (rs6000_init_hard_regno_mode_ok): Likewise.
22610 (rs6000_debug_reg_global): Add debugging for PTImode and wt
22611 constraint. Print if LRA is turned on.
22612 (rs6000_option_override_internal): Give an error if -mvsx-timode
22613 and VSX is not enabled.
22614 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
22615 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
22616 to reg+offset addressing. Use PTImode when checking offset
22617 addresses for validity.
22618 (reg_offset_addressing_ok_p): Likewise.
22619 (rs6000_legitimate_offset_address_p): Likewise.
22620 (rs6000_legitimize_address): Likewise.
22621 (rs6000_legitimize_reload_address): Likewise.
22622 (rs6000_legitimate_address_p): Likewise.
22623 (rs6000_eliminate_indexed_memrefs): Likewise.
22624 (rs6000_emit_move): Likewise.
22625 (rs6000_secondary_reload): Likewise.
22626 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
22627 reloads to fpr registers to continue to use reg+offset addressing,
22628 but 64-bit reloads to altivec registers need reg+reg addressing.
22629 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
22630 it. Treat LO_SUM like a PLUS operation.
22631 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
22632 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
22633 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
22634 registers to share a register with a smaller sized type, since VSX
22635 puts scalars in the upper 64-bits.
22636 (print_operand): Add support for PTImode.
22637 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
22638 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
22639 registers, but don't have arithmetic support.
22640 (rs6000_memory_move_cost): Add test for VSX.
22641 (rs6000_opt_masks): Add -mvsx-timode.
22643 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
22646 (VSr): Use wt constraint for TImode.
22647 (VSv): Drop TImode support.
22648 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
22649 (vsx_movti_64bit): Likewise.
22650 (vsx_movti_32bit): Likewise.
22651 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
22652 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
22653 one '?' on the appropriate output constraint. Do not allow TImode
22654 logical operations on 32-bit systems.
22655 (vsx_ior<mode>3): Likewise.
22656 (vsx_xor<mode>3): Likewise.
22657 (vsx_one_cmpl<mode>2): Likewise.
22658 (vsx_nor<mode>3): Likewise.
22659 (vsx_andc<mode>3): Likewise.
22660 (vsx_concat_<mode>): Likewise.
22661 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
22663 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
22664 OPTION_MASK_VSX_TIMODE.
22665 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
22666 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
22668 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
22669 (TI2 iterator): New iterator for TImode, PTImode.
22670 (wd mode attribute): Add values for vector types.
22671 (movti_string): Replace TI move operations with operations for TImode
22672 and PTImode. Add support for TImode being allowed in VSX registers.
22673 (mov<mode>_string, TImode/PTImode): Likewise.
22674 (movti_ppc64): Likewise.
22675 (mov<mode>_ppc64, TImode/PTImode): Likewise.
22676 (TI mode splitters): Likewise.
22678 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
22681 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
22683 PR tree-optimization/56355
22684 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
22685 Also handle integers with undefined overflow.
22687 2013-03-20 Catherine Moore <clm@codesourcery.com>
22688 Maciej W. Rozycki <macro@codesourcery.com>
22689 Tom de Vries <tom@codesourcery.com>
22690 Nathan Sidwell <nathan@codesourcery.com>
22691 Iain Sandoe <iain@codesourcery.com>
22692 Nathan Froyd <froydnj@codesourcery.com>
22693 Chao-ying Fu <fu@mips.com>
22695 * doc/extend.texi: (micromips, nomicromips, nocompression):
22696 Document new function attributes.
22697 * doc/invoke.texi (minterlink-compressed, mmicromips,
22698 m14k, m14ke, m14kec): Document new options.
22699 (minterlink-mips16): Update documentation.
22700 * doc/md.texi (ZC, ZD): Document new constraints.
22701 * configure.ac (gcc_cv_as_micromips): Check if linker
22702 supports the .set micromips directive.
22703 * configure: Regenerate.
22704 * config.in: Regenerate.
22705 * config/mips/mips-tables.opt: Regenerate.
22706 * config/mips/micromips.md: New file.
22707 * constraints.md (ZC, ZD): New constraints.
22708 * config/mips/predicates.md (movep_src_register): New predicate.
22709 (movep_src_operand): New predicate.
22710 (non_volatile_mem_operand): New predicate.
22711 * config/mips/mips.md (multimem): New type.
22712 (length): Differentiate between 17-bit and 18-bit branch offsets.
22713 (MOVEP1, MOVEP2): New mode iterator.
22714 (mov_<load>l): Use ZC constraint.
22715 (mov_<load>r): Likewise.
22716 (mov_<store>l): Likewise.
22717 (mov_<store>r): Likewise.
22718 (*branch_equality<mode>_inverted): Add microMIPS support.
22719 (*branch_equality<mode>): Likewise.
22720 (*jump_absolute): Likewise.
22721 (indirect_jump_<mode>): Likewise.
22722 (tablejump_<mode>): Likewise.
22723 (<optab>_internal): Likewise.
22724 (sibcall_internal): Likewise.
22725 (sibcall_value_internal): Likewise.
22726 (prefetch): Use constraint ZD.
22727 * config/mips/mips.opt (minterlink-compressed): New option.
22728 (minterlink-mips16): Now an alias for minterlink-compressed.
22729 (mmicromips): New option.
22730 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
22731 (compare_and_swap_12): Likewise.
22732 (sync_add<mode>): Likewise.
22733 (sync_<optab>_12): Likewise.
22734 (sync_old_<optab>_12): Likewise.
22735 (sync_new_<optab>_12): Likewise.
22736 (sync_nand_12): Likewise.
22737 (sync_old_nand_12): Likewise.
22738 (sync_new_nand_12): Likewise.
22739 (sync_sub<mode>): Likewise.
22740 (sync_old_add<mode>): Likewise.
22741 (sync_old_sub<mode>): Likewise.
22742 (sync_new_add<mode>): Likewise.
22743 (sync_new_sub<mode>): Likewise.
22744 (sync_<optab><mode>): Likewise.
22745 (sync_old_<optab><mode>): Likewise.
22746 (sync_new_<optab><mode>): Likewise.
22747 (sync_nand<mode>): Likewise.
22748 (sync_old_nand<mode>): Likewise.
22749 (sync_new_nand<mode>): Likewise.
22750 (sync_lock_test_and_set<mode>): Likewise.
22751 (test_and_set_12): Likewise.
22752 (atomic_compare_and_swap<mode>): Likewise.
22753 (atomic_exchange<mode>_llsc): Likewise.
22754 (atomic_fetch_add<mode>_llsc): Likewise.
22755 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
22756 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
22757 (umips_save_restore_pattern_p): Likewise.
22758 (umips_load_store_pair_p): Likewise.
22759 (umips_output_load_store_pair): Likewise.
22760 (umips_movep_target_p): Likewise.
22761 (umips_12bit_offset_address_p): Likewise.
22762 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
22763 (mips_base_mips16): Rename this...
22764 (mips_base_compression_flags): ...to this. Update all uses.
22765 (mips_attribute_table): Add micromips, nomicromips and nocompression.
22766 (mips_mips16_decl_p): Delete.
22767 (mips_nomips16_decl_p): Delete.
22768 (mips_get_compress_on_flags): New function.
22769 (mips_get_compress_off_flags): New function.
22770 (mips_get_compress_mode): New function.
22771 (mips_get_compress_on_name): New function.
22772 (mips_get_compress_off_name): New function.
22773 (mips_insert_attributes): Support multiple compression types.
22774 (mips_merge_decl_attributes): Likewise.
22775 (umips_12bit_offset_address_p): New function.
22776 (mips_start_function_definition): Emit .set micromips directive.
22777 (mips_call_may_need_jalx_p): New function.
22778 (mips_function_ok_for_sibcall): Add microMIPS support.
22779 (mips_print_operand_punctuation): Support short delay slots and
22781 (umips_swm_mask, umips_swm_encoding): New.
22782 (umips_build_save_restore): New function.
22783 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
22784 (was_mips16_p): Remove.
22785 (old_compression_mode): New.
22786 (mips_set_compression_mode): New function.
22787 (mips_set_current_function): Add microMIPS support.
22788 (mips_option_override): Likewise.
22789 (umips_save_restore_pattern_p): New function.
22790 (umips_output_save_restore): New function.
22791 (umips_load_store_pair_p_1): New function.
22792 (umips_load_store_pair_p): New function.
22793 (umips_output_load_store_pair_1): New function.
22794 (umips_output_load_store_pair): New function.
22795 (umips_movep_target_p) New function.
22796 (mips_prepare_pch_save): Add microMIPS support.
22797 * config/mips/mips.h (TARGET_COMPRESSION): New.
22798 (TARGET_CPU_CPP_BUILTINS): Update macro
22799 to use new compression flags and to support microMIPS.
22800 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
22801 (MIPS_ARCH_FLOAT_SPEC): Likewise.
22802 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
22803 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
22804 (ASM_SPEC): Support mmicromips and mno-micromips.
22805 (M16STORE_REG_P): New macro.
22806 (MIPS_CALL): Support TARGET_MICROMIPS.
22807 (MICROMIPS_J): New macro.
22808 (mips_base_mips16): Rename this...
22809 (mips_base_compression_flags): ...to this.
22810 (UMIPS_12BIT_OFFSET_P): New macro.
22811 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
22812 (MULTILIB_DIRNAMES): Likewise.
22813 2013-03-20 Richard Biener <rguenther@suse.de>
22815 PR tree-optimization/56661
22816 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
22817 the result does not have to be distinct.
22819 2013-03-20 Richard Biener <rguenther@suse.de>
22821 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
22824 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22825 Steven Bosscher <steven@gcc.gnu.org>
22827 PR rtl-optimization/56605
22828 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
22830 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22833 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
22834 that require movd instead of movq.
22836 2013-03-20 Richard Biener <rguenther@suse.de>
22838 * tree-ssa-structalias.c (struct variable_info): Add pointer
22839 to the first field of an aggregate with sub-vars. Make
22840 this and the pointer to the next subfield its ID.
22841 (vi_next): New function.
22842 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
22843 storedanything_id, integer_id): Increment by one.
22844 (new_var_info, get_call_vi, lookup_call_clobber_vi,
22845 get_call_clobber_vi): Adjust.
22846 (solution_set_expand): Simplify and speedup.
22847 (solution_set_add): Inline into ...
22848 (set_union_with_increment): ... this. Adjust accordingly.
22849 (do_sd_constraint): Likewise.
22850 (do_ds_constraint): Likewise.
22851 (do_complex_constraint): Simplify.
22852 (build_pred_graph): Adjust.
22853 (solve_graph): Likewise. Simplify and speedup.
22854 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
22855 get_constraint_for_component_ref, get_constraint_for_1,
22856 first_vi_for_offset, first_or_preceding_vi_for_offset,
22857 create_function_info_for, create_variable_info_for_1,
22858 create_variable_info_for, intra_create_variable_infos): Adjust.
22859 (init_base_vars): Push NULL for ID zero.
22860 (compute_points_to_sets): Adjust.
22862 2013-03-20 Richard Biener <rguenther@suse.de>
22864 * cfgloop.c (verify_loop_structure): Streamline and avoid
22865 ICEing on corrupt loop tree.
22866 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
22869 2013-03-20 Richard Biener <rguenther@suse.de>
22871 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
22872 check whether an SSA update is needed.
22874 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
22876 * config/mips/constraints.md (T): Rename to...
22880 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
22881 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
22883 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22885 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
22886 (*subsi3_carryin_uxtw): Likewise.
22888 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22890 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
22891 (*rorsi3_insn_uxtw): Likewise.
22893 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22895 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
22896 (*extrsi5_insn_uxtw): Likewise.
22898 2013-03-19 Richard Biener <rguenther@suse.de>
22900 PR tree-optimization/56273
22901 * passes.c (init_optimization_passes): Move second VRP after DOM.
22903 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
22905 * config/i386/i386.md (*movti_internal): Merge from
22906 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
22907 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
22908 nox64 isa attributes.
22910 2013-03-18 Richard Biener <rguenther@suse.de>
22912 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
22914 (merge_node_constraints): Likewise.
22915 (build_succ_graph): Likewise.
22916 (valid_graph_edge): Inline into single caller.
22917 (unify_nodes): Likewise. Use bitmap_set_bit return value
22919 (scc_visit): Fix formatting and variable use.
22920 (do_sd_constraint): Use gcc_checking_assert.
22921 (do_ds_constraint): Likewise.
22922 (do_complex_constraint): Likewise.
22923 (condense_visit): Likewise. Cleanup.
22924 (dump_pred_graph): New function.
22925 (perform_var_substitution): Dump the pred-graph before
22926 variable substitution.
22927 (find_equivalent_node): Use gcc_checking_assert.
22928 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
22930 2013-03-18 Richard Biener <rguenther@suse.de>
22932 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
22933 Remove cond_expr_stmt_list argument and do not gimplify the
22935 (vect_loop_versioning): Adjust.
22936 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
22937 Cleanup to use less temporaries.
22938 (vect_create_data_ref_ptr): Cleanup.
22940 2013-03-18 Jakub Jelinek <jakub@redhat.com>
22942 PR tree-optimization/56635
22943 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
22944 require types_compatible_p types.
22946 2013-03-18 Nick Clifton <nickc@redhat.com>
22948 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
22949 spurious backslash.
22951 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
22952 Add missing line to comment describing function.
22954 2013-03-18 Richard Biener <rguenther@suse.de>
22956 PR tree-optimization/56210
22957 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
22958 Handle string / character search functions.
22959 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
22961 2013-03-18 Richard Biener <rguenther@suse.de>
22963 PR middle-end/56483
22964 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
22965 and implement properly.
22966 * gimple.h (gimple_cond_single_var_p): Remove.
22968 2013-03-18 Richard Biener <rguenther@suse.de>
22970 * tree-data-ref.h (find_data_references_in_loop): Declare.
22971 * tree-data-ref.c (get_references_in_stmt): Use a stack
22972 vector pre-allocated in the callers.
22973 (find_data_references_in_stmt): Adjust.
22974 (graphite_find_data_references_in_stmt): Likewise.
22975 (create_rdg_vertices): Likewise.
22976 (find_data_references_in_loop): Export.
22977 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
22978 Compute dependences here...
22979 (vect_analyze_data_refs): ...not here. When we encounter
22980 a non-vectorizable data reference in basic-block vectorization
22981 truncate the data reference vector. Do not bother to
22982 fixup data-dependence information for gather loads.
22983 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
22984 of data references, as reported.
22986 2013-03-18 Richard Biener <rguenther@suse.de>
22988 PR tree-optimization/3713
22989 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
22990 has_constants and expr.
22991 (stmt_has_constants): Properly valueize SSA names when deciding
22992 whether the stmt has constants.
22994 2013-03-18 Richard Biener <rguenther@suse.de>
22996 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
22997 whole function when there is nothing to do.
22998 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
22999 * tree-vectorizer.c (vectorize_loops): Update virtual and
23000 loop-closed SSA once.
23001 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
23003 2013-03-18 Richard Biener <rguenther@suse.de>
23005 PR middle-end/56113
23006 * domwalk.c (bb_postorder): New global static.
23007 (cmp_bb_postorder): New function.
23008 (walk_dominator_tree): Replace scheme imposing an order for
23009 visiting dominator sons by one sorting them at the time they
23010 are pushed on the stack.
23012 2013-03-18 Richard Biener <rguenther@suse.de>
23014 PR tree-optimization/39326
23015 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
23016 (struct mem_ref): Replace mem member with ao_ref typed member.
23017 (MEM_ANALYZABLE): Adjust.
23018 (memref_eq): Likewise.
23019 (mem_ref_alloc): Likewise.
23020 (gather_mem_refs_stmt): Likewise.
23021 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
23022 (execute_sm_if_changed_flag_set): Adjust.
23023 (execute_sm): Likewise.
23024 (ref_always_accessed_p): Likewise.
23025 (refs_independent_p): Likewise.
23026 (can_sm_ref_p): Likewise.
23028 2013-03-18 Jakub Jelinek <jakub@redhat.com>
23031 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
23032 return 1 even for !unsignedp.
23034 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23036 * config/i386/i386.md (isa): Add x64 and nox64.
23037 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
23038 (*pushtf): Enable *roF alternative for x64 isa only.
23039 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
23040 mode attribute of integer alternatives to DImode for TARGET_64BIT.
23041 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
23042 (*movtf_internal): Merge from *movtf_internal_rex64 and
23043 *movtf_internal_sse. Use x64 and nox64 isa attributes.
23044 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
23045 nox64 isa attributes.
23046 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
23047 nox64 isa attributes.
23048 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
23050 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23052 * config/alpha/alpha.c (TARGET_LRA_P): New define.
23054 2013-03-17 Jakub Jelinek <jakub@redhat.com>
23057 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
23058 class names. Remove trailing comma after "ALL_REGS".
23060 2013-03-16 Jan Hubicka <jh@suse.cz>
23062 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
23063 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
23064 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
23065 of cgraph_get_create_node.
23066 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
23068 2013-03-16 Jason Merrill <jason@redhat.com>
23071 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
23072 with DW_AT_default_value.
