1 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
3 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
4 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
5 (enum move_pos): Move to tree-ssa-loop-im.h
6 * cfgloop.h: Move some prototypes.
7 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
8 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
9 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
10 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
11 (loop_containing_stmt): Relocate from tree-flow-inline.h.
12 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
13 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
14 (enum move_pos): Relocate here.
15 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
17 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
18 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim, make_pass_lim):
19 Relocate here from tree-ssa-loop.c.
20 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
22 (loop_edge_to_cancel, unloop_loops): Make static.
23 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
24 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
25 (tree_complete_unroll, gate_tree_complete_unroll,
26 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
27 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
28 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
29 * tree-ssa-loop-ivopts.c: Remove local prototypes.
30 (stmt_invariant_in_loop_p): Remove unused function.
31 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
32 * tree-ssa-loop-manip.h: New file. Add prototypes.
33 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
34 (gcov_type_to_double_int): Move to cfgloop.h.
35 (double_int_cmp, bound_index,
36 estimate_numbers_of_iterations_loop): Make static.
37 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
38 (max_loop_iterations): Factor out get_max_loop_iterations.
39 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
41 * tree-ssa-loop-niter.h: New file. Add prototypes.
42 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
43 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
44 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
45 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
46 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
47 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
48 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
49 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
50 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
51 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
52 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
53 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
54 pass_data_complete_unroll, make_pass_complete_unroll,
55 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
56 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
57 tree-ssa-loop-ivcanon.c.
58 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
59 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
60 tree-ssa-loop-prefetch.c.
61 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
63 (get_lsm_tmp_name): Relocate and add suffix parameter.
64 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
65 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
66 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
67 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
68 (get_estimated_loop_iterations): Factor out accessor from
69 estimated_loop_iterations in tree-ssa-loop-niter.c.
70 (get_max_loop_iterations): Factor out accessor from _max_loop_iterations
72 * loop-unroll.c (decide_unroll_constant_iterations,
73 decide_unroll_runtime_iterations, decide_peel_simple,
74 decide_unroll_stupid): Use new get_* accessors.
76 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
78 PR tree-optimization/20318
79 * doc/extend.texi (returns_nonnull): New function attribute.
80 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
82 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
83 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
85 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
88 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
89 false if both components are bitfields.
91 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
93 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
94 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
95 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
96 * config/aarch64/aarch64.h
97 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
98 * config/aarch64/aarch64-simd-builtins.def
99 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
101 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
103 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
105 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
107 * config/aarch64/arm_neon.h (vdiv_f64): Added.
109 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
111 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
112 (vneg_f64): New intrinsic.
113 (vneg_s8): Asm replaced with C.
114 (vneg_s16): Likewise.
115 (vneg_s32): Likewise.
116 (vneg_s64): New intrinsic.
117 (vnegq_f32): Asm replaced with C.
118 (vnegq_f64): Likewise.
119 (vnegq_s8): Likewise.
120 (vnegq_s16): Likewise.
121 (vnegq_s32): Likewise.
122 (vnegq_s64): Likewise.
124 2013-10-09 Renlin Li <Renlin.Li@arm.com>
126 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
128 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
130 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
131 packed stack special handling.
132 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
133 back to fixed stack slots for FPRs saved due to stdarg.
135 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
137 * config/s390/s390.c (s390_frame_info): Restructure function.
139 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
141 * config/s390/s390.c (struct s390_frame_layout): New field
143 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
144 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
145 regs_ever_clobbered to char*.
146 (s390_regs_ever_clobbered): Check crtl->saves_all_registers
147 instead of cfun->has_nonlocal_label. Ignore frame related restore
149 (s390_register_info): Enable FPR save slots. Move/Copy some
150 functionality into ...
151 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
152 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
154 (s390_frame_info): Do gpr slot allocation here now. stdarg does
155 not imply a stack frame.
156 (s390_init_frame_layout): Remove variable clobbered_regs.
157 (s390_update_register_info): Remove function.
158 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
159 cannot be used for register renaming.
160 (s390_hard_regno_scratch_ok): New function.