23074 2013-03-16 Jakub Jelinek <jakub@redhat.com>
23076 * BASE-VER: Set to 4.9.0.
23078 2013-03-14 Andi Kleen <ak@linux.intel.com>
23081 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
23082 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
23083 Document _x* TSX intrinsics.
23085 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
23086 David Holsgrove <david.holsgrove@xilinx.com>
23088 * configure.ac: Add MicroBlaze TLS support detection.
23089 * configure: Regenerate.
23090 * config/microblaze/microblaze-protos.h
23091 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
23092 symbol_mentioned_p, label_mentioned_p): Add prototypes.
23093 * config/microblaze/microblaze.c (microblaze_address_type): Add
23094 ADDRESS_TLS and tls_reloc address types.
23095 (microblaze_address_info): Add tls_reloc.
23096 (TARGET_HAVE_TLS): Define.
23097 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
23098 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
23099 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
23100 load_tls_operand, microblaze_call_tls_get_addr,
23101 microblaze_legitimize_tls_address): New functions.
23102 (microblaze_classify_unspec): Handle UNSPEC_TLS.
23103 (get_base_reg): Use microblaze_tls_symbol_p.
23104 (microblaze_classify_address): Handle TLS.
23105 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
23106 label_mentioned_p and microblaze_tls_referenced_p.
23107 (microblaze_legitimize_address): Handle TLS.
23108 (microblaze_address_insns): Handle ADDRESS_TLS.
23109 (pic_address_needs_scratch): Handle TLS.
23110 (print_operand_address): Handle TLS.
23111 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
23112 (microblaze_expand_move): Handle TLS.
23113 (microblaze_legitimate_constant_p): Check
23114 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
23115 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
23116 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
23117 (PIC_OFFSET_TABLE_REGNUM): Set.
23118 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
23119 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
23120 (addsi3, movsi_internal2, movdf_internal): Update constraints
23121 * config/microblaze/predicates.md (arith_plus_operand): Define
23122 (move_operand): Redefine as move_src_operand,
23123 check microblaze_tls_referenced_p.
23125 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23127 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
23128 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
23130 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23132 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
23135 2013-03-14 Jakub Jelinek <jakub@redhat.com>
23137 PR tree-optimization/53265
23138 * common.opt (Waggressive-loop-optimizations): New option.
23139 * tree-ssa-loop-niter.c: Include tree-pass.h.
23140 (do_warn_aggressive_loop_optimizations): New function.
23141 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
23142 if number_of_latch_executions returned constant.
23143 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
23144 early. If number_of_latch_executions returned constant, set
23145 nb_iterations_upper_bound back to it.
23146 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
23148 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
23149 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
23151 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
23152 (MULTILIB_OSDIRNAMES): Set.
23153 * genmultilib: If defaultosdirname doesn't start with :: , set
23154 defaultosdirname2 instead, clear it and emit two . multilib_raw
23155 entries instead of just one.
23157 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
23159 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23160 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
23161 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23162 (SUBTARGET_OVERRIDE_OPTIONS): New.
23164 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
23167 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
23168 (musermode): Convert to Var(TARGET_USERMODE).
23169 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
23170 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
23171 * config/sh/sh.c (sh_option_override): Use
23172 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
23173 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
23175 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
23177 (udivsi3_i4_single, divsi3_i4_single): Use
23178 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
23180 2013-03-13 Dave Korn <dave.korn.cygwin@....>
23182 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
23185 2013-03-13 Richard Biener <rguenther@suse.de>
23187 PR tree-optimization/56608
23188 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
23189 calls when vectorizing basic-blocks.
23191 2013-03-13 Jakub Jelinek <jakub@redhat.com>
23194 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
23197 2013-03-12 Jakub Jelinek <jakub@redhat.com>
23199 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
23201 2013-03-11 Jan Hubicka <jh@suse.cz>
23204 * lto-streamer-out.c (output_symbol_p): Skip references from
23205 constructors of external variables.
23207 2013-03-11 Jan Hubicka <jh@suse.cz>
23209 PR middle-end/56571
23210 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
23212 * emit-rtl.c (verify_rtx_sharing): Likewise.
23213 (copy_insn_1): Likewise.
23214 * rtl.c (copy_rtx): Likewise.
23216 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
23219 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
23220 output_operand_lossage message.
23222 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
23225 * arm.c (shift_op): Validate RTL pattern on the fly.
23226 (arm_print_operand, case 'S'): Don't use shift_operator to validate
23229 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23232 * config/pa/pa.md (call_value): Check for calls to powf and direct to
23233 new call patterns that clobber %fr12.
23234 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
23235 split and postreload patterns.
23236 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
23237 registers %fr12 and %fr12R as call used.
23239 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
23241 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
23242 (canon_address, record_store, replace_read, check_mem_read_rtx,
23243 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
23244 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
23245 rest_of_handle_dse): Likewise.
23247 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
23249 PR middle-end/56524
23250 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
23252 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
23253 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
23254 (save_optabs_if_changed): Replace with...
23255 (init_tree_optimization_optabs): ...this.
23256 * optabs.c (save_optabs_if_changed): Rename to...
23257 (init_tree_optimization_optabs): ...this. Take the optimization node
23258 as argument. Do nothing if the base optabs are already correct.
23259 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
23260 to recompute optabs.
23261 * function.h (function): Remove optabs field.
23262 * function.c (invoke_set_current_function_hook): Call
23263 init_tree_optimization_optabs. Use the result to initialize
23266 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
23268 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
23269 if GTMA_HAS_NO_INSTRUMENTATION.
23270 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
23271 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
23272 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
23273 * gimple-pretty-print.c (dump_gimple_transaction): Handle
23274 GTMA_HAS_NO_INSTRUMENTATION.
23276 2013-03-08 Jakub Jelinek <jakub@redhat.com>
23278 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
23281 2013-03-08 Marek Polacek <polacek@redhat.com>
23282 Jakub Jelinek <jakub@redhat.com>
23284 PR tree-optimization/56478
23285 * predict.c (is_comparison_with_loop_invariant_p): Change the
23286 type of loop_step to tree.
23287 (predict_loops): Adjust.
23288 (predict_iv_comparison): Perform the computations on double_ints.
23290 2013-03-08 Richard Biener <rguenther@suse.de>
23292 PR tree-optimization/56570
23293 * tree-cfg.c (verify_expr_location_1): Verify locations for
23295 * tree-sra.c (create_access_replacement): Strip locations
23296 from DECL_DEBUG_EXPRs.
23298 2013-03-08 Richard Biener <rguenther@suse.de>
23300 * tree-inline.c (expand_call_inline): Do not associate
23301 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
23302 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
23304 2013-03-08 Richard Biener <rguenther@suse.de>
23306 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
23307 or block changes with -Og. Fix for location / block encoding
23308 changes and PHI arguments with locations.
23310 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
23312 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
23314 (struct output_info): Likewise.
23315 (register_overhead): Remove bad gcc_assert.
23316 (bitmap_find_bit): If there is only a single bitmap element, do not
23317 count a miss as a search.
23318 (print_statistics): Update for counter type changes.
23319 (dump_bitmap_statistics): Likewise. Print headers such that they
23320 are properly lined up with the printed counters.
23322 2013-03-07 Jakub Jelinek <jakub@redhat.com>
23324 PR tree-optimization/56559
23325 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
23326 check that it has only a single use.
23328 2013-03-07 Richard Biener <rguenther@suse.de>
23330 * doc/invoke.texi (fwhole-program): Discourage use in combination
23333 2013-03-06 Jakub Jelinek <jakub@redhat.com>
23335 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
23337 PR tree-optimization/56539
23338 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
23339 instead of GSI_CONTINUE_LINKING as last argument to
23340 force_gimple_operand_gsi. Adjust function comment.
23342 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
23345 PR middle-end/56548
23346 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
23347 promoted mode, convert the result back to the original mode.
23349 2013-03-06 Richard Biener <rguenther@suse.de>
23351 PR middle-end/56294
23352 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
23353 (insert_updated_phi_nodes_compare_uids): New function.
23354 (update_ssa): Sort symbols_to_rename after UID before
23355 traversing it to insert PHI nodes.
23357 2013-03-06 Richard Biener <rguenther@suse.de>
23359 PR middle-end/50494
23360 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
23361 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
23364 2013-02-13 Richard Biener <rguenther@suse.de>
23367 * varasm.c (output_constant_def_1): Get the decl representing
23368 the constant as argument.
23369 (output_constant_def): Wrap output_constant_def_1.
23370 (make_decl_rtl): Use output_constant_def_1 with the decl
23371 representing the constant.
23372 (build_constant_desc): Optionally re-use a decl already
23373 representing the constant.
23374 (tree_output_constant_def): Adjust.
23376 2013-03-06 Joey Ye <joey.ye@arm.com>
23379 * gcc.c (convert_white_space): New function.
23380 (main): Handles white space in function name.
23382 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
23385 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
23386 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
23387 to SH_DIV_CALL_TABLE for TARGET_SH2.
23388 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
23390 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
23391 call-table options.
23393 2013-03-05 Sterling Augustine <saugustine@google.com>
23394 Cary Coutant <ccoutant@google.com>
23397 * dwarf2out.c (resolve_addr): Don't call
23398 remove_loc_list_addr_table_entries a second time for the same
23401 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23404 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
23405 (avoid_complex_debug_insns): New function.
23406 (expand_debug_locations): Call it.
23408 PR rtl-optimization/56484
23409 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
23410 lifetimes of hard registers on small register class machines.
23412 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23414 * config/microblaze/microblaze-protos.h: Rename
23415 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
23416 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
23418 (microblaze_fast_interrupt_function_p): New function.
23419 (microblaze_is_interrupt_handler): Rename to
23420 microblaze_is_interrupt_variant and add fast_interrupt check.
23421 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
23422 (save_restore_insns): Likewise.
23423 (compute_frame_size): Likewise.
23424 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
23425 (microblaze_globalize_label): Likewise.
23426 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
23427 * config/microblaze/microblaze.md: Use wrapper
23428 microblaze_is_interrupt_variant.
23430 2013-03-05 Kai Tietz <ktietz@redhat.com>
23432 * sdbout.c (sdbout_one_type): Switch to current function's section
23433 supporting cold/hot.
23435 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23437 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
23440 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23442 PR middle-end/56461
23443 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
23444 if VALGRIND_GET_VBITS is defined, temporarily make object
23445 memory all defined, and restore previous valgrind addressability
23446 and definability afterwards. Free this_object at the end.
23448 PR middle-end/56461
23449 * lra.c (lra): Call lra_clear_live_ranges if live_p,
23450 right before calling lra_create_live_ranges, also call it
23451 when clearing live_p. Only call lra_clear_live_ranges
23452 at the end if live_p.
23454 PR middle-end/56461
23455 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
23457 2013-03-05 Richard Biener <rguenther@suse.de>
23459 PR tree-optimization/56521
23460 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
23463 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
23466 * except.h (remove_unreachable_eh_regions): New prototype.
23467 * except.c (remove_eh_handler_splicer): New function, split out
23468 of remove_eh_handler.
23469 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
23470 warning about running it on many EH regions one at a time.
23471 (remove_unreachable_eh_regions_worker): New function, walk the
23472 EH tree in depth-first order and remove non-marked regions.
23473 (remove_unreachable_eh_regions): New function.
23474 * tree-eh.c (mark_reachable_handlers): New function, split out
23475 from remove_unreachable_handlers.
23476 (remove_unreachable_handlers): Use mark_reachable_handlers and
23477 remove_unreachable_eh_regions.
23478 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
23479 and remove_unreachable_eh_regions.
23481 2013-03-05 Richard Biener <rguenther@suse.de>
23483 PR middle-end/56525
23484 * loop-init.c (fix_loop_structure): Remove loops in two stages,
23485 not freeing them until the end.
23487 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23489 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
23491 2013-03-05 Richard Biener <rguenther@suse.de>
23493 PR tree-optimization/56270
23494 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
23495 of loads after scheduling an SLP instance.
23497 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23499 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
23501 (check_gcc_parallelize): Run guality.exp as a separate job from
23502 vect.exp with unsorted.exp and $(dg_target_exps) separately from
23503 struct-layout-1.exp with stackalign.exp.
23505 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
23507 PR middle-end/56461
23508 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
23509 load_index sbitmap even if some bit in it isn't set.
23511 PR middle-end/56461
23512 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
23513 (discover_iteration_bound_by_body_walk): Change queues to
23514 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
23515 spelling in comment. Call safe_push on queues[bound_index] directly.
23516 Release queues[queue_index] in every iteration unconditionally.
23517 Release bounds vector.
23519 PR middle-end/56461
23520 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
23521 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
23522 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
23525 2013-03-05 Richard Biener <rguenther@suse.de>
23528 * tree-inline.c (remap_blocks_to_null): New function.
23529 (expand_call_inline): When expanding a call stmt without
23530 an associated block inline remap all callee blocks to NULL.
23532 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23534 PR rtl-optimization/56494
23535 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
23536 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
23537 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
23539 PR middle-end/56461
23540 * sel-sched-ir.c (free_sched_pools): Release
23541 succs_info_pool.stack[succs_info_pool.max_top] vectors too
23542 if succs_info_pool.max_top isn't -1.
23545 * opts.c (opts_obstack, opts_concat): Moved to...
23546 * opts-common.c (opts_obstack, opts_concat): ... here.
23548 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23550 PR middle-end/56461
23551 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
23553 2013-03-04 Martin Jambor <mjambor@suse.cz>
23555 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
23556 all appropriate places.
23558 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
23560 PR tree-optimization/56424
23561 * ipa-split.c (split_function): Do not set the RSO flag if result is
23562 not by reference and its type is a register type.
23564 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
23566 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
23567 (microblaze_legitimate_pic_operand): Likewise
23568 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
23569 new function microblaze_legitimate_pic_operand
23570 * config/microblaze/microblaze-protos.h
23571 (microblaze_legitimate_pic_operand): Declare.
23573 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23575 * config/microblaze/predicates.md (call_insn_simple_operand):
23576 New predicate for supported rtx code types.
23577 * config/microblaze/microblaze.md (call_internal1): Use
23578 call_insn_simple_operand predicate.
23580 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23582 PR middle-end/56461
23583 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
23584 partitions.ordered_remove.
23586 PR middle-end/56461
23587 * tree-vect-stmts.c (vectorizable_conversion): Don't call
23588 vec_oprnds0.create (1) for modifier == NONE.
23590 PR middle-end/56461
23591 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
23592 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
23593 vec_oprnds1 right before pushing anything to it for
23596 PR middle-end/56461
23597 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
23598 set nbbs to 0 instead of having separate code path.
23599 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
23600 instead of false as last argument if returning NULL.
23602 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
23604 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
23605 the attribute is now called "target" instead of "option".
23606 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
23607 * doc/tm.texi.in (Target Attributes): Likewise document the correct
23608 attribute/pragma name for TARGET_OPTION_VALID_P and
23609 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
23610 * doc/tm.texi: Regenerated.
23612 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
23614 * config/microblaze/microblaze.c:
23615 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
23616 * config/microblaze/microblaze.h: Add -mxl-reorder to
23618 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
23619 instructions emitted if TARGET_REORDER.
23620 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
23621 or 0 for -m/-mno case, but initialises as 2 to detect default use case
23624 2013-03-01 Xinliang David Li <davidxl@google.com>
23626 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
23629 2013-03-01 Jakub Jelinek <jakub@redhat.com>
23631 PR middle-end/56461
23632 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
23633 vector even when returning true. Fix up function comment formatting.
23635 PR middle-end/56461
23636 * ira-build.c (ira_loop_nodes_count): New variable.
23637 (create_loop_tree_nodes): Initialize it.
23638 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
23640 PR middle-end/56461
23641 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
23642 method on dr_chain and result_chain.
23643 * tree-vect-stmts.c (vectorizable_store): Only call
23644 result_chain.create if j == 0.
23646 PR middle-end/56461
23647 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
23648 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
23649 before overwriting it.
23651 2013-03-01 Tobias Burnus <burnus@net-b.de>
23653 * doc/extended.texi (C Extensions): Change order in @menu
23655 (Other MIPS Built-in Functions): Move last MIPS entry before
23656 "picoChip Built-in Functions".
23657 (SH Built-in Functions): Move after RX Built-in Functions.
23658 * doc/gcc.texi (Introduction): Change order in @menu
23660 * doc/md.texi (Constraints): Ditto.
23661 * gty.texi (Type Information): Ditto.
23662 (User-provided marking routines for template types): Make
23664 * doc/invoke.texi (AArch64 Options): Move before
23665 "Adapteva Epiphany Options".
23667 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
23668 Jakub Jelinek <jakub@redhat.com>
23671 * asan.c (gate_asan): Lookup no_sanitize_address instead of
23672 no_address_safety_analysis attribute.
23673 * doc/extend.texi (no_address_safety_attribute): Rename to
23674 no_sanitize_address attribute, mention no_address_safety_analysis
23675 attribute as deprecated alias.
23677 2013-02-28 Jakub Jelinek <jakub@redhat.com>
23679 PR middle-end/56461
23680 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
23681 type to vec<vec<tree> > *.
23682 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
23683 to be vec<tree> instead of vec<tree> *, set vec_defs
23684 to vNULL and call vec_defs.create (number_of_vects), adjust other
23686 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
23687 vectorizable_condition): Adjust vect_get_slp_defs callers.
23689 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23691 * config/aarch64/aarch64.c
23692 (aarch64_float_const_representable): Remove unused variable.
23694 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23696 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
23698 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23700 * config/aarch64/aarch64-builtins.c
23701 (aarch64_init_simd_builtins): Make static.
23703 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23705 * config/aarch64/aarch64.c
23706 (aarch64_simd_make_constant): Make static.
23708 2013-02-28 Martin Jambor <mjambor@suse.cz>
23710 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
23711 with no initialization to the RHS of debug statements.
23713 2013-02-28 Martin Jambor <mjambor@suse.cz>
23715 PR tree-optimization/56294
23716 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
23718 (get_access_replacement): Do not call create_access_replacement.
23719 Assert a replacement exists.
23720 (get_repl_default_def_ssa_name): Create the replacement declaration
23723 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23725 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
23726 final_end_function.
23728 2013-02-28 Marek Polacek <polacek@redhat.com>
23730 PR rtl-optimization/56466
23731 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
23732 if we're changing a loop.
23733 (peel_loops_completely): Likewise.
23735 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
23738 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
23740 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23743 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
23744 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
23745 INTX_FTYPE_FX, FX_FTYPE_INTX.
23746 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
23748 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23750 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
23751 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
23752 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
23753 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
23754 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
23755 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
23756 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
23757 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
23758 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
23759 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
23760 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
23761 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
23762 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
23763 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
23764 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
23765 (avrxmega6): Increase max flash segments from 5 to 6.
23766 * config/avr/t-multilib: Regenerate.
23767 * config/avr/avr-tables.opt: Regenerate.
23768 * doc/avr-mmcu.texi: Regenerate.
23770 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23772 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
23773 (avr_device_to_arch): Rename to avr_device_to_ld.
23774 (avr_device_to_as): New prototype.
23775 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
23776 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
23777 * config/avr/driver-avr.c (avr_device_to_as): New.
23778 (avr_device_to_arch): Rename to avr_device_to_ld.
23780 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23782 PR middle-end/56461
23783 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
23784 method on dr_chain and result_chain.
23786 PR middle-end/56461
23787 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
23788 pointer_set_destroy on not_executed_last_iteration.
23790 PR middle-end/56461
23791 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
23793 PR middle-end/56461
23794 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
23795 FOR_EACH_DEFINED_FUNCTION when freeing state.
23797 PR middle-end/56461
23798 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
23800 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
23803 PR middle-end/56461
23804 * ipa-cp.c (decide_whether_version_node): Call vec_free on
23805 known_aggs[i].items and release known_aggs vector.
23807 PR middle-end/56461
23808 * ipa-reference.c (propagate): Free node_info even for alias nodes.
23810 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23812 * config/microblaze/microblaze.c (microblaze_emit_compare):
23813 Use xor for EQ/NE comparisions.
23814 * config/microblaze/microblaze.md (cstoresf4): Add constraints
23815 (cbranchsf4): Adjust operator to comparison_operator.
23817 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23819 PR middle-end/56461
23820 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
23822 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
23823 vec_safe_push, always update *slot.
23824 (redirect_edge_var_map_clear): Use vec_free.
23825 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
23826 (free_var_map_entry): Use vec_free.
23827 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
23828 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
23830 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
23832 PR middle-end/45472
23833 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
23834 when the may_trap_p bit of the exprs being merged differs.
23835 Reorder tests for speculativeness in the logical and operator.
23837 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23839 * incpath.c (add_standard_paths): Use reconcat instead of concat
23840 where appropriate and avoid leaking memory.
23842 * opts.h: Include obstack.h.
23843 (opts_concat): New prototype.
23844 (opts_obstack): New declaration.
23845 * opts.c (opts_concat): New function.
23846 (opts_obstack): New variable.
23847 (init_options_struct): Call gcc_init_obstack on opts_obstack.
23848 (finish_options): Use opts_concat instead of concat
23849 and XOBNEWVEC instead of XNEWVEC.
23850 * opts-common.c (generate_canonical_option, decode_cmdline_option,
23851 generate_option): Likewise.
23852 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
23853 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
23856 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
23857 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
23859 2013-02-26 Jakub Jelinek <jakub@redhat.com>
23861 PR middle-end/56461
23862 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
23864 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23866 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
23867 (arm_block_move_unaligned_straight): Likewise.
23868 (arm_adjust_block_mem): Likewise.
23870 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23873 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
23874 temp, cond and label.
23875 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
23878 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
23879 * config/c6x/c6x.h (dbx_register_map): Update declaration.
23882 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
23883 of prologue/epilogue functions.
23886 * config/tilegx/tilegx.c (tilegx_expand_prologue):
23887 Remove unused variable cfa_offset.
23888 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
23891 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
23892 type promotion to unsigned.
23895 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
23896 for HOST_WIDE_INT of 32 bit / same size as int.
23897 (arm_block_move_unaligned_straight): Likewise.
23898 (arm_adjust_block_mem): Likewise.
23901 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
23904 2013-02-26 Marek Polacek <polacek@redhat.com>
23906 PR tree-optimization/56426
23907 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
23909 2013-02-26 Richard Biener <rguenther@suse.de>
23912 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
23913 unused variable loops.
23915 2013-02-26 Jakub Jelinek <jakub@redhat.com>
23917 PR tree-optimization/56448
23918 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
23919 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
23920 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
23921 later operands of the references, or even first operand for
23922 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
23924 PR tree-optimization/56443
23925 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
23926 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
23927 to type_for_mode langhook.
23929 2013-02-25 Matt Turner <mattst88@gmail.com>
23931 * doc/invoke.texi: Document r4700.
23933 2013-02-25 Richard Biener <rguenther@suse.de>
23935 PR tree-optimization/56175
23936 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
23938 (simplify_bitwise_binary): ... here. Also guard the conversion
23939 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
23941 2013-02-25 Catherine Moore <clm@codesourcery.com>
23944 2013-02-24 Catherine Moore <clm@codesourcery.com>
23945 Maciej W. Rozycki <macro@codesourcery.com>
23946 Tom de Vries <tom@codesourcery.com>
23947 Nathan Sidwell <nathan@codesourcery.com>
23948 Iain Sandoe <iain@codesourcery.com>
23949 Nathan Froyd <froydnj@codesourcery.com>
23950 Chao-ying Fu <fu@mips.com>
23952 * doc/extend.texi: (micromips, nomicromips, nocompression):
23953 Document new function attributes.
23954 * doc/invoke.texi (minterlink-compressed, mmicromips,
23955 m14k, m14ke, m14kec): Document new options.
23956 (minterlink-mips16): Update documentation.
23957 * doc/md.texi (ZC, ZD): Document new constraints.
23958 * configure.ac (gcc_cv_as_micromips): Check if linker
23959 supports the .set micromips directive.
23960 * configure: Regenerate.
23961 * config.in: Regenerate.
23962 * config/mips/mips-tables.opt: Regenerate.
23963 * config/mips/micromips.md: New file.
23964 * constraints.md (ZC, AD): New constraints.
23965 * config/mips/predicates.md (movep_src_register): New predicate.
23966 (movep_src_operand): New predicate.
23967 (non_volatile_mem_operand): New predicate.
23968 * config/mips/mips.md (multimem): New type.
23969 (length): Differentiate between 17-bit and 18-bit branch offsets.
23970 (MOVEP1, MOVEP2): New mode iterator.
23971 (mov_<load>l): Use ZC constraint.
23972 (mov_<load>r): Likewise.
23973 (mov_<store>l): Likewise.
23974 (mov_<store>r): Likewise.
23975 (*branch_equality<mode>_inverted): Add microMIPS support.
23976 (*branch_equality<mode>): Likewise.
23977 (*jump_absolute): Likewise.
23978 (indirect_jump_<mode>): Likewise.
23979 (tablejump_<mode>): Likewise.
23980 (<optab>_internal): Likewise.
23981 (sibcall_internal): Likewise.
23982 (sibcall_value_internal): Likewise.
23983 (prefetch): Use constraint ZD.
23984 * config/mips/mips.opt (minterlink-compressed): New option.
23985 (minterlink-mips16): Now an alias for minterlink-compressed.
23986 (mmicromips): New option.
23987 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23988 (compare_and_swap_12): Likewise.
23989 (sync_add<mode>): Likewise.
23990 (sync_<optab>_12): Likewise.
23991 (sync_old_<optab>_12): Likewise.
23992 (sync_new_<optab>_12): Likewise.
23993 (sync_nand_12): Likewise.
23994 (sync_old_nand_12): Likewise.
23995 (sync_new_nand_12): Likewise.
23996 (sync_sub<mode>): Likewise.
23997 (sync_old_add<mode>): Likewise.
23998 (sync_old_sub<mode>): Likewise.
23999 (sync_new_add<mode>): Likewise.
24000 (sync_new_sub<mode>): Likewise.
24001 (sync_<optab><mode>): Likewise.
24002 (sync_old_<optab><mode>): Likewise.
24003 (sync_new_<optab><mode>): Likewise.
24004 (sync_nand<mode>): Likewise.
24005 (sync_old_nand<mode>): Likewise.
24006 (sync_new_nand<mode>): Likewise.
24007 (sync_lock_test_and_set<mode>): Likewise.
24008 (test_and_set_12): Likewise.
24009 (atomic_compare_and_swap<mode>): Likewise.
24010 (atomic_exchange<mode>_llsc): Likewise.
24011 (atomic_fetch_add<mode>_llsc): Likewise.
24012 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24013 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24014 (umips_save_restore_pattern_p): Likewise.
24015 (umips_load_store_pair_p): Likewise.
24016 (umips_output_load_store_pair): Likewise.
24017 (umips_movep_target_p): Likewise.
24018 (umips_12bit_offset_address_p): Likewise.
24019 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24020 (mips_base_mips16): Rename this...
24021 (mips_base_compression_flags): ...to this. Update all uses.
24022 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24023 (mips_mips16_decl_p): Delete.
24024 (mips_nomips16_decl_p): Delete.
24025 (mips_get_compress_on_flags): New function.
24026 (mips_get_compress_off_flags): New function.
24027 (mips_get_compress_mode): New function.
24028 (mips_get_compress_on_name): New function.
24029 (mips_get_compress_off_name): New function.
24030 (mips_insert_attributes): Support multiple compression types.
24031 (mips_merge_decl_attributes): Likewise.
24032 (umips_12bit_offset_address_p): New function.
24033 (mips_start_function_definition): Emit .set micromips directive.
24034 (mips_call_may_need_jalx_p): New function.
24035 (mips_function_ok_for_sibcall): Add microMIPS support.
24036 (mips_print_operand_punctuation): Support short delay slots and
24038 (umips_swm_mask, umips_swm_encoding): New.
24039 (umips_build_save_restore): New function.
24040 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24041 (was_mips16_p): Remove.
24042 (old_compression_mode): New.
24043 (mips_set_compression_mode): New function.
24044 (mips_set_current_function): Add microMIPS support.
24045 (mips_option_override): Likewise.
24046 (umips_save_restore_pattern_p): New function.
24047 (umips_output_save_restore): New function.
24048 (umips_load_store_pair_p_1): New function.
24049 (umips_load_store_pair_p): New function.
24050 (umips_output_load_store_pair_1): New function.
24051 (umips_output_load_store_pair): New function.
24052 (umips_movep_target_p) New function.
24053 (mips_prepare_pch_save): Add microMIPS support.
24054 * config/mips/mips.h (TARGET_COMPRESSION): New.
24055 (TARGET_CPU_CPP_BUILTINS): Update macro
24056 to use new compression flags and to support microMIPS.
24057 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24058 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24059 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24060 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24061 (ASM_SPEC): Support mmicromips and mno-micromips.
24062 (M16STORE_REG_P): New macro.
24063 (MIPS_CALL): Support TARGET_MICROMIPS.
24064 (MICROMIPS_J): New macro.
24065 (mips_base_mips16): Rename this...
24066 (mips_base_compression_flags): ...to this.
24067 (UMIPS_12BIT_OFFSET_P): New macro.
24068 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24069 (MULTILIB_DIRNAMES): Likewise.
24071 2013-02-25 Tom de Vries <tom@codesourcery.com>
24073 PR rtl-optimization/56131
24074 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
24075 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
24076 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
24078 2013-02-25 Tobias Burnus <burnus@net-b.de>
24080 * doc/invoke.texi (-fsanitize=): Move from optimization
24081 to debugging options.
24083 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
24085 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
24087 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
24088 Alexander Monakov <amonakov@ispras.ru>
24090 PR middle-end/56077
24091 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
24092 flush pending lists also on non-jumps. Adjust comment.
24094 2013-02-24 Catherine Moore <clm@codesourcery.com>
24095 Maciej W. Rozycki <macro@codesourcery.com>
24096 Tom de Vries <tom@codesourcery.com>
24097 Nathan Sidwell <nathan@codesourcery.com>
24098 Iain Sandoe <iain@codesourcery.com>
24099 Nathan Froyd <froydnj@codesourcery.com>
24100 Chao-ying Fu <fu@mips.com>
24102 * doc/extend.texi: (micromips, nomicromips, nocompression):
24103 Document new function attributes.
24104 * doc/invoke.texi (minterlink-compressed, mmicromips,
24105 m14k, m14ke, m14kec): Document new options.
24106 (minterlink-mips16): Update documentation.
24107 * doc/md.texi (ZC, ZD): Document new constraints.
24108 * configure.ac (gcc_cv_as_micromips): Check if linker
24109 supports the .set micromips directive.
24110 * configure: Regenerate.
24111 * config.in: Regenerate.
24112 * config/mips/mips-tables.opt: Regenerate.
24113 * config/mips/micromips.md: New file.
24114 * constraints.md (ZC, AD): New constraints.
24115 * config/mips/predicates.md (movep_src_register): New predicate.
24116 (movep_src_operand): New predicate.
24117 (non_volatile_mem_operand): New predicate.
24118 * config/mips/mips.md (multimem): New type.
24119 (length): Differentiate between 17-bit and 18-bit branch offsets.
24120 (MOVEP1, MOVEP2): New mode iterator.
24121 (mov_<load>l): Use ZC constraint.
24122 (mov_<load>r): Likewise.
24123 (mov_<store>l): Likewise.
24124 (mov_<store>r): Likewise.
24125 (*branch_equality<mode>_inverted): Add microMIPS support.
24126 (*branch_equality<mode>): Likewise.
24127 (*jump_absolute): Likewise.
24128 (indirect_jump_<mode>): Likewise.
24129 (tablejump_<mode>): Likewise.
24130 (<optab>_internal): Likewise.
24131 (sibcall_internal): Likewise.
24132 (sibcall_value_internal): Likewise.
24133 (prefetch): Use constraint ZD.
24134 * config/mips/mips.opt (minterlink-compressed): New option.
24135 (minterlink-mips16): Now an alias for minterlink-compressed.
24136 (mmicromips): New option.
24137 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24138 (compare_and_swap_12): Likewise.
24139 (sync_add<mode>): Likewise.
24140 (sync_<optab>_12): Likewise.
24141 (sync_old_<optab>_12): Likewise.
24142 (sync_new_<optab>_12): Likewise.
24143 (sync_nand_12): Likewise.
24144 (sync_old_nand_12): Likewise.
24145 (sync_new_nand_12): Likewise.
24146 (sync_sub<mode>): Likewise.
24147 (sync_old_add<mode>): Likewise.
24148 (sync_old_sub<mode>): Likewise.
24149 (sync_new_add<mode>): Likewise.
24150 (sync_new_sub<mode>): Likewise.
24151 (sync_<optab><mode>): Likewise.
24152 (sync_old_<optab><mode>): Likewise.
24153 (sync_new_<optab><mode>): Likewise.
24154 (sync_nand<mode>): Likewise.
24155 (sync_old_nand<mode>): Likewise.
24156 (sync_new_nand<mode>): Likewise.
24157 (sync_lock_test_and_set<mode>): Likewise.
24158 (test_and_set_12): Likewise.
24159 (atomic_compare_and_swap<mode>): Likewise.
24160 (atomic_exchange<mode>_llsc): Likewise.
24161 (atomic_fetch_add<mode>_llsc): Likewise.
24162 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24163 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24164 (umips_save_restore_pattern_p): Likewise.
24165 (umips_load_store_pair_p): Likewise.
24166 (umips_output_load_store_pair): Likewise.
24167 (umips_movep_target_p): Likewise.
24168 (umips_12bit_offset_address_p): Likewise.
24169 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24170 (mips_base_mips16): Rename this...
24171 (mips_base_compression_flags): ...to this. Update all uses.
24172 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24173 (mips_mips16_decl_p): Delete.
24174 (mips_nomips16_decl_p): Delete.
24175 (mips_get_compress_on_flags): New function.
24176 (mips_get_compress_off_flags): New function.