161 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
162 (s390_initial_elimination_offset): Change offset calculation of
163 the return address pointer.
164 (save_gprs): Deal with only r6 being saved from the call-saved
166 (restore_gprs): Set frame related flag.
167 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New
169 (s390_emit_prologue): Call s390_register_info instead of
170 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
171 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
172 (s390_optimize_prologue): Call s390_optimize_register_info.
173 Try to remove also FPR slot save/restore INSNs. Remove frame
174 related flags from restore INSNs.
176 2013-10-08 DJ Delorie <dj@redhat.com>
178 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
181 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
182 avoid conflict with the MI use of %c.
183 * config/rl78/rl78-real.md: change %c to %C throughout.
184 * config/rl78/rl78-virt.md: Likewise.
186 2013-10-08 Jan Hubicka <jh@suse.cz>
188 * config/i386/i386.c (ix86_option_override_internal): Switch
189 to SSE math for -ffast-math when target ISA supports SSE2.
191 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
193 * tree-flow.h: Remove some prototypes.
194 * tree.h: Remove some protypes, add a couple.
195 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
196 using_eh_for_cleanups_p): Add interface routines for front ends.
197 * tree-eh.h: New file. Add protoptyes.
198 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
199 (add_stmt_to_eh_lp_fn): Make static.
200 (lower_try_finally): Use new using_eh_for_cleanups_p.
201 * emit-rtl.c: Include tree-eh.h.
202 * gimple.h: Include tree-eh.h.
204 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
206 PR tree-optimization/58480
207 * tree-vrp.c (infer_nonnull_range): New function.
208 (infer_value_range): Call infer_nonnull_range.
210 2013-10-08 Dehao Chen <dehao@google.com>
212 PR tree-optimization/58619
213 * tree-inline.c (copy_phis_for_bb): Combine location data
216 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
219 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
220 RTX_FRAME_RELATED_P on INSN.
222 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
224 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
225 (altivec_expand_vec_perm_const): Call it.
227 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
229 * config/rs6000/vector.md (mov<mode>): Emit permuted move
230 sequences for LE VSX loads and stores at expand time.
231 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
233 * config/rs6000/rs6000.c (rs6000_const_vec): New.
234 (rs6000_gen_le_vsx_permute): New.
235 (rs6000_gen_le_vsx_load): New.
236 (rs6000_gen_le_vsx_store): New.
237 (rs6000_gen_le_vsx_move): New.
238 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
239 (*vsx_le_perm_load_v4si): New.
240 (*vsx_le_perm_load_v8hi): New.
241 (*vsx_le_perm_load_v16qi): New.
242 (*vsx_le_perm_store_v2di): New.
243 (*vsx_le_perm_store_v4si): New.
244 (*vsx_le_perm_store_v8hi): New.
245 (*vsx_le_perm_store_v16qi): New.
246 (*vsx_xxpermdi2_le_<mode>): New.
247 (*vsx_xxpermdi4_le_<mode>): New.
248 (*vsx_xxpermdi8_le_V8HI): New.
249 (*vsx_xxpermdi16_le_V16QI): New.
250 (*vsx_lxvd2x2_le_<mode>): New.
251 (*vsx_lxvd2x4_le_<mode>): New.
252 (*vsx_lxvd2x8_le_V8HI): New.
253 (*vsx_lxvd2x16_le_V16QI): New.
254 (*vsx_stxvd2x2_le_<mode>): New.
255 (*vsx_stxvd2x4_le_<mode>): New.
256 (*vsx_stxvd2x8_le_V8HI): New.
257 (*vsx_stxvd2x16_le_V16QI): New.
259 2013-10-07 Renlin Li <Renlin.Li@arm.com>
261 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
263 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
265 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
266 loop to work also for 31bit ABI.
267 Save the stack pointer for frame_size > 0.
269 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
271 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
272 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
273 constraint letters from expanders.
274 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
275 retry count to general_operand.
276 ("tabort"): Give operand 0 a mode.
277 ("tabort_1"): Add mode and constraint letter for operand 0.
278 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
280 2013-10-04 Jeff Law <law@redhat.com>
282 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
284 * tree-ssa-threadedge.c (thread_through_normal_block): Broken out of ...
285 (thread_across_edge): Here. Call it.
287 2013-10-04 Cary Coutant <ccoutant@google.com>
289 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
290 discarding a location list expression (or a piece of one).
292 2013-10-03 Jan Hubicka <jh@suse.cz>
294 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
295 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
296 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
298 2013-10-03 Jan Hubicka <jh@suse.cz>
300 * config/i386/i386.c (ix86_option_override_internal): Do not enable
301 accumulate-outgoing-args when producing unwind info.
303 2013-10-03 Wei Mi <wmi@google.com>
305 * lra-constraints.c (insert_move_for_subreg): New function
306 extracted from simplify_operand_subreg.
307 (simplify_operand_subreg): Add reload for paradoxical subreg.
309 2013-10-03 Rong Xu <xur@google.com>
311 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
312 the candidate of builtin_expect such that we should fix the
313 size/time estimation.
314 (estimate_function_body_sizes): Do the acutally size/time fix-up
317 2013-10-03 Rong Xu <xur@google.com>
319 * predict.c (tree_predict_by_opcode): Get the probability
320 for builtin_expect from param builtin_expect_probability.
321 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
322 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
323 * doc/invoke.texi: Add documentation for builtin-expect-probability.
325 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
328 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
329 * calls.c (alloca_call_p): Use get_callee_fndecl.
330 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
331 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
333 (vrp_visit_stmt): Remove duplicated code.
335 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
337 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
338 ceildf2, btruncdf2, instead of vsx_* name.
340 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
341 iterators to only do V2DF and V4SF here. Move the DF code to
342 rs6000.md where it is combined with SF mode. Replace <VSv> with
343 just 'v' since only vector operations are handled with these insns
344 after moving the DF support to rs6000.md.
345 (vsx_sub<mode>3): Likewise.
346 (vsx_mul<mode>3): Likewise.
347 (vsx_div<mode>3): Likewise.
348 (vsx_fre<mode>2): Likewise.
349 (vsx_neg<mode>2): Likewise.
350 (vsx_abs<mode>2): Likewise.
351 (vsx_nabs<mode>2): Likewise.
352 (vsx_smax<mode>3): Likewise.
353 (vsx_smin<mode>3): Likewise.
354 (vsx_sqrt<mode>2): Likewise.
355 (vsx_rsqrte<mode>2): Likewise.
356 (vsx_fms<mode>4): Likewise.
357 (vsx_nfma<mode>4): Likewise.
358 (vsx_copysign<mode>3): Likewise.
359 (vsx_btrunc<mode>2): Likewise.
360 (vsx_floor<mode>2): Likewise.
361 (vsx_ceil<mode>2): Likewise.
362 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
363 (vsx_sminsf3): Likewise.
364 (vsx_fmadf4): Likewise.
365 (vsx_fmsdf4): Likewise.
366 (vsx_nfmadf4): Likewise.
367 (vsx_nfmsdf4): Likewise.
368 (vsx_cmpdf_internal1): Likewise.
370 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
371 simpler to select whether a target has SPE or traditional floating
372 point support in iterators.
373 (TARGET_DF_SPE): Likewise.
374 (TARGET_SF_FPR): Likewise.
375 (TARGET_DF_FPR): Likewise.
376 (TARGET_SF_INSN): Macros to say whether floating point support
377 exists for a given operation for expanders.
378 (TARGET_DF_INSN): Likewise.
380 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
381 combining of SF/DF mode operations, using both traditional and VSX
389 (abs<mode>2): Combine SF/DF modes using traditional floating point
390 instructions. Add support for using the upper DF registers with
391 VSX support, and SF registers with power8-vector support. Update
392 expanders for operations supported by both the SPE and traditional
393 floating point units.
394 (abs<mode>2_fpr): Likewise.
395 (nabs<mode>2): Likewise.
396 (nabs<mode>2_fpr): Likewise.