24177 (mips_get_compress_mode): New function.
24178 (mips_get_compress_on_name): New function.
24179 (mips_get_compress_off_name): New function.
24180 (mips_insert_attributes): Support multiple compression types.
24181 (mips_merge_decl_attributes): Likewise.
24182 (umips_12bit_offset_address_p): New function.
24183 (mips_start_function_definition): Emit .set micromips directive.
24184 (mips_call_may_need_jalx_p): New function.
24185 (mips_function_ok_for_sibcall): Add microMIPS support.
24186 (mips_print_operand_punctuation): Support short delay slots and
24188 (umips_swm_mask, umips_swm_encoding): New.
24189 (umips_build_save_restore): New function.
24190 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24191 (was_mips16_p): Remove.
24192 (old_compression_mode): New.
24193 (mips_set_compression_mode): New function.
24194 (mips_set_current_function): Add microMIPS support.
24195 (mips_option_override): Likewise.
24196 (umips_save_restore_pattern_p): New function.
24197 (umips_output_save_restore): New function.
24198 (umips_load_store_pair_p_1): New function.
24199 (umips_load_store_pair_p): New function.
24200 (umips_output_load_store_pair_1): New function.
24201 (umips_output_load_store_pair): New function.
24202 (umips_movep_target_p) New function.
24203 (mips_prepare_pch_save): Add microMIPS support.
24204 * config/mips/mips.h (TARGET_COMPRESSION): New.
24205 (TARGET_CPU_CPP_BUILTINS): Update macro
24206 to use new compression flags and to support microMIPS.
24207 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24208 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24209 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24210 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24211 (ASM_SPEC): Support mmicromips and mno-micromips.
24212 (M16STORE_REG_P): New macro.
24213 (MIPS_CALL): Support TARGET_MICROMIPS.
24214 (MICROMIPS_J): New macro.
24215 (mips_base_mips16): Rename this...
24216 (mips_base_compression_flags): ...to this.
24217 (UMIPS_12BIT_OFFSET_P): New macro.
24218 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24219 (MULTILIB_DIRNAMES): Likewise.
24221 2013-02-24 Jakub Jelinek <jakub@redhat.com>
24224 * target-globals.c (save_target_globals): For init_reg_sets and
24225 target_reinit remporarily set this_fn_optabs to this_target_optabs.
24227 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
24229 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
24230 * config/aarch64/t-aarch64
24231 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
24233 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
24235 PR inline-asm/56148
24236 * lra-constraints.c (process_alt_operands): Reload operand
24237 conflicting with earlier clobber only if no more other conflicting
24240 2013-02-22 Jakub Jelinek <jakub@redhat.com>
24243 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
24244 if not linking a shared library.
24246 2013-02-22 Seth LaForge <sethml@google.com>
24248 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
24250 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
24252 * config/arm/arm.md (split for extendsidi): Update condition.
24253 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
24254 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
24255 (qhs_zextenddi_cstr): Likewise.
24257 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24259 PR middle-end/56420
24260 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
24261 avoid signed wrapping.
24262 (expand_mult): Handle properly multiplication by
24263 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
24264 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
24265 in the compiler if coeff is HOST_WIDE_INT_MIN.
24266 (expand_divmod): Don't make ext_op1 static, change it's type to
24267 uhwi. Avoid undefined behavior in -INTVAL (op1).
24269 PR rtl-optimization/50339
24270 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
24272 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
24273 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
24274 into splitting_ashiftrt field.
24275 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
24277 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
24280 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
24282 PR middle-end/56108
24283 * trans-mem.c (execute_tm_mark): Do not expand transactions that
24284 are sure to go irrevocable.
24286 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
24288 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
24289 scalars are valid operands.
24291 2013-02-21 Martin Jambor <mjambor@suse.cz>
24293 PR tree-optimization/56310
24294 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
24295 only matching indices and non-negative final offsets.
24296 (intersect_aggregates_with_edge): Pass src_idx to
24297 agg_replacements_to_vector. Pass src_idx insstead of index to
24298 intersect_with_agg_replacements.
24300 2013-02-21 Martin Jambor <mjambor@suse.cz>
24302 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
24303 instead of hard-wired defaults.
24305 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
24307 * doc/invoke.texi (MIPS Options): Update documentation of the
24308 floating-point multiply-accumulate instruction restrictions.
24310 2013-02-21 Kostya Serebryany <kcc@google.com>
24312 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
24313 asan_shadow_offset on x86_64 linux.
24315 2013-02-21 Richard Biener <rguenther@suse.de>
24317 PR tree-optimization/56415
24319 2013-02-11 Richard Biener <rguenther@suse.de>
24321 PR tree-optimization/56273
24322 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24325 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24328 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
24331 PR inline-asm/56405
24332 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
24333 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
24335 2013-02-20 Jan Hubicka <jh@suse.cz>
24337 PR tree-optimization/56265
24338 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
24339 when target is referenced for first time.
24341 2013-02-20 Richard Biener <rguenther@suse.de>
24343 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
24344 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
24345 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
24346 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
24347 not return anything.
24348 (rename_ssa_copies): Do not remove unused locals.
24349 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
24350 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
24351 * passes.c (execute_function_todo): Do not schedule unused locals
24352 removal if cleanup_tree_cfg did something.
24353 * tree-ssa-live.c (remove_unused_locals): Dump statistics
24354 about the number of removed locals.
24356 2013-02-20 Richard Biener <rguenther@suse.de>
24358 PR tree-optimization/56398
24359 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
24361 2013-02-20 Martin Jambor <mjambor@suse.cz>
24363 PR tree-optimization/55334
24364 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
24365 restricted pointers to arrays.
24367 2013-02-20 Richard Biener <rguenther@suse.de>
24368 Jakub Jelinek <jakub@redhat.com>
24370 PR tree-optimization/56396
24371 * tree-ssa-ccp.c (n_const_val): New static variable.
24372 (get_value): Return NULL for SSA names we don't have a lattice
24374 (ccp_initialize): Initialize n_const_val.
24375 * tree-ssa-copy.c (n_copy_of): New static variable.
24376 (init_copy_prop): Initialize n_copy_of.
24377 (get_value): Return NULL_TREE for SSA names we don't have a
24380 2013-02-20 Martin Jambor <mjambor@suse.cz>
24382 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
24384 2013-02-20 Richard Biener <rguenther@suse.de>
24386 * genpreds.c (write_lookup_constraint): Do not compare first
24387 letter of the constraint again.
24389 2013-02-20 Richard Biener <rguenther@suse.de>
24391 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
24393 (get_use_iv_cost): Terminate hashtable walk when coming across
24396 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
24398 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
24399 reassociation for avx2 targets.
24401 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24403 * config/microblaze/microblaze.c: microblaze_has_clz = 0
24404 Add version check for v8.10.a to enable microblaze_has_clz
24405 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
24406 version and TARGET_PATTERN_COMPARE check
24407 * config/microblaze/microblaze.md: New clzsi2 instruction
24409 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24411 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
24412 function before branching.
24414 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24416 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
24418 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
24420 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24422 PR middle-end/55889
24423 * sel-sched.c: Include ira.h.
24424 (implicit_clobber_conflict_p): New function.
24425 (moveup_expr): Use it.
24426 * Makefile.in (sel-sched.o): Depend on ira.h.
24428 2013-02-19 Richard Biener <rguenther@suse.de>
24430 PR tree-optimization/56384
24431 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
24432 (vn_hash_type): Split out from ...
24433 (vn_hash_constant_with_type): ... here.
24434 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
24435 (vn_phi_eq): Compare types from vn_phi_s structure.
24436 (vn_phi_lookup): Populate vn_phi_s type.
24437 (vn_phi_insert): Likewise.
24439 2013-02-19 Jakub Jelinek <jakub@redhat.com>
24441 PR tree-optimization/56350
24442 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
24443 if haven't found reduction or nested cycle operand, rather than
24444 asserting we must find it.
24446 PR tree-optimization/56381
24447 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
24450 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
24451 Jakub Jelinek <jakub@redhat.com>
24454 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
24455 (swap_optab_enable): Same.
24456 (init_all_optabs): Use argument instead of global.
24457 * tree.h (struct tree_optimization_option): New field target_optabs.
24458 * expr.h (init_all_optabs): Add argument to prototype.
24459 (TREE_OPTIMIZATION_OPTABS): New.
24460 (save_optabs_if_changed): Protoize.
24461 * optabs.h: Declare this_fn_optabs.
24462 * optabs.c (save_optabs_if_changed): New.
24463 Declare this_fn_optabs.
24464 (init_optabs): Add argument to init_all_optabs() call.
24465 * function.c (invoke_set_current_function_hook): Handle per
24467 * function.h (struct function): New field optabs.
24468 * config/mips/mips.c (mips_set_mips16_mode): Handle when
24469 optimization_current_node has changed.
24470 * target-globals.h (save_target_globals_default_opts): Protoize.
24471 * target-globals.c (save_target_globals_default_opts): New.
24473 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24476 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
24477 registers %fr12 and %fr12R as call used.
24480 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
24481 and HImode, require all displacements to be an integer multiple of
24483 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
24484 only allow QImode and HImode when reload is in progress and strict is
24485 true. Likewise for symbolic addresses. Use base14_operand to check
24486 displacements in REG+BASE addresses.
24488 2013-02-18 Richard Biener <rguenther@suse.de>
24490 PR tree-optimization/56366
24491 * tree-vect-loop.c (get_initial_def_for_induction): Properly
24492 handle sign-conversion of outer-loop initial induction value.
24494 2013-02-18 Richard Biener <rguenther@suse.de>
24496 PR middle-end/56349
24497 * cfghooks.c (merge_blocks): If we merge a latch into another
24498 block adjust references to it.
24499 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
24500 (verify_loop_structure): Verify that a recorded latch is in fact
24503 2013-02-18 Richard Biener <rguenther@suse.de>
24505 PR tree-optimization/56321
24506 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
24507 order SSA name release and virtual operand unlinking.
24509 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24511 * config/microblaze/microblaze.md (save_stack_block): Define.
24512 (restore_stack_block): Likewise.
24514 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24516 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
24517 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
24518 * config/microblaze/microblaze.c (microblaze_option_override):
24519 Bail out early for PIC modes when target does not support PIC.
24521 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24523 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
24524 Replace with a microblaze version.
24525 (microblaze_trampoline_init): Adapt for microblaze.
24526 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
24529 2013-02-16 Jakub Jelinek <jakub@redhat.com>
24530 Dodji Seketeli <dodji@redhat.com>
24533 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
24534 (instrument_mem_region_access): Do not forget to always put
24535 instrumentation of the of 'base' and 'base + len' in a "if (len !=
24536 0) statement, even for cases where either 'base' or 'base + len'
24537 are not instrumented -- because they have been previously
24538 instrumented. Simplify the logic by putting all the statements
24539 instrument 'base + len' inside a sequence, and then insert that
24540 sequence right before the current insertion point. Then, to
24541 instrument 'base + len', just get an iterator on that statement.
24542 And do not forget to update the pointer to iterator the function
24543 received as argument.
24545 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
24547 PR rtl-optimization/56348
24548 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
24550 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
24552 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
24553 (clean_graph_dump_file): Pass base to start_graph_dump.
24555 2013-02-14 Richard Henderson <rth@redhat.com>
24558 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
24560 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
24562 * collect2-aix.h: Define F_LOADONLY.
24564 2013-02-14 Richard Biener <rguenther@suse.de>
24567 * varasm.c (output_constant_def_1): Get the decl representing
24568 the constant as argument.
24569 (output_constant_def): Wrap output_constant_def_1.
24570 (make_decl_rtl): Use output_constant_def_1 with the decl
24571 representing the constant.
24572 (build_constant_desc): Optionally re-use a decl already
24573 representing the constant.
24574 (tree_output_constant_def): Adjust.
24576 2013-02-14 Dodji Seketeli <dodji@redhat.com>
24579 * asan.c (instrument_builtin_call): Really put the length of the
24580 second source argument into src1_len.
24582 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24584 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
24585 argument. If it is false, don't create edge from then_bb to
24587 (insert_if_then_before_iter): Pass true to it.
24588 (build_check_stmt): Pass false to it.
24589 (transform_statements): Flush hash table only on extended basic
24590 block boundaries, rather than at the beginning of every bb.
24591 Don't flush hash table on nonfreeing_call_p calls.
24592 * tree-flow.h (nonfreeing_call_p): New prototype.
24593 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
24595 2013-02-13 David S. Miller <davem@davemloft.net>
24597 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
24599 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
24602 * ira.c (max_regno_before_ira): Move from ...
24604 (fix_reg_equiv_init): Use max_regno_before_ira instead of
24607 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24609 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
24611 2013-02-13 Richard Biener <rguenther@suse.de>
24614 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
24615 globals in MEM_REFs.
24617 2013-02-13 Richard Biener <rguenther@suse.de>
24619 * loop-init.c (loop_optimizer_init): Clear loop state when
24620 re-initializing preserved loops.
24621 * loop-unswitch.c (unswitch_single_loop): Return whether
24622 we unswitched the loop. Do not verify loop state here.
24623 (unswitch_loops): When we unswitched a loop discover new loops.
24625 2013-02-13 Kostya Serebryany <kcc@google.com>
24627 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
24629 * sanitizer.def: Rename __asan_init to __asan_init_v1.
24631 2013-02-12 Dodji Seketeli <dodji@redhat.com>
24633 Avoid instrumenting duplicated memory access in the same basic block
24634 * Makefile.in (asan.o): Add new dependency on hash-table.h
24635 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
24636 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
24637 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
24638 (free_mem_ref_resources, has_mem_ref_been_instrumented)
24639 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
24640 (get_mem_ref_of_assignment): New functions.
24641 (get_mem_refs_of_builtin_call): Extract from
24642 instrument_builtin_call and tweak a little bit to make it fit with
24644 (instrument_builtin_call): Use the new
24645 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
24646 of is_gimple_builtin_call.
24647 (instrument_derefs, instrument_mem_region_access): Insert the
24648 instrumented memory reference into the hash table.
24649 (maybe_instrument_assignment): Renamed instrument_assignment into
24650 this, and change it to advance the iterator when instrumentation
24651 actually happened and return true in that case. This makes it
24652 homogeneous with maybe_instrument_assignment, and thus give a
24653 chance to callers to be more 'regular'.
24654 (transform_statements): Clear the memory reference hash table
24655 whenever we enter a new BB, when we cross a function call, or when
24656 we are done transforming statements. Use
24657 maybe_instrument_assignment instead of instrumentation. No more
24658 need to special case maybe_instrument_assignment and advance the
24659 iterator after calling it; it's now handled just like
24660 maybe_instrument_call. Update comment.
24662 2013-02-13 Richard Biener <rguenther@suse.de>
24664 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
24665 Fix loop discovery code.
24667 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
24669 PR inline-asm/56148
24670 * lra-constraints.c (process_alt_operands): Match early clobber
24671 operand with itself. Check conflicts with earlyclobber only if
24672 the operand is not reloaded. Prefer to reload conflicting operand
24673 if earlyclobber and matching operands are the same.
24675 2013-02-12 Richard Biener <rguenther@suse.de>
24678 * lto-streamer-out.c (write_symbol): Do not output symbols
24679 for hard register variables.
24681 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
24684 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
24685 (umulsidi3_insn, mulsidi3_insn): New insns.
24687 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
24689 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
24690 (struct tune_params): Add vec_costs field.
24691 * config/arm/arm.c (arm_builtin_vectorization_cost)
24692 (arm_add_stmt_cost): New functions.
24693 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
24694 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
24695 (arm_default_vec_cost): New struct of type cpu_vec_costs.
24696 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
24697 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
24698 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
24699 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
24701 2013-02-12 Richard Biener <rguenther@suse.de>
24704 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
24705 decls again if possible.
24707 2013-02-12 Richard Biener <rguenther@suse.de>
24709 PR middle-end/56288
24710 * tree-ssa.c (verify_ssa_name): Fix check, move
24711 SSA_NAME_IN_FREE_LIST check up.
24713 2013-02-12 Jakub Jelinek <jakub@redhat.com>
24714 Steven Bosscher <steven@gcc.gnu.org>
24716 PR rtl-optimization/56151
24717 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
24718 equal to op0 or op1, and last_insn pattern is CODE operation
24719 with MEM dest and one of the operands matches that MEM.
24721 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
24723 * doc/extend.texi: Document Function Multiversioning and "default"
24724 parameter string to target attribute.
24725 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
24726 target attribute parameter is "default".
24727 (ix86_compare_version_priority): Remove checks for target attribute.
24728 (ix86_mangle_function_version_assembler_name): Change error to sorry.
24729 Remove check for target attribute equal to NULL. Add assert.
24730 (ix86_generate_version_dispatcher_body): Change error to sorry.
24732 2013-02-11 Iain Sandoe <iain@codesourcery.com>
24733 Jack Howarth <howarth@bromo.med.uc.edu>
24734 Patrick Marlier <patrick.marlier@gmail.com>
24737 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
24738 define ENDFILE_SPEC as TM_DESTRUCTOR.