397 (neg<mode>2): Likewise.
398 (neg<mode>2_fpr): Likewise.
399 (add<mode>3): Likewise.
400 (add<mode>3_fpr): Likewise.
401 (sub<mode>3): Likewise.
402 (sub<mode>3_fpr): Likewise.
403 (mul<mode>3): Likewise.
404 (mul<mode>3_fpr): Likewise.
405 (div<mode>3): Likewise.
406 (div<mode>3_fpr): Likewise.
407 (sqrt<mode>3): Likewise.
408 (sqrt<mode>3_fpr): Likewise.
410 (rsqrt<mode>2): Likewise.
411 (cmp<mode>_fpr): Likewise.
412 (smax<mode>3): Likewise.
413 (smin<mode>3): Likewise.
414 (smax<mode>3_vsx): Likewise.
415 (smin<mode>3_vsx): Likewise.
416 (negsf2): Delete SF operations that are merged with DF.
423 (fmasf4_fpr): Likewise.
424 (fmssf4_fpr): Likewise.
425 (nfmasf4_fpr): Likewise.
426 (nfmssf4_fpr): Likewise.
428 (rsqrtsf_internal1): Likewise.
431 (cmpsf_internal1): Likewise.
432 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
433 (negdf2): Delete DF operations that are merged with SF.
441 (rsqrtdf_internal1): Likewise.
442 (fmadf4_fpr): Likewise.
443 (fmsdf4_fpr): Likewise.
444 (nfmadf4_fpr): Likewise.
445 (nfmsdf4_fpr): Likewise.
449 (cmpdf_internal1): Likewise.
450 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
451 (btrunc<mode>2): Delete separate expander, and combine with the
452 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
453 (btrunc<mode>2_fpr): Likewise.
454 (ceil<mode>2): Likewise.
455 (ceil<mode>2_fpr): Likewise.
456 (floor<mode>2): Likewise.
457 (floor<mode>2_fpr): Likewise.
458 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
459 Add support for using the upper registers with VSX and
460 power8-vector. Move insns to be closer to the define_expands. On
461 VSX systems, prefer the traditional form of FMA over the VSX
462 version, since the traditional form allows the target not to
463 overlap with the inputs.
464 (fms<mode>4_fpr): Likewise.
465 (nfma<mode>4_fpr): Likewise.
466 (nfms<mode>4_fpr): Likewise.
468 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
469 Richard Earnshaw <richard.earnshaw@arm.com>
471 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
472 (struct mult_cost_table): Likewise.
473 (struct mem_cost_table): Likewise.
474 (struct fp_cost_table): Likewise.
475 (struct vector_cost_table): Likewise.
476 (cpu_cost_table): Likewise.
477 * config/arm/arm.opt (mold-rts-costs): New option.
478 (mnew-generic-costs): Likewise.
479 * config/arm/arm.c (generic_extra_costs): New table.
480 (cortexa15_extra_costs): Likewise.
481 (arm_slowmul_tune): Use NULL as new costs.
482 (arm_fastmul_tune): Likewise.
483 (arm_strongarm_tune): Likewise.
484 (arm_xscale_tune): Likewise.
485 (arm_9e_tune): Likewise.
486 (arm_v6t2_tune): Likewise.
487 (arm_cortex_a5_tune): Likewise.
488 (arm_cortex_a9_tune): Likewise.
489 (arm_v6m_tune): Likewise.
490 (arm_fa726te_tune): Likewise.
491 (arm_cortex_a15_tune): Use cortex15_extra_costs.
492 (arm_cortex_tune): Use generict_extra_costs.
493 (shifter_op_p): New function.
494 (arm_unspec_cost): Likewise.
495 (LIBCALL_COST): Define.
496 (arm_new_rtx_costs): New function.
497 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
498 table is available. Use old costs otherwise unless mnew-generic-costs
500 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
501 (cpu_cost_table): Declare.