24739 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
24741 2013-02-11 Alexander Potapenko <glider@google.com>
24742 Jack Howarth <howarth@bromo.med.uc.edu>
24743 Jakub Jelinek <jakub@redhat.com>
24746 * config/darwin.c (cdtor_record): Rename ctor_record.
24747 (sort_cdtor_records): Rename sort_ctor_records.
24748 (finalize_dtors): New routine to sort destructors by
24749 priority before use in assemble_integer.
24750 (machopic_asm_out_destructor): Use finalize_dtors if needed.
24752 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
24754 PR rtl-optimization/56275
24755 * simplify-rtx.c (avoid_constant_pool_reference): Check that
24756 offset is non-negative and less than cmode size before
24757 calling simplify_subreg.
24759 2013-02-11 Richard Biener <rguenther@suse.de>
24761 PR tree-optimization/56264
24762 * cfgloop.h (fix_loop_structure): Adjust prototype.
24763 * loop-init.c (fix_loop_structure): Return the number of
24764 newly discovered loops.
24765 * tree-cfgcleanup.c (repair_loop_structures): When new loops
24766 are discovered, do a full loop-closed SSA rewrite.
24768 2013-02-11 Richard Biener <rguenther@suse.de>
24770 PR tree-optimization/56273
24771 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24773 (check_array_ref): Fix missing newline in dumps.
24774 (search_for_addr_array): Likewise.
24776 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
24778 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
24780 2013-02-09 Jakub Jelinek <jakub@redhat.com>
24783 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
24785 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
24787 PR rtl-optimization/56246
24788 * lra-constraints.c (simplify_operand_subreg): Try to reuse
24790 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
24791 constraints are satisfied.
24793 2013-02-08 Jeff Law <law@redhat.com>
24796 * emit-rtl.c (reg_is_parm_p): New function.
24797 * regs.h (reg_is_parm_p): New prototype.
24798 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
24799 callee-clobbered registers.
24801 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
24804 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
24805 If there is no implicit builtin declaration, just return NULL.
24807 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24809 * config/i386/sse.md (FMAMODEM): New mode iterator.
24810 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
24811 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
24813 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24815 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
24816 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
24817 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
24819 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24821 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
24822 (microblaze*-*-elf): Likewise.
24823 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
24825 * config/microblaze/microblaze-c.c: Add builtin defines for
24826 _LITTLE_ENDIAN and _BIG_ENDIAN.
24827 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
24828 add to TARGET_DEFAULT flags.
24829 Expand ASM_SPEC and LINK_SPEC.
24830 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
24831 * config/microblaze/microblaze.md: Update extendsidi2 and
24832 movdi_internal instructions to use low-order / high-order reg
24834 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
24835 options and inversemask / mask of LITTLE_ENDIAN.
24836 * config/microblaze/t-microblaze: Expand multilib options to
24837 include mlittle-endian (le) and update exceptions patterns.
24839 2013-02-08 Jakub Jelinek <jakub@redhat.com>
24841 PR rtl-optimization/56195
24842 * lra-constraints.c (get_reload_reg): Don't reuse regs
24843 if they have smaller mode than requested, if they have
24844 wider mode than requested, try to return a SUBREG.
24846 PR tree-optimization/56250
24847 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
24848 if type is unsigned and code isn't MULT_EXPR.
24850 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
24852 PR tree-optimization/56064
24853 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
24854 bits according to mode.
24855 * fixed-value.h (fixed_from_double_int)
24856 (const_fixed_from_double_int): Adjust comments.
24858 2013-02-08 Richard Biener <rguenther@suse.de>
24861 * lto-streamer.h (struct data_in): Remove current_file, current_line
24862 and current_col members.
24863 * lto-streamer-out.c (lto_output_location): Stream changed bits
24864 en-block for efficiency.
24865 * lto-streamer-in.c (clear_line_info): Remove.
24866 (lto_input_location): Cache current file, line and column
24867 globally via local statics. Read changed bits en-block.
24868 (input_function): Do not call clear_line_info.
24869 (lto_read_body): Likewise.
24870 (lto_input_toplevel_asms): Likewise.
24872 2013-02-08 Michael Matz <matz@suse.de>
24874 PR tree-optimization/52448
24875 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
24876 (nt_call_phase): New static.
24877 (add_or_mark_expr): Only mark accesses with newer phase than any
24879 (nonfreeing_call_p): New.
24880 (nt_init_block): Update nt_call_phase, mark blocks as visited.
24881 (nt_fini_block): Keep blocks marked as visited.
24882 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
24884 2013-02-08 Richard Biener <rguenther@suse.de>
24886 * ira.c (ira): Free broken dominator information.
24888 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24890 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
24892 2013-02-08 Marek Polacek <polacek@redhat.com>
24894 * cfgloop.c (verify_loop_structure): Add more checking of headers.
24896 2013-02-08 Richard Biener <rguenther@suse.de>
24898 PR middle-end/56181
24899 * cfgloop.h (flow_loops_find): Adjust.
24900 (bb_loop_header_p): Declare.
24901 * cfgloop.c (bb_loop_header_p): New function split out from ...
24902 (flow_loops_find): ... here. Adjust function signature,
24903 support incremental loop structure update.
24904 (verify_loop_structure): Cleanup. Verify a loop is a loop.
24905 * cfgloopmanip.c (fix_loop_structure): Move ...
24906 * loop-init.c (fix_loop_structure): ... here.
24907 (apply_loop_flags): Split out from ...
24908 (loop_optimizer_init): ... here.
24909 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
24910 in incremental mode, only remove dead loops here.
24912 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
24915 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
24916 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
24917 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
24918 (*round<mode>3.libgcc): New insns for fixed-modes.
24919 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
24920 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
24921 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
24922 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
24923 implementations. Define to __builtin_avr_absFX,
24924 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
24925 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
24926 __builtin_avr_countlsFX, respectively.
24927 * config/avr/avr-c.c (target.h): Include it.
24928 (enum avr_builtin_id): New enum.
24929 (avr_resolve_overloaded_builtin): New static function.
24930 (avr_register_target_pragmas): Use it to set
24931 targetm.resolve_overloaded_builtin.
24932 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
24933 tree nodes used by DEF_BUILTIN.
24934 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
24935 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
24936 <AVR_BUILTIN_xxBITS>: Same.
24938 2013-02-08 Richard Biener <rguenther@suse.de>
24940 * cfgloop.c (verify_loop_structure): Properly handle
24941 a loop exiting to another loop header.
24942 * ira-int.h (ira_loops): Remove.
24943 * ira.c (ira_loops): Remove.
24944 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
24945 (do_reload): Use loop_optimizer_finalize.
24946 * ira-build.c (create_loop_tree_nodes): Use get_loops and
24947 number_of_loops to access the loop tree.
24948 (more_one_region_p): Likewise.
24949 (finish_loop_tree_nodes): Likewise.
24950 (rebuild_regno_allocno_maps): Likewise.
24951 (mark_loops_for_removal): Likewise.
24952 (mark_all_loops_for_removal): Likewise.
24953 (remove_unnecessary_regions): Likewise.
24954 (ira_build): Likewise.
24955 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
24957 2013-02-08 Richard Biener <rguenther@suse.de>
24959 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
24960 * ipa-pure-const.c (analyze_function): Avoid calling
24961 mark_irreducible_loops twice.
24962 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
24964 2013-02-07 David S. Miller <davem@davemloft.net>
24966 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
24968 * var-tracking.c (vt_add_function_parameter): Test the presence of
24969 HAVE_window_save properly and do not remap argument registers when
24970 we have a leaf function.
24972 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
24975 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
24977 * config/i386/i386.c (ix86_print_operand): Ditto.
24979 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24981 * lra-constraints.c (process_alt_operands): Fix recently added comment.
24983 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24985 PR rtl-optimization/56225
24986 * lra-constraints.c (process_alt_operands): Check that reload hard
24987 reg can hold value for strict_low_part.
24989 2013-02-07 Jakub Jelinek <jakub@redhat.com>
24992 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
24993 dwarf2out_end_function.
24994 (in_first_function_p, maybe_at_text_label_p,
24995 first_loclabel_num_not_at_text_label): New variables.
24996 (dwarf2out_var_location): In the first function find out
24997 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
24998 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
25001 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
25003 PR rtl-optimization/56178
25004 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
25005 SUBREG of a register. Tidy up related block of code.
25006 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
25007 note if the source is a register or a SUBREG of a register.
25009 2013-02-07 Jakub Jelinek <jakub@redhat.com>
25012 * config/rs6000/rs6000.md (ptrm): New mode attr.
25013 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
25014 call_value_indirect_aix<pttrsize>,
25015 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
25018 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25020 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
25021 if -bnortl. Convert to strcmp and strncmp.
25023 2013-02-07 Alan Modra <amodra@gmail.com>
25026 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
25027 addresses won't wrap when offsetting.
25028 (rs6000_secondary_reload): Provide secondary reloads needed for
25029 wrapping LO_SUM addresses.
25031 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
25033 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
25034 MACH, just __MACH__.
25036 2013-02-06 Richard Biener <rguenther@suse.de>
25038 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
25039 instead of calling fix_loop_structure.
25041 2013-02-06 Jakub Jelinek <jakub@redhat.com>
25043 PR middle-end/56217
25044 * omp-low.c (use_pointer_for_field): Return false if
25045 lower_send_shared_vars doesn't generate any copy-out code.
25047 2013-02-06 Tom de Vries <tom@codesourcery.com>
25049 PR rtl-optimization/56131
25050 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
25051 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
25052 of the label is NULL. Add comment.
25054 2013-02-05 Jakub Jelinek <jakub@redhat.com>
25056 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
25059 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
25060 (STATIC_LIBTSAN_LIBS): Likewise.
25061 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
25062 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
25063 is defined, don't add anything else beyond that.
25064 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
25065 (LINK_COMMAND_SPEC): Use them.
25067 PR tree-optimization/56205
25068 * tree-stdarg.c (check_all_va_list_escapes): Return true if
25069 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
25070 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
25072 2013-02-05 Richard Biener <rguenther@suse.de>
25074 PR tree-optimization/53342
25075 PR tree-optimization/53185
25076 * tree-vectorizer.h (vect_check_strided_load): Remove.
25077 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
25078 not disallow peeling for vectorized strided loads.
25079 (vect_check_strided_load): Make static and simplify.
25080 (vect_analyze_data_refs): Adjust.
25081 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
25082 correctly when vectorizing strided loads.
25084 2013-02-05 Richard Biener <rguenther@suse.de>
25086 * doc/install.texi: Refer to ISL, not PPL.
25088 2013-02-05 Jan Hubicka <jh@suse.cz>
25090 PR tree-optimization/55789
25091 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
25093 2013-02-05 Jan Hubicka <jh@suse.cz>
25095 PR tree-optimization/55789
25096 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
25097 the dead call anyway.
25099 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
25102 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
25104 2013-02-04 Alexander Potapenko <glider@google.com>
25105 Jack Howarth <howarth@bromo.med.uc.edu>
25106 Jakub Jelinek <jakub@redhat.com>
25109 * config/darwin.c (sort_ctor_records): Stabilized qsort
25110 on constructor priority by using original position.
25111 (finalize_ctors): New routine to sort constructors by
25112 priority before use in assemble_integer.
25113 (machopic_asm_out_constructor): Use finalize_ctors if needed.
25115 2013-02-04 Jakub Jelinek <jakub@redhat.com>
25118 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
25119 about visibility on artificial decls.
25120 * config/sol2.c (solaris_assemble_visibility): Likewise.
25122 2013-02-04 Kai Tietz <ktietz@redhat.com>
25125 * config/i386/i386.c (function_value_ms_64): Add additional valtype
25126 argument and improve checking of return-argument types for 16-byte
25128 (ix86_function_value_1): Add additional valtype argument on call
25129 of function_value_64.
25130 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
25131 handling infunction_value_64 function.
25133 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
25135 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
25137 2013-02-04 Richard Biener <rguenther@suse.de>
25139 PR tree-optimization/56188
25140 * tree-ssa-structalias.c (label_visit): Consider case with
25141 initially non-empty points-to set.
25142 (perform_var_substitution): Dump node mapping and clean up.
25144 2013-02-04 Richard Guenther <rguenther@suse.de>
25147 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
25148 node prevail as last resort.
25149 (lto_symtab_merge_decls): Remove guard on LTRANS here.
25150 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
25152 2013-02-04 Richard Biener <rguenther@suse.de>
25154 PR tree-optimization/56113
25155 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
25157 (equiv_class_lookup_or_add): ... this.
25158 (label_visit): Adjust and fix error in previous patch.
25159 (perform_var_substitution): Adjust.
25161 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
25163 * config/sh/divtab.c: Fix formatting and comments throughout the file.
25164 * config/sh/sh4-300.md: Likewise.
25165 * config/sh/sh4a.md: Likewise.
25166 * config/sh/constraints.md: Likewise.
25167 * config/sh/sh.md: Likewise.
25168 * config/sh/netbsd-elf.h: Likewise.
25169 * config/sh/predicates.md: Likewise.
25170 * config/sh/sh-protos.h: Likewise.
25171 * config/sh/ushmedia.h: Likewise.
25172 * config/sh/linux.h: Likewise.
25173 * config/sh/sh.c: Likewise.
25174 * config/sh/superh.h: Likewise.
25175 * config/sh/elf.h: Likewise.
25176 * config/sh/sh4.md: Likewise.
25177 * config/sh/sh.h: Likewise.
25179 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25181 * config/pa/constraints.md: Adjust unused letters. Change "T"
25182 constraint to match_test floating_point_store_memory_operand().
25183 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
25184 (base14_operand): New.
25185 (floating_point_store_memory_operand): New.
25186 (integer_store_memory_operand): Revise to use base14_operand and
25187 reg_plus_base_memory_operand.
25188 (move_dest_operand): Allow symbolic_memory_operands.
25189 (symbolic_memory_operand): Check for LO_SOM.
25190 (symbolic_operand): Change default case to break.
25191 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
25192 CONST_DOUBLE values to be reloaded by putting them into memory when
25193 the destination is a floating point register.
25194 (movdf): Remove code to handle CONST_DOUBLE.
25196 (reload_indf_r1): New.
25197 (reload_insf_r1): New.
25198 Consistently use "Q" and "T" constraints with integer and floating
25199 point move instructions, respectively.
25200 (movdi): Remove FAIL.
25201 Change predicate for source operand unamed DImode move from
25202 general_operand to move_src_operand.
25203 (umulsidi3): Change predicate for destination operand to
25205 Likewise for similar unamed patterns.
25206 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
25207 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
25208 (hppa_legitimize_address): Simplify mask calculation.
25209 (pa_emit_move_sequence): Revised handling of secondary reloads from
25210 REG+D addresses for floating point loads and stores. Directly handle
25211 loading CONST0_RTX (mode) to a floating point register.
25212 (pa_secondary_reload): Handle reloading DF and SFmode constant values
25213 to floating point registers. Don't restrict secondary reloads to
25214 floating point registers to integer modes. Revise some comments and
25216 (TARGET_LEGITIMATE_ADDRESS_P): Define.
25217 (pa_legitimate_address_p): New.
25218 (pa_legitimize_reload_address): New.
25219 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
25220 (STRICT_REG_OK_FOR_BASE_P): New.
25221 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
25222 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
25224 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
25225 Andrew Dixie <andrewd@gentrack.com>
25227 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
25230 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
25232 * expmed.c (extract_bit_field_1): Pass the full width of the
25233 structure to get_best_reg_extraction_insn.
25235 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25238 * configure.ac (use_cxa_atexit): Add AIX.
25239 * configure: Regenerate.
25241 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
25243 2013-02-01 Jakub Jelinek <jakub@redhat.com>
25246 * final.c (need_profile_function): New variable.
25247 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
25248 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
25249 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
25250 notes, targetm.asm_out.function_prologue doesn't emit anything,
25251 HAVE_prologue and profiler should be emitted before prologue,
25252 set need_profile_function instead of emitting it.
25253 (final_scan_insn): If need_profile_function, emit
25254 profile_function on the first NOTE_INSN_BASIC_BLOCK or
25255 NOTE_INSN_FUNCTION_BEG note.
25257 2013-02-01 Richard Henderson <rth@redhat.com>
25259 * config/rs6000/rs6000.md (smulditi3): New.
25262 * config/alpha/alpha.md (umulditi3): New.
25264 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25266 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
25267 (ASM_OUTPUT_ALIGNED_LOCAL): New.
25269 2013-02-01 Richard Biener <rguenther@suse.de>
25271 PR tree-optimization/56113
25272 * tree-ssa-structalias.c (label_visit): Reduce work for
25273 single-predecessor nodes.
25275 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
25277 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
25278 range isn't testing for zero.
25280 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
25282 PR middle-end/56113
25283 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
25285 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
25286 Nick Clifton <nickc@redhat.com>
25288 * config/v850/constraints.md (Q): Define as a memory constraint.
25289 * config/v850/predicates.md (label_ref_operand): New predicate.
25290 (e3v5_shift_operand): New predicate.
25291 (ior_operator): New predicate.
25292 * config/v850/t-v850: Add e3v5 multilib.
25293 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
25294 (v850_gen_movdi): Prototype.
25295 * config/v850/v850.c: Add support for e3v5 architecture.
25296 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
25298 (construct_save_jarl): Add e3v5 long JARL support.
25299 (v850_adjust_insn_length): New function. Adjust length of call
25300 insns when using e3v5 instructions.
25301 (v850_gen_movdi): New function: Generate instructions to move a
25303 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
25304 (CPP_SPEC): Define __v850e3v5__ as appropriate.
25305 (TARGET_USE_FPU): Enable for e3v5.
25306 (CONST_OK_FOR_W): New macro.
25307 (ADJUST_INSN_LENGTH): Define.
25308 * config/v850/v850.md (UNSPEC_LOOP): Define.
25309 (attr cpu): Add v850e3v5.
25310 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
25311 (movdi): New pattern.
25312 (movdi_internal): New pattern.
25313 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
25314 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
25315 (cstoresf4): Likewise.
25316 (cstoredf4): Likewise.
25317 (insv): New pattern.
25318 (rotlso3_a): New pattern.
25319 (rotlsi3_b): New pattern
25320 (rotlsi3_v850e3v5): New pattern.
25321 (doloop_begin): New pattern.
25322 (fix_loop_counter): New pattern.
25323 (doloop_end): New pattern.
25324 (branch_normal): Add e3v5 long branch support.
25325 (branch_invert): Likewise.
25326 (branch_z_normal): Likewise.
25327 (branch_z_invert): Likewise.
25328 (branch_nz_normal): Likewise.
25329 (branch_nz_invert): Likewise.
25330 (call_internal_short): Add e3v5 register-indirect JARL support.
25331 (call_internal_long): Likewise.
25332 (call_value_internal_short): Likewise.
25333 (call_value_internal_long): Likewise.
25334 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
25335 (mloop): New option.
25336 * config.gcc: Add support for configuring v840e3v5 target.
25337 * doc/invoke.texi: Document new v850 specific command line options.
25339 2013-01-31 Paul Koning <ni1d@arrl.net>
25343 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
25344 children if parent is a class.
25345 (prune_unused_types_prune): Don't add DW_AT_declaration.
25347 2013-01-31 Richard Biener <rguenther@suse.de>
25349 PR tree-optimization/56157
25350 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
25351 match up operand with SLP child.
25353 2013-01-31 Jason Merrill <jason@redhat.com>
25356 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
25357 parameters the first time.
25358 (gen_scheduled_generic_parms_dies): Check completeness here.
25360 2013-01-31 Richard Biener <rguenther@suse.de>
25362 PR middle-end/53073
25363 * common.opt (faggressive-loop-optimizations): New flag,
25364 enabled by default.
25365 * doc/invoke.texi (faggressive-loop-optimizations): Document.
25366 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
25367 infer_loop_bounds_from_undefined by it.
25369 2013-01-31 Richard Biener <rguenther@suse.de>
25371 PR tree-optimization/56150
25372 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
25373 visit virtual operands.
25374 (find_uses_to_rename_bb): Likewise.
25376 2013-01-31 Richard Biener <rguenther@suse.de>
25378 PR tree-optimization/56150
25379 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
25380 mixed store non-store stmts.
25382 2013-01-30 Jakub Jelinek <jakub@redhat.com>
25385 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
25386 LIBASAN_EARLY_SPEC is defined.
25387 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
25388 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
25390 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
25393 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
25394 invalid args instead of ICEing on it.
25395 (ix86_valid_target_attribute_tree): Return error_mark_node if
25396 ix86_valid_target_attribute_inner_p failed.
25397 (ix86_valid_target_attribute_p): Return false only if
25398 ix86_valid_target_attribute_tree returned error_mark_node. Allow
25399 target("default") attribute.
25400 (sorted_attr_string): Change argument from const char * to tree,
25401 merge in all target attribute arguments rather than just one.
25402 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
25403 instead of free. Avoid using strcat.
25404 (ix86_mangle_function_version_assembler_name): Mangle
25405 target("default") as if no target attribute is present. Adjust
25406 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
25407 instead of xmalloc and XDELETEVEC instead of free.
25408 (ix86_function_versions): Don't return true if one of the decls
25409 doesn't have target attribute. If they don't and one of the decls
25410 is DECL_FUNCTION_VERSIONED, report an error. Adjust
25411 sorted_attr_string caller. Use XDELETEVEC instead of free.
25412 (ix86_supports_function_versions): Remove.
25413 (make_name): Fix up formatting.
25414 (make_dispatcher_decl): Remove resolver_name and its initialization.
25415 Avoid leaking memory.
25416 (is_function_default_version): Return true if there is
25417 target("default") attribute rather than no target attribute at all.
25418 (make_resolver_func): Avoid leaking memory.
25419 (ix86_generate_version_dispatcher_body): Likewise.
25420 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
25421 * target.def (supports_function_versions): Remove.
25422 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
25423 * doc/tm.texi: Regenerated.
25425 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
25427 PR rtl-optimization/56144
25428 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
25429 for values with side effects.
25431 2013-01-30 Richard Biener <rguenther@suse.de>
25433 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
25434 (sparseset_pop): Likewise.
25435 * cfganal.c (compute_idf): Likewise. Increase work-stack size
25436 to be able to use quick_push in the worker loop.
25438 2013-01-30 Marek Polacek <polacek@redhat.com>
25440 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
25442 2013-01-30 Richard Biener <rguenther@suse.de>
25445 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
25447 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
25449 PR tree-optimization/56064
25450 * fixed-value.c (fixed_from_double_int): New function.
25451 * fixed-value.h (fixed_from_double_int): New prototype.
25452 (const_fixed_from_double_int): New static inline function.
25453 * fold-const.c (native_interpret_fixed): New static function.
25454 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
25455 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
25456 (native_encode_fixed): New static function.
25457 (native_encode_expr) <FIXED_CST>: Use it.
25458 (native_interpret_int): Move double_int worker code to...
25459 * double-int.c (double_int::from_buffer): ...this new static method.
25460 * double-int.h (double_int::from_buffer): Prototype it.
25462 2013-01-30 Richard Biener <rguenther@suse.de>
25464 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
25465 New pointer-map and obstack.
25466 (init_alias_vars): Allocate pointer-map and obstack.
25467 (delete_points_to_sets): Free them.
25468 (find_what_var_points_to): Cache result.
25469 (find_what_p_points_to): Adjust for changed interface of
25470 find_what_var_points_to.
25471 (compute_points_to_sets): Likewise.
25472 (ipa_pta_execute): Likewise.
25474 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25476 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
25477 * configure: Regenerate.
25478 * config.in: Regenerate.
25479 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
25480 #nobits/#progbits if supported.
25482 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
25485 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
25486 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
25487 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
25489 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25491 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
25492 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
25494 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25496 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
25497 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
25499 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25501 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
25503 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
25504 * config/arm/cortex-a7.md: New bypasses using
25505 arm_mac_accumulator_is_result.
25507 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25509 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
25510 (cortex_a7_neon_mla): Likewise.
25511 (cortex_a7_fpfmad): New reservation.
25512 (cortex_a7_fpmacs): Use ffmas and update required units.
25513 (cortex_a7_fpmuld): Update required units and latency.
25514 (cortex_a7_fpmacd): Likewise.
25515 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
25516 (cortex_a7_neon). Likewise.
25517 (bypass) Update participating units.
25519 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25521 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
25522 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
25524 * config/arm/vfp11.md (vfp_farith): Use ffmas.
25525 (vfp_fmul): Use ffmad.
25526 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
25527 (cortex_r4_fmacd): Use ffmad.
25528 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
25529 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
25530 (cortex_a9_fmacd): Use ffmad.
25531 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
25532 (cortex_a8_vfp_macd): Use ffmad.
25533 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
25534 (cortex_a5_fpmacd): Use ffmad.
25535 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
25536 (cortex_a15_vfp_macd): Use ffmad.
25537 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
25539 2013-01-29 Jason Merrill <jason@redhat.com>
25542 * varasm.c (default_assemble_visibility): Don't warn about
25543 visibility on artificial decls.
25545 2013-01-29 Richard Biener <rguenther@suse.de>
25547 PR tree-optimization/56113
25548 * tree-ssa-structalias.c (equiv_class_lookup): Also return
25550 (label_visit): Free duplicate bitmaps and record the leader instead.
25551 (perform_var_substitution): Adjust.
25553 2013-01-29 Richard Biener <rguenther@suse.de>
25555 PR tree-optimization/55270
25556 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
25557 the CFG, schedule loops for fixup.
25559 2013-01-29 Nick Clifton <nickc@redhat.com>
25561 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
25564 2013-01-28 Leif Ekblad <leif@rdos.net>
25566 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
25567 * config/i386/i386.h (TARGET_RDOS): New macro.
25568 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
25569 * config/i386/i386.c (ix86_option_override_internal): For 64bit
25570 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
25571 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
25572 DEFAULT_LARGE_SECTION_THRESHOLD.
25573 * config/i386/i386.md (R14_REG, R15_REG): New constants.
25574 * config/i386/rdos.h: New file.
25575 * config/i386/rdos64.h: New file.
25577 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
25580 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
25583 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25585 PR rtl-optimization/56117
25586 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
25587 call cselib_lookup_from_insn on the MEM before calling
25588 add_insn_mem_dependence.
25590 2013-01-28 Richard Biener <rguenther@suse.de>
25592 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
25593 to a stmt that didn't have one.
25594 (copy_phis_for_bb): Likewise for PHI arguments.
25595 (copy_debug_stmt): Likewise for debug stmts.
25597 2013-01-28 Richard Biener <rguenther@suse.de>
25599 PR tree-optimization/56034
25600 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
25601 (partition_builtin_p): Adjust.
25602 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
25603 it is the last partition.
25604 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
25605 up the vertex for the definition.
25606 (classify_partition): Classify whether a partition is a
25607 PKIND_REDUCTION, thus has uses outside of the loop.
25608 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
25609 Merge all PKIND_REDUCTION partitions into the last partition.
25610 (tree_loop_distribution): Seed partitions from reductions as well.
25612 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25614 PR tree-optimization/56125
25615 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
25616 pow(x,c) into sqrt(x) * powi(x, n/2) or
25617 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
25618 optimizing for size.
25619 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
25620 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
25623 PR tree-optimization/56094
25624 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
25625 to UNKNOWN_LOCATION while gimplifying expr.
25627 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
25630 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
25631 operand 0 in movabs insn template for -masm=intel asm alternative.
25632 (*movabs<mode>_2): Ditto for operand 1.
25634 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
25637 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
25640 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25642 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
25645 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25647 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
25648 Undef to avoid warning.
25650 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25652 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
25653 * configure: Regenerate.
25655 2013-01-25 Jakub Jelinek <jakub@redhat.com>
25657 PR tree-optimization/56098
25658 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
25659 for stmts with volatile ops.
25660 (cond_store_replacement): Don't optimize if assign has volatile ops.
25661 (cond_if_else_store_replacement_1): Don't optimize if either
25662 then_assign or else_assign have volatile ops.
25663 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
25666 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25668 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
25670 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25672 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
25673 missing ':' in asm example.
25675 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
25677 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
25678 entries into lane and laneq entries.
25679 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
25680 Remove AdvSIMD scalar modes.
25681 (aarch64_sq<r>dmulh_laneq<mode>): New.
25682 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
25684 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
25685 builtin implementations to relfect changes in RTL in aarch64-simd.md.
25686 * config/aarch64/iterators.md (VCOND): New.
25689 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25692 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
25693 Add NULL LIBNAME argument to existing definitions.
25694 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
25695 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
25696 * config/avr/avr.c (DEF_BUILTIN): Same.
25697 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
25698 (avr_expand_builtin): Expand to a vanilla call if a libgcc
25699 implementation is available (DECL_ASSEMBLER_NAME is set).
25700 (avr_fold_absfx): New static function.
25701 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
25702 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
25703 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
25704 AVR_BUILTIN_ABSLLK.
25705 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
25706 (abshk, absk, abslk, absllk): Provide as static inline functions.
25708 2013-01-25 Marek Polacek <polacek@redhat.com>
25710 PR tree-optimization/56035
25711 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
25713 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
25715 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
25716 (*movtf_internal_rex64): Add (!o,C) alternative
25717 (*movxf_internal_rex64): Ditto.
25718 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
25720 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
25722 * doc/invoke.texi: fix typo.
25723 * doc/objc.texi: fix typo.
25725 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
25727 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
25728 for the first two alternatives.
25730 2013-01-24 Diego Novillo <dnovillo@google.com>
25732 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
25733 (ggc-zone.o): Remove.
25734 * configure.ac: Remove option --with-gc.
25735 * configure: Re-generate.
25736 * doc/install.texi: Remove documentation for --with-gc.
25737 * gengtype.c (write_enum_defn): Remove. Update all users.
25738 (write_Types_process_field): Remove generation of gt_e_* argument.
25739 (output_type_enum): Remove. Update all users.
25740 (write_enum_defn): Remove. Update all users.
25741 (enum alloc_zone): Remove. Update all users.
25742 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
25743 * ggc-common.c (ggc_splay_alloc): Remove first argument.
25744 Update all callers.
25745 (struct ptr_data): Remove field TYPE. Update all users.
25746 (gt_pch_note_object): Remove argument TYPE. Update all users.
25747 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
25749 * ggc-none.c (ggc_alloc_typed_stat): Remove.
25750 (struct alloc_zone): Remove.
25751 (ggc_internal_alloc_zone_stat): Remove.
25752 (ggc_internal_cleared_alloc_zone_stat): Remove.
25753 * ggc-page.c (ggc_alloc_typed_stat): Remove.
25754 (ggc_pch_count_object): Remove last argument. Update all users.
25755 (ggc_pch_alloc_object): Remove last argument. Update all users.
25756 (struct alloc_zone): Remove.
25757 * ggc-zone.c: Remove.
25758 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
25759 (struct alloc_zone): Remove.
25760 (ggc_alloc_typed_stat): Remove.
25761 (ggc_alloc_typed): Remove.
25762 (ggc_splay_alloc): Remove first argument.
25763 (rtl_zone): Remove. Update all users.
25764 (tree_zone): Remove. Update all users.
25765 (tree_id_zone): Remove. Update all users.
25766 (ggc_internal_zone_alloc_stat): Remove. Update all users.
25767 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
25768 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
25769 * tree-ssanames.c: Remove references to zone allocator in comments.
25771 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
25773 * config/avr/avr.c (avr_out_fract): Make register numbers that
25774 might be outside of source operand signed.
25776 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
25778 * config/i386/constraints.md (Yf): New constraint.
25779 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
25780 of f constraint to conditionaly disable x87 register preferences.
25781 (*movdf_internal): Ditto.
25782 (*movsf_internal): Ditto.
25784 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
25786 PR inline-asm/55934
25787 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
25788 that have operands with impossible constraints.
25789 Add a FIXME for a speed-up opportunity.
25790 * lra-constraints.c (process_alt_operands): Verify that a class
25791 selected from constraints on asms is valid for the operand mode.
25792 (curr_insn_transform): Remove incorrect comment.
25794 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
25796 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
25797 TOC operand is a valid symbol ref in the constant pool.
25799 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25801 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
25803 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
25806 * config/avr/stdfix.h: New file.
25807 * t-avr (stdfix-gcc.h): New rule to build it.
25808 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
25810 2013-01-23 Kostya Serebryany <kcc@google.com>
25812 * config/darwin.h: remove dependency on
25813 CoreFoundation (asan on Mac OS).
25815 2013-01-23 Jakub Jelinek <jakub@redhat.com>
25818 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
25819 instead of cmpdi_operand for first comparison operand.
25820 Don't assert that comparison operands aren't both constants.
25822 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25824 * doc/install.texi (Downloading the Source): Update references to
25825 downloading separate components.
25827 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25829 * doc/extend.texi (__int128): Improve grammar.
25831 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
25834 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
25835 alternative to (o,r).
25836 (*movdi_internal_rex64): Remove (!o,n) alternative.
25837 (DImode immediate->memory splitter): Remove.
25838 (DImode immediate->memory peephole2): Remove.
25839 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
25840 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
25841 alternative to (!o,*r).
25842 (*movtf_internal_sse): New pattern.
25843 (*movxf_internal_rex64): New pattern.
25844 (*movxf_internal): Disable for TARGET_64BIT.
25845 (*movdf_internal_rex64): Remove (!o,F) alternative.
25847 2013-01-22 Jakub Jelinek <jakub@redhat.com>
25849 PR middle-end/56074
25850 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
25851 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
25852 * tree-vect-loop-manip.c (find_loop_location): Also ignore
25853 stmt locations where LOCATION_LOCUS of the stmt location is
25854 UNKNOWN_LOCATION or BUILTINS_LOCATION.
25857 * config/i386/i386.md (UNSPEC_STOS): New.
25858 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
25859 *strsetqi_1): Add UNSPEC_STOS.
25861 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
25864 * doc/invoke.texi: Remove left over -Wsynth example.