503 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
506 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
507 (*subs_mul_imm_<mode>)
508 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
509 (*sub_<shift>_<mode>)
510 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
513 2013-10-03 Ian Bolton <ian.bolton@arm.com>
515 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
517 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
519 2013-10-02 Teresa Johnson <tejohnson@google.com>
521 * predict.c (probably_never_executed): New function.
522 (probably_never_executed_bb_p): Invoke probably_never_executed.
523 (probably_never_executed_edge_p): Ditto.
524 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
525 Treat profile insanities conservatively.
527 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
529 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
531 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
533 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
534 use it. Use smaller increase for scratch. Don't increase reject
535 for early clobber scratch.
536 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
537 setting eliminated regs except setting fp from hfp.
538 (lra_eliminate): Check lra_insn_recog_data on NULL.
540 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
543 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
544 setting -mvsx-timode by default until the underlying problem is fixed.
545 (RS6000_CPU, power7 defaults): Likewise.
547 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
549 * config/x-linux (host-linux.o): Remove header dependencies.
550 Use $(COMPILE) and $(POSTCOMPILE).
551 * config/t-linux-android (linux-android.o): Ditto.
553 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
555 * Makefile.in (expmed.o-warn): Remove.
557 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
559 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
560 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
562 2013-10-02 Teresa Johnson <tejohnson@google.com>
564 * dojump.c (do_jump_1): Divide probability between
565 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
567 2013-10-02 Tom Tromey <tromey@redhat.com>
569 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
571 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
573 * tree-flow.h: Remove some prototypes.
574 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
575 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
576 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
577 mark_virtual_phi_result_for_renaming): Relocate here.
578 * tree-into-ssa.h: Add prototypes.
579 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
580 single_pred_before_succ_order.
581 (blocks_in_phiopt_order): Rename and move to cfganal.c.
582 (nonfreeing_call_p) Move to gimple.c.
583 * cfganal.c (single_pred_before_succ_order): Move and renamed from
585 * basic-block.h (single_pred_before_succ_order): Add prototype.
586 * gimple.c (nonfreeing_call_p): Relocate here.
587 * gimple.h: Add prototype.
588 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
589 * tree-ssa-dom.h: New file. Relocate prototypes here.
590 * tree-ssa.h: Include tree-ssa-dom.h.
592 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
594 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
595 Use $(COMPILE) and $(POSTCOMPILE).
597 * config/alpha/x-alpha (driver-alpha.o): Ditto.
599 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
601 * tree-flow.h: Remove some prototypes.
602 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
603 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
604 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
605 propagate_tree_value*): Move from here to...
606 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
607 propagate_tree_value*): Relocate here.
608 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
609 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
610 * gimple-fold.c: Remove gimple-fold.h from include list.
611 * tree-vrp.c: Remove gimple-fold.h from include list.
612 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
613 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
614 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
615 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
616 * sese.c: Add tree-ssa-propagate.h to include list.
618 2013-10-02 Richard Biener <rguenther@suse.de>
620 * tree-loop-distribution.c: Include tree-vectorizer.h for
622 (enum partition_kind): Remove PKIND_REDUCTION.
623 (struct partition_s): Remove has_writes member, add reduction_p member.
624 (partition_alloc): Adjust.
625 (partition_builtin_p): Likewise.
626 (partition_has_writes): Remove.
627 (partition_reduction_p): New function.
628 (partition_merge_into): Likewise.
629 (generate_code_for_partition): Commonize builtin partition
631 (rdg_cannot_recompute_vertex_p): Remove.
632 (already_processed_vertex_p): Likewise.
633 (rdg_flag_vertex): Do not set has_writes.
634 (classify_partition): Adjust.
635 (rdg_build_partitions): Do not set has_writes, treat all
636 partitions as useful.
637 (distribute_loop): Record number of library calls generated. Adjust.
638 (tree_loop_distribution): Report number of loops and library
639 calls generated as opt-info.
641 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
643 * tree-flow.h: Include new .h files. Move prototypes.
644 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
645 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
646 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
647 * tree-pretty-print.h: Add prototypes from tree-flow.h.
648 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
649 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
650 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
651 * tree.h (get_ref_base_and_extent): Move prototype out.