25866 2013-01-21 Jakub Jelinek <jakub@redhat.com>
25868 PR tree-optimization/56051
25869 * fold-const.c (fold_binary_loc): Don't fold
25870 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
25871 a narrowing conversion, or widening conversion from signed
25874 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
25876 PR rtl-optimization/56023
25877 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
25878 dependent on debug instruction.
25880 2013-01-21 Martin Jambor <mjambor@suse.cz>
25882 PR middle-end/56022
25883 * function.c (allocate_struct_function): Call
25884 invoke_set_current_function_hook earlier.
25886 2013-01-21 Jakub Jelinek <jakub@redhat.com>
25888 * reload1.c (init_reload): Only initialize reload_obstack
25889 during the first call.
25891 2013-01-21 Marek Polacek <polacek@redhat.com>
25893 * cfgloop.c (verify_loop_structure): Fix up grammar.
25895 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
25897 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
25898 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
25900 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25903 * config/arm/marvell-pj4.md: Update copyright year.
25904 Fix up use of alu to alu_reg and simple_alu_imm.
25906 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
25908 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
25910 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
25913 * lra-constraints.c (curr_insn_transform): Don't reuse original
25914 insn for secondary memory move when memory mode should be different.
25916 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25918 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
25919 atomic_storedi_1): New patterns.
25921 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
25923 btver2 pipeline descriptions.
25924 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
25926 * config/i386/i386.md (btver2_decode): New type attributes.
25927 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
25929 * config/i386/btver2.md: New file describing btver2 pipelines.
25931 2013-01-19 Andrew Pinski <apinski@cavium.com>
25933 PR tree-optimization/52631
25934 * tree-ssa-sccvn (visit_use): Before looking up the original
25935 statement, try looking up the simplified expression.
25937 2013-01-19 Anthony Green <green@moxielogic.com>
25939 * config/moxie/moxie.c (moxie_expand_prologue): Set
25940 current_function_static_stack_size.
25942 2013-01-18 Jakub Jelinek <jakub@redhat.com>
25944 PR tree-optimization/56029
25945 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
25946 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
25948 2013-01-18 Sharad Singhai <singhai@google.com>
25950 PR tree-optimization/55995
25951 * dumpfile.c (dump_loc): Print location only if available.
25952 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
25954 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
25957 * lra-constraints.c (curr_insn_transform): Reuse original insn for
25958 secondary memory move.
25959 (inherit_reload_reg): Use rclass instead of cl for
25960 check_secondary_memory_needed_p.
25962 2013-01-18 Jakub Jelinek <jakub@redhat.com>
25964 PR middle-end/56015
25965 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
25966 the case where writing real complex part of target modifies op1.
25968 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
25970 * config/aarch64/aarch64-simd.md
25971 (aarch64_vcond_internal<mode>): Handle unordered cases.
25972 * config/aarch64/iterators.md (v_cmp_result): New.
25974 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
25975 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25977 * config/arm/marvell-pj4.md: New file.
25978 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
25979 * config/arm/arm.md (generic_sched): Add marvell_pj4.
25980 (generic_vfp): Likewise.
25981 * config/arm/arm-cores.def: Add marvell-pj4.
25982 * config/arm/arm-tune.md: Regenerate.
25983 * config/arm/arm-tables.opt: Regenerate.
25984 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
25985 * doc/invoke.texi: Document marvell-pj4.
25987 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
25989 * config/aarch64/arm_neon.h: Map scalar types to standard types.
25991 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
25996 * var-tracking.c (negative_power_of_two_p): New.
25997 (global_get_addr_cache, local_get_addr_cache): New.
25998 (get_addr_from_global_cache, get_addr_from_local_cache): New.
25999 (vt_canonicalize_addr): Rewrite using the above. Adjust the
26001 (vt_stack_offset_p): Remove.
26002 (vt_canon_true_dep): Always canonicalize loc's address.
26003 (clobber_overlapping_mems): Make sure we have a MEM.
26004 (local_get_addr_clear_given_value): New.
26005 (val_reset): Clear local cached entries.
26006 (compute_bb_dataflow): Create and release the local cache.
26007 Disable duplicate MEMs clobbering.
26008 (emit_notes_in_bb): Clobber MEMs likewise.
26009 (vt_emit_notes): Create and release the local cache.
26010 (vt_initialize, vt_finalize): Create and release the global
26011 cache, respectively.
26012 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
26014 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26016 PR libmudflap/53359
26017 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
26018 not found in the symtab.
26020 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26023 PR rtl-optimization/55547
26024 PR rtl-optimization/53827
26027 * alias.c (offset_overlap_p): New, factored out of...
26028 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
26029 the conservative special case for symbolic constants. Don't
26030 adjust zero sizes on alignment.
26032 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
26034 PR rtl-optimization/52573
26035 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
26036 REG_UNUSED for the same register.
26038 2013-01-17 Richard Biener <rguenther@suse.de>
26039 Marek Polacek <polacek@redhat.com>
26041 PR rtl-optimization/55833
26042 * loop-unswitch.c (unswitch_loops): Move loop verification...
26043 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
26044 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
26045 Set it to true when we're removing a loop from hierarchy tree in
26046 an irreducible region.
26047 (fix_bb_placements): Adjust caller.
26048 (fix_loop_placements): Likewise.
26050 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
26052 * config/avr/builtins.def (DEF_BUILTIN): Factor out
26053 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
26054 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
26055 Remove ID. Adjust comments.
26056 * config/avr/avr-c.c (avr_builtin_name): Remove.
26057 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
26058 * config/avr/avr.c (avr_tolower): New static function.
26059 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
26060 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
26061 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
26064 2013-01-17 Jan Hubicka <jh@suse.cz>
26066 PR tree-optimization/55273
26067 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
26069 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
26072 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
26073 store through atomic_store<mode>_1.
26074 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
26076 2013-01-17 Martin Jambor <mjambor@suse.cz>
26078 PR tree-optimizations/55264
26079 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
26080 for virtual methods.
26081 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
26082 virtual methods before inlining is over.
26083 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
26085 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
26088 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
26090 PR rtl-optimization/56005
26091 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
26092 pending reads for prefetch.
26094 2013-01-16 Ian Bolton <ian.bolton@arm.com>
26096 * config/aarch64/aarch64.md
26097 (*cstoresi_neg_uxtw): New pattern.
26098 (*cmovsi_insn_uxtw): New pattern.
26099 (*<optab>si3_uxtw): New pattern.
26100 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
26101 (*<optab>si3_insn_uxtw): New pattern.
26102 (*bswapsi2_uxtw): New pattern.
26104 2013-01-16 Richard Biener <rguenther@suse.de>
26106 * tree-inline.c (tree_function_versioning): Remove set but
26107 never used variable.
26109 2013-01-16 Richard Biener <rguenther@suse.de>
26111 PR tree-optimization/55964
26112 * tree-flow.h (rename_variables_in_loop): Remove.
26113 (rename_variables_in_bb): Likewise.
26114 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
26115 (copy_loop_before): Adjust and delete update-ssa status.
26116 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
26117 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
26118 (rename_variables_in_loop): Remove.
26119 (slpeel_update_phis_for_duplicate_loop): Likewise.
26120 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
26121 use available cfg machinery instead of duplicating it.
26122 Update PHI nodes and perform poor-mans SSA update here.
26123 (slpeel_tree_peel_loop_to_edge): Adjust.
26125 2013-01-16 Richard Biener <rguenther@suse.de>
26127 PR tree-optimization/54767
26128 PR tree-optimization/53465
26129 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
26130 (vrp_visit_phi_node): For PHI arguments coming via backedges
26131 drop all symbolical range information.
26132 (execute_vrp): Compute backedges.
26134 2013-01-16 Richard Biener <rguenther@suse.de>
26136 * doc/install.texi: Update CLooG and ISL requirements to
26139 2013-01-16 Christian Bruel <christian.bruel@st.com>
26142 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
26143 (broken_move): Handle UNSPECV_SP_SWITCH_B.
26144 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
26146 2013-01-16 DJ Delorie <dj@redhat.com>
26148 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
26149 (UNSPECV_SP_SWITCH_E): New.
26150 (sp_switch_1): Change to an unspec.
26151 (sp_switch_2): Change to an unspec. Don't use post-inc when we
26154 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
26156 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
26157 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
26158 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
26159 (expand_mem_thread_fence): Ditto.
26160 (expand_mem_signal_fence): Ditto.
26161 (expand_atomic_load): Ditto.
26162 (expand_atomic_store): Ditto.
26164 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
26166 PR rtl-optimization/55547
26167 PR rtl-optimization/53827
26170 * alias.c (memrefs_conflict_p): Set sizes to negative after
26173 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26176 * function.c (thread_prologue_and_epilogue_insns): Always
26177 add crtl->drap_reg to set_up_by_prologue.set, even if
26178 stack_realign_drap is false.
26180 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
26182 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
26183 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
26184 *call): Fix indention.
26186 2013-01-15 Tom de Vries <tom@codesourcery.com>
26189 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
26192 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
26194 PR rtl-optimization/55153
26195 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
26197 2013-01-15 Martin Jambor <mjambor@suse.cz>
26199 PR tree-optimization/55920
26200 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
26201 accesses as grp_to_be_debug_replaced.
26203 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26205 PR tree-optimization/55920
26206 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
26207 there is non-useless type conversion needed from debug rhs to lhs,
26208 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
26210 2013-01-15 Joseph Myers <joseph@codesourcery.com>
26211 Mikael Pettersson <mikpe@it.uu.se>
26214 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
26216 (ASM_OUTPUT_CASE_LABEL): Remove.
26217 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
26218 * final.c (shorten_branches): Update alignment of labels before
26219 jump tables if CASE_VECTOR_SHORTEN_MODE.
26221 2013-01-15 Richard Biener <rguenther@suse.de>
26224 * system.h: Do not include gmp.h for building host tools.
26226 2013-01-15 Richard Biener <rguenther@suse.de>
26228 PR middle-end/55882
26229 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
26230 account for bitpos when computing alignment.
26232 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
26234 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
26235 (ix86_target_macros_internal): Likewise.
26237 * config/i386/i386.c (m_CORE2I7): Removed.
26238 (m_CORE_HASWELL): New macro.
26239 (m_CORE_ALL): Likewise.
26240 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
26241 (initial_ix86_arch_features): Likewise.
26242 (processor_target_table): Initializations for Core avx2.
26243 (cpu_names): New names "core-avx2".
26244 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
26245 PROCESSOR_CORE_HASWELL.
26246 (ix86_issue_rate): New case.
26247 (ia32_multipass_dfa_lookahead): Likewise.
26248 (ix86_sched_init_global): Likewise.
26250 * config/i386/i386.h (TARGET_HASWELL): New macro.
26251 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
26252 (processor_type): New PROCESSOR_HASWELL.
26254 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26256 PR tree-optimization/55955
26257 * tree-vect-loop.c (vectorizable_reduction): Give up early on
26258 *SHIFT_EXPR and *ROTATE_EXPR codes.
26260 PR tree-optimization/48766
26261 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
26262 -ftrapv disable -fwrapv.
26264 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26267 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
26268 etc. to 1 and not to __flash.
26269 Use LL suffix for __INT24_MAX__ with -mint8.
26270 Use ULL suffix for __UINT24_MAX__ with -mint8.
26272 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26274 * config/avr/avr-arch.h
26275 (struct base_arch_s): Use typedef avr_arch_t instead.
26276 (struct arch_info_s): Use typedef avr_arch_info_t instead.
26277 (struct mcu_type_s): Use typedef avr_mcu_t instead.
26278 * config/avr/avr.c: Same.
26279 * config/avr/avr-devices.c: Same.
26280 * config/avr/driver-avr.c: Same.
26281 * config/avr/gen-avr-mmcu-texi.c: Same.
26282 * config/avr/avr-mcus.def: Adjust comment.
26284 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
26286 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
26287 * config/aarch64/iterators.md (VALLDI): New.
26289 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
26290 Andi Kleen <ak@linux.intel.com>
26293 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
26294 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
26297 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26299 * config/avr/avr-stdint.h: Remove trailing blanks.
26300 * config/avr/avr-log.h: Same.
26301 * config/avr/avr-arch.h: Same.
26302 * config/avr/avr-devices.c: Same.
26303 * config/avr/avr-dimode.md: Same.
26304 * config/avr/predicates.md: Same.
26305 * config/avr/avr-c.c: Same. And fix typo.
26307 * config/avr/avr-protos.h: Same. And:
26308 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
26309 (init_cumulative_args): Rename to avr_init_cumulative_args.
26310 (expand_prologue): Rename to avr_expand_prologue.
26311 (expand_epilogue): Rename to avr_expand_epilogue.
26312 (adjust_insn_length): Rename to avr_adjust_insn_length.
26313 (notice_update_cc): Rename to avr_notice_update_cc.
26314 (final_prescan_insn): Rename to avr_final_prescan_insn.
26315 * config/avr/avr.c: Same.
26316 * config/avr/avr.h: Same.
26317 * config/avr/avr.md: Remove trailing blanks.
26318 (prologue): Use avr_expand_prologue.
26319 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
26321 2013-01-14 Richard Biener <rguenther@suse.de>
26323 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
26324 verify_location, collect_subblocks): New functions.
26325 (verify_gimple_in_cfg): Verify that locations only reference
26326 BLOCKs in the functions BLOCK tree.
26328 2013-01-14 Richard Biener <rguenther@suse.de>
26330 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
26332 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
26334 (insert_out_of_ssa_copy_on_edge): Likewise.
26335 (rewrite_close_phi_out_of_ssa): Likewise.
26336 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
26338 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
26339 propagated constants.
26340 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
26343 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26345 * config/avr/avr-modes.def: Add GPL copyright notice.
26347 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
26349 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
26350 MEMMODEL_MASK to determine memory model.
26351 (atomic_store<mode>): Ditto from operands[2].
26352 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
26354 2013-01-13 Jakub Jelinek <jakub@redhat.com>
26357 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
26358 (fold_gimple_assign): Don't call unshare_expr here.
26359 (fold_ctor_reference): Call unshare_expr.
26361 2013-01-13 Terry Guo <terry.guo@arm.com>
26363 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
26364 * doc/fragments.texi: Document MULTILIB_REUSE.
26365 * gcc.c (multilib_reuse): New internal spec.
26366 (set_multilib_dir): Also search multilib from multilib_reuse.
26367 * genmultilib (tmpmultilib3): Refactor code.
26368 (tmpmultilib4): Ditto.
26369 (multilib_reuse): New multilib argument.
26371 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
26373 * Makefile.in: Update copyright.
26375 2013-01-12 Tom de Vries <tom@codesourcery.com>
26377 PR middle-end/55890
26378 * calls.c (expand_call): Check if arg_nr is valid.
26380 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
26382 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
26383 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
26384 documentation. Add missing '__' in front of
26385 __builtin_ia32_packssdw256.
26387 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26390 * config/s390/s390.c (s390_preferred_reload_class): Do not return
26391 NO_REGS for larl operands.
26392 (s390_reload_larl_operand): Use s390_load_address instead of
26395 2013-01-11 Richard Biener <rguenther@suse.de>
26397 * tree-cfg.c (verify_node_sharing_1): Split out from ...
26398 (verify_node_sharing): ... here.
26399 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
26401 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
26403 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
26404 Substitute TREECHECKING.
26405 * configure: Regenerate.
26406 * Makefile.in (TREECHECKING): New.
26408 2013-01-11 Richard Guenther <rguenther@suse.de>
26410 PR tree-optimization/44061
26411 * tree-vrp.c (extract_range_basic): Compute zero as
26412 value-range for __builtin_constant_p of function parameters.
26414 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
26416 Update copyright years.
26418 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
26420 PR rtl-optimization/55672
26421 * lra-eliminations.c (mark_not_eliminable): Permit addition with
26422 const to be eliminable.
26424 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
26426 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
26427 * configure: Regenerate.
26429 2013-01-10 Richard Biener <rguenther@suse.de>
26431 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
26433 2013-01-10 Richard Biener <rguenther@suse.de>
26436 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
26437 locations for virtual PHI arguments.
26438 (rewrite_update_phi_arguments): Likewise.
26440 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
26442 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
26445 2013-01-10 Jakub Jelinek <jakub@redhat.com>
26447 PR tree-optimization/55921
26448 * tree-complex.c (expand_complex_asm): New function.
26449 (expand_complex_operations_1): Call it for GIMPLE_ASM.
26451 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26454 * config/s390/s390.c (s390_symref_operand_p)
26455 (s390_loadrelative_operand_p): Merge the two functions.
26456 (s390_check_qrst_address, print_operand_address): Add parameters
26457 to s390_loadrelative_operand_p invokation.
26458 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
26459 (s390_reload_larl_operand, s390_secondary_reload): Use
26460 s390_loadrelative_operand_p instead of s390_symref_operand_p.
26461 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
26463 2013-01-09 Mike Stump <mikestump@comcast.net>
26465 * dse.c (record_store): Remove unnecessary assert.
26467 2013-01-09 Jan Hubicka <jh@suse.cz>
26469 PR tree-optimization/55569
26470 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
26471 * cfgloop.h (scale_loop_profile): Likewise.
26473 2013-01-09 Jan Hubicka <jh@suse.cz>
26476 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
26478 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
26480 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
26482 PR middle-end/55114
26483 * expr.h (maybe_emit_group_store): Declare.
26484 * expr.c (maybe_emit_group_store): New function.
26485 * builtins.c (expand_builtin_int_roundingfn): Call it.
26486 (expand_builtin_int_roundingfn_2): Likewise.
26488 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
26490 PR rtl-optimization/55829
26491 * lra-constraints.c (match_reload): Add code for absent output.
26492 (curr_insn_transform): Add code for reloads of matched inputs
26495 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
26497 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
26498 attribute of movddup insn to DF.
26499 (*vec_interleave_lowv2df): Ditto.
26500 (vec_dupv2df): Ditto.
26502 2013-01-09 Jan Hubicka <jh@suse.cz>
26504 PR tree-optimiation/55875
26505 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
26506 EVERY_ITERATION parameter.
26507 (number_of_iterations_exit): Check if exit is executed every iteration.
26508 (idx_infer_loop_bounds): Similarly here.
26509 (n_of_executions_at_most): Simplify
26510 to only test for cases where statement is dominated by the
26511 particular bound; handle correctly the "postdominance" test.
26512 (scev_probably_wraps_p): Use max loop iterations info
26513 as a global bound first.
26515 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
26516 Nick Clifton <nickc@redhat.com>
26518 * config/v850/v850.md (cbranchsf4): New pattern.
26519 (cstoresf4): New pattern.
26520 (cbranchdf4): New pattern.
26521 (cstoredf4): New pattern.
26522 (movsicc): Disallow floating point comparisons.
26523 (cmpsf_le_insn): Fix order of operators.
26524 (cmpsf_lt_insn): Likewise.
26525 (cmpsf_eq_insn): Likewise.
26526 (cmpdf_le_insn): Likewise.
26527 (cmpdf_lt_insn): Likewise.
26528 (cmpdf_eq_insn): Likewise.
26529 (cmpsf_ge_insn): Use LE comparison.
26530 (cmpdf_ge_insn): Likewise.
26531 (cmpsf_gt_insn): Use LT comparison.
26532 (cmpdf_gt_insn): Likewise.
26533 (cmpsf_ne_insn): Delete pattern.
26534 (cmpdf_ne_insn): Delete pattern.
26535 * config/v850/v850.c (v850_gen_float_compare): Use
26536 gen_cmpdf_eq_insn for NE comparison.
26537 (v850_float_z_comparison_operator)
26538 (v850_float_nz_comparison_operator): Move from here ...
26539 * config/v850/predicates.md: ... to here. Move GT and GE
26540 comparisons into v850_float_z_comparison_operator.
26541 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
26543 (v850_float_nz_comparison_operator): Likewise.
26545 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26547 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
26548 with calls to gen_insvsi/gen_insvdi.
26550 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26552 * config/i386/i386.c (initial_ix86_tune_features): Set up
26553 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
26555 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
26556 Jakub Jelinek <jakub@redhat.com>
26558 PR tree-optimization/48189
26559 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
26560 If nitercst is 0, don't predict the exit edge.
26562 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26564 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
26565 in asm_fprintf with reg_names.
26566 (aarch64_print_operand_address): Likewise.
26567 (aarch64_return_addr): Likewise.
26568 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
26570 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26572 * config/pa/pa.h (VAL_U6_BITS_P): Define.
26573 (INT_U6_BITS): Likewise.
26574 * config/pa/predicates.md (uint6_operand): New predicate.
26575 (shift5_operand, shift6_operand): Likewise.
26576 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
26578 (lshrdi3): Use shift6_operand.
26579 (shrpsi4, shrpdi4): New insn patterns.
26580 (extzv): Delete expander.
26581 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
26582 predicates in unamed zero extract patterns. Tighten common constraint.
26583 (extv): Delete expander.
26584 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
26585 predicates in unamed sign extract patterns. Tighten common constraint.
26586 (insv): Delete expander.
26587 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
26588 predicates in unamed insert patterns. Tighten common constraint.
26589 Change uint32_operand predicate to uint6_operand predicate in unamed
26590 DImode pattern to insert constant values of type 1...1xxxx.
26592 2013-01-04 Jan Hubicka <jh@suse.cz>
26594 PR tree-optimization/55823
26595 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
26598 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26599 Uros Bizjak <ubizjak@gmail.com>
26601 PR rtl-optimization/55845
26602 * df-problems.c (can_move_insns_across): Stop scanning at
26603 volatile_insn_p source instruction or give up if
26604 across_from .. across_to range contains any volatile_insn_p
26607 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26609 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
26610 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
26612 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
26613 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
26615 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26618 * asan.c (asan_clear_shadow): New function.
26619 (asan_emit_stack_protection): Use it.
26621 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26623 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
26624 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
26625 with tab instead of space.
26627 2013-01-08 Nick Clifton <nickc@redhat.com>
26629 * config/rl78/rl78.c (rl78_expand_prologue): Always select
26630 register bank 0 at the start of an interrupt handler.
26631 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
26634 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26636 * config/aarch64/aarch64-simd.md
26637 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
26638 (aarch64_simd_bsl): Likewise.
26639 (aarch64_vcond_internal<mode>): Likewise.
26640 (vcond<mode><mode>): Likewise.
26641 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
26642 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
26644 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26646 * config/aarch64/aarch64-builtins.c
26647 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
26649 2013-01-08 Martin Jambor <mjambor@suse.cz>
26652 * tree-sra.c (analyze_access_subtree): Return true also after
26653 potentially creating a debug-only replacement.
26655 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26657 PR middle-end/55890
26658 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
26660 PR tree-optimization/54120
26661 * tree-vrp.c (range_fits_type_p): Don't allow
26662 src_precision < precision from signed vr to unsigned_p
26663 if vr->min or vr->max is negative.
26664 (simplify_float_conversion_using_ranges): Test can_float_p
26665 against CODE_FOR_nothing.
26667 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26668 Richard Biener <rguenther@suse.de>
26670 PR middle-end/55851
26671 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
26672 types instead of just INTEGER_TYPE types.
26674 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
26676 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
26679 2013-01-07 Steve Ellcey <sellcey@mips.com>
26682 * config/mips/mips.opt: Change mad to mmad to match documentation.
26684 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26687 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
26688 .progmemx.data now.
26690 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26693 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
26694 (avr_addrspace_t): Add .section_name field.
26695 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
26697 (avr_addrspace): Same. Initialize .section_name. Remove last
26698 NULL entry. Put __memx into .progmemx.data.
26699 (progmem_section_prefix): Remove.
26700 (avr_asm_init_sections): No need to initialize progmem_section.
26701 (avr_asm_named_section): Use avr_addrspace[].section_name to get
26702 section name prefix.
26703 (avr_asm_select_section): Ditto. And use get_unnamed_section to
26704 retrieve the progmem section.
26705 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
26706 boundary to run over avr_addrspace[].
26707 (avr_register_target_pragmas): Ditto.
26709 2013-01-06 Jakub Jelinek <jakub@redhat.com>
26711 * varasm.c (output_constant_def_contents): For asan_protect_global
26712 protected strings, adjust DECL_ALIGN if needed, before testing for
26714 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
26715 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
26717 (output_object_block): For asan protected decls, emit asan padding
26718 after their contents.
26719 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
26720 (asan_finish_file): Test it here instead.
26722 2013-01-07 Nick Clifton <nickc@redhat.com>
26723 Matthias Klose <doko@debian.org>
26724 Doug Kwan <dougkwan@google.com>
26725 H.J. Lu <hongjiu.lu@intel.com>
26728 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
26730 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
26732 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
26734 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
26736 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
26738 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26741 * doc/install.texi (Cross-Compiler-Specific Options): Document
26744 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
26746 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
26747 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
26748 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
26749 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
26750 vqmovun_high_s64): Fix source operand number and update copyright.
26752 2013-01-07 Richard Biener <rguenther@suse.de>
26754 PR middle-end/55890
26755 * gimple.h (gimple_call_builtin_p): New overload.
26756 * gimple.c (validate_call): New function.
26757 (gimple_call_builtin_p): Likewise.
26758 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26759 Use gimple_call_builtin_p.
26760 (find_func_clobbers): Likewise.
26761 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
26762 (strlen_optimize_stmt): Likewise.
26764 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26766 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
26767 (vld1q_dup_*): Likewise.
26768 (vld1_*): Likewise.
26769 (vld1q_*): Likewise.
26770 (vld1_lane_*): Likewise.
26771 (vld1q_lane_*): Likewise.
26773 2013-01-07 Richard Biener <rguenther@suse.de>
26775 * lto-streamer.h (LTO_minor_version): Bump to 2.
26777 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26779 * config/aarch64/aarch64-protos.h
26780 (aarch64_const_double_zero_rtx_p): Rename to...
26781 (aarch64_float_const_zero_rtx_p): ...this.
26782 (aarch64_float_const_representable_p): New.
26783 (aarch64_output_simd_mov_immediate): Likewise.
26784 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
26785 move immediate case.
26786 * config/aarch64/aarch64.c
26787 (aarch64_const_double_zero_rtx_p): Rename to...
26788 (aarch64_float_const_zero_rtx_p): ...this.
26789 (aarch64_print_operand): Allow printing of new constants.
26790 (aarch64_valid_floating_const): New.
26791 (aarch64_legitimate_constant_p): Check for valid floating-point
26793 (aarch64_simd_valid_immediate): Likewise.
26794 (aarch64_vect_float_const_representable_p): New.
26795 (aarch64_float_const_representable_p): Likewise.
26796 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
26797 (aarch64_output_simd_mov_immediate): New.
26798 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
26799 (*movdf_aarch64): Likewise.
26800 * config/aarch64/constraints.md (Ufc): New.
26801 (Y): call aarch64_float_const_zero_rtx.
26802 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
26804 2013-01-07 Richard Biener <rguenther@suse.de>
26806 PR tree-optimization/55888
26807 PR tree-optimization/55862
26808 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
26809 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
26810 not if it is contained therein.
26812 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26814 * config/avr/t-avr: Typo.
26816 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26819 * config/avr/t-avr: Don't automatically rebuild
26820 $(srcdir)/config/avr/t-multilib
26821 $(srcdir)/config/avr/avr-tables.opt
26822 $(srcdir)/doc/avr-mmcu.texi
26823 (avr-mcus): New phony target to build them on request.
26824 (s-avr-mlib, s-avr-mmcu-texi): Remove.
26825 * avr/avr-mcus.def: Adjust comments.
26827 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
26829 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
26831 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26833 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
26835 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26837 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
26839 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
26841 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
26842 to generate profiling.
26843 * config/rs6000/aix64.h (LIB_SPEC): Same.
26845 2013-01-04 Andrew Pinski <apinski@cavium.com>
26847 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
26849 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
26851 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
26853 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
26855 (ix86_expand_move): Ditto.
26856 (ix86_zero_extend_to_Pmode): Ditto.
26857 (ix86_expand_call): Ditto.
26858 (ix86_expand_special_args_builtin): Ditto.
26859 (ix86_expand_builtin): Ditto.
26861 2013-01-04 Richard Biener <rguenther@suse.de>
26863 PR tree-optimization/55862
26864 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
26865 translating them through PHI nodes.
26867 2013-01-04 Martin Jambor <mjambor@suse.cz>
26869 PR tree-optimization/55755
26870 * tree-sra.c (sra_modify_assign): Do not check that an access has no
26871 children when trying to avoid producing a VIEW_CONVERT_EXPR.
26873 2013-01-04 Marek Polacek <polacek@redhat.com>
26875 PR middle-end/55859
26876 * opts.c (default_options_optimization): Clarify error message.
26878 2013-01-04 Richard Biener <rguenther@suse.de>
26880 PR middle-end/55863
26881 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
26884 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26887 * config/pa/pa.md (movsi): Revert previous change.
26888 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
26891 2013-01-03 Richard Henderson <rth@redhat.com>
26893 * config/i386/i386.c (ix86_expand_move): Always assign to op1
26894 after eliminating TLS symbols.
26896 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
26899 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
26900 * graphite-poly.c (debug_gmp_value): Likewise.
26902 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
26905 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
26906 selected code model, define __code_mode_small__, __code_model_medium__,
26907 __code_model_large__, __code_model_32__ or __code_model_kernel__.
26908 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
26909 xchg temporary register with %k. Declare temporary register as
26911 [__x86_64__]: For medium and large code models, preserve %rbx register.
26913 2013-01-03 Richard Biener <rguenther@suse.de>
26915 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
26916 (dump_subscript): Adjust.
26917 (finalize_ddr_dependent): Do not dump redundant info.
26918 (analyze_siv_subscript): Adjust.
26919 (subscript_dependence_tester): Likewise.
26920 (compute_affine_dependence): Likewise.
26922 2013-01-03 Richard Biener <rguenther@suse.de>
26925 2013-01-03 Richard Biener <rguenther@suse.de>
26927 PR tree-optimization/55857
26928 * tree-vect-stmts.c (vectorizable_load): Do not setup
26929 re-alignment for invariant loads.
26931 2013-01-02 Richard Biener <rguenther@suse.de>
26933 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
26934 invariant load do not generate a vector load from the scalar location.
26936 2013-01-03 Richard Biener <rguenther@suse.de>
26938 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
26939 for not vectorizing.
26940 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
26941 not build INDIRECT_REFs, call get_name once only.
26942 (vect_create_data_ref_ptr): Likewise. Dump base object kind
26943 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
26945 2013-01-03 Richard Biener <rguenther@suse.de>
26947 PR tree-optimization/55857
26948 * tree-vect-stmts.c (vectorizable_load): Do not setup
26949 re-alignment for invariant loads.
26951 2013-01-03 Richard Biener <rguenther@suse.de>
26954 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
26955 prefer a built-in decl.
26957 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26959 * gcc.c (process_command): Update copyright notice dates.
26960 * gcov.c (print_version): Likewise.
26961 * gcov-dump.c (print_version): Likewise.
26963 PR rtl-optimization/55838
26964 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
26965 iv0.step, iv1.step and step.
26967 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26968 Marc Glisse <marc.glisse@inria.fr>
26970 PR tree-optimization/55832
26971 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
26972 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
26973 integer_{one,zero}_node.
26975 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26978 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
26979 * var-tracking.c (reverse_op): Don't add reverse ops to
26980 VALUEs that have already
26981 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
26983 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
26985 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
26987 2013-01-02 Teresa Johnson <tejohnson@google.com>
26989 * dumpfile.c (dump_loc): Print filename with location.
26990 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
26991 new location_t parameter to emit complete unroll message with
26992 new dump framework.
26993 (canonicalize_loop_induction_variables): Compute loops location
26994 and pass to try_unroll_loop_completely.
26995 * loop-unroll.c (report_unroll_peel): New function.
26996 (peel_loops_completely): Use new dump format with location
26997 for main dumpfile message, and invoke report_unroll_peel on success.
26998 (decide_unrolling_and_peeling): Ditto.
26999 (decide_peel_once_rolling): Remove old dumpfile message subsumed
27000 by report_unroll_peel.
27001 (decide_peel_completely): Ditto.
27002 (decide_unroll_constant_iterations): Ditto.
27003 (decide_unroll_runtime_iterations): Ditto.
27004 (decide_peel_simple): Ditto.
27005 (decide_unroll_stupid): Ditto.
27006 * cfgloop.c (get_loop_location): New function.
27007 * cfgloop.h (get_loop_location): Declare.
27009 2013-01-02 Sriraman Tallam <tmsriram@google.com>
27011 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
27014 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27016 PR middle-end/55198
27017 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
27018 BLKmode objects when EXPAND_MEMORY is specified.
27020 2013-01-02 Sriraman Tallam <tmsriram@google.com>
27022 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
27024 (fold_builtin_cpu): Do not share cpu model decls across statements.
27026 2013-01-02 Jason Merrill <jason@redhat.com>
27029 * tree.c (build_array_type_1): Revert earlier change.
27031 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
27033 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
27035 * config/aarch64/aarch64-tune.md: Re-generate.
27037 2013-01-02 Richard Biener <rguenther@suse.de>
27039 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
27040 invariant load do not generate a vector load from the scalar location.
27042 2013-01-02 Richard Biener <rguenther@suse.de>
27045 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
27046 * configure: Regenerate.
27048 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27050 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
27051 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
27052 (expand_builtin_int_roundingfn_2): Keep the original target around
27053 for the fallback case.
27055 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27057 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
27058 to be clear for sign changes.
27060 2013-01-01 Jan Hubicka <jh@suse.cz>
27062 * ipa-inline-analysis.c: Fix formatting.
27064 2013-01-01 Jakub Jelinek <jakub@redhat.com>
27066 PR tree-optimization/55831
27067 * tree-vect-loop.c (get_initial_def_for_induction): Use
27068 gsi_after_labels instead of gsi_start_bb.
27070 Copyright (C) 2013 Free Software Foundation, Inc.
27072 Copying and distribution of this file, with or without modification,
27073 are permitted in any medium without royalty provided the copyright
27074 notice and this notice are preserved.