652 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
654 * gimple-low.h: New File. Add prototypes from tree-flow.h.
655 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
656 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
657 * tree-scalar-evolution.c: Include tree.h.
658 * sese.c: Include tree.h.
659 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
660 * dwarf2out.c: Include tree-dfa.h.
661 * tree-chrec.c: Include tree.h.
662 * tree-data-ref.c: Include tree.h.
664 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
666 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
669 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
671 * config/t-sol2 (sol2-c.o): Remove header dependencies.
672 Use $(COMPILE) and $(POSTCOMPILE).
673 (sol2-cxx.o): Likewise.
674 (sol2-stubs.o): Likewise.
676 * config/x-solaris (host-solaris.o): Likewise.
678 * config/sparc/t-sparc (sparc.o): Remove.
679 (sparc-c.o): Remove header dependencies.
680 Use $(COMPILE) and $(POSTCOMPILE).
681 * config/sparc/x-sparc: Likewise.
683 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
685 * config/arc/arc-opts.h: Add 2013 to Copyright years.
686 * config/arc/arc700.md: Likewise.
687 * config/arc/arc-modes.def: Likewise.
688 * config/arc/arc-simd.h: Likewise.
689 * config/arc/t-arc-uClibc: Likewise.
690 * config/arc/t-arc-newlib: Likewise.
692 2013-10-02 Renlin Li <renlin.li@arm.com>
694 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
696 (aarch64_expand_epilogue): Likewise.
698 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
699 Yufeng Zhang <yufeng.zhang@arm.com>
701 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
703 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
704 'base_in' represent a conversion and legal_cast_p_1 holds; set
705 'base_in' with the returned value from get_unwidened.
707 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
709 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
710 plus_constant is not used.
712 2013-10-01 Wei Mi <wmi@google.com>
714 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
715 * config/i386/i386.md: Add define_peephole2 to
716 break partial reg stall for cvtss2sd/cvtsd2ss.
718 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
720 * config/arc/arc.c (pass_arc_ifcvt::clone):
721 Update for ctxt_ -> m_ctxt change.
723 2013-10-01 Jeff Law <law@redhat.com>
725 * tree-ssa-threadupdate.c (struct redirection_data): Delete
726 outgoing_edge and intermediate_edge fields. Instead store the path.
727 (redirection_data::hash): Hash on the last edge's destination index.
728 (redirection_data::equal): Check the entire thread path.
729 (lookup_redirectio_data): Corresponding changes.
730 (create_edge_and_update_destination_phis): Likewise.
731 (thread_single_edge): Likewise.
733 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
734 Diego Novillo <dnovillo@google.com>
736 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
737 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
738 (vld32wh_insn, vld32wl_insn): Delete commented-out old
739 versions of these patterns.
741 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
742 (__builtin_arc_aligned): Likewise.
744 * config/arc/arc.md: Expand adc_0 comment stating the intended
745 purpose and why it isn't ready.
746 Replace commented out call_value_via_label_mixed with a
747 plain comment about bl_s.
749 * config/arc/arc.c (stdio.h): Don't include directly.
750 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
751 Remove if (1) condition.
752 (arc_encode_section_info): Fix comment.
754 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
756 * config/arc/arc.c (arc_conditional_register_usage):
757 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
758 Also set reg_alloc_order for DMA config regs.
760 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
761 Jeremy Bennett <jeremy.bennett@embecosm.com>
763 * doc/install.texi (--with-cpu): Mention ARC.
764 (arc-*-elf32): New paragraph.
765 (arc-linux-uclibc): Likewise.
766 * doc/md.texi (Machine Constraints): Add ARC part.
767 * doc/invoke.texi: (menu): Add ARC Options.
768 (Machine Dependent Options) <ARC Options>: Add synopsis.
769 (node ARC Options): Add.
770 * doc/extend.texi (long_call / short_call attribute): Add ARC.
771 (ARC Built-in Functions): New section defining
772 generic ARC built-in functions.
773 (ARC SIMD Built-in Functions): New section defining SIMD specific
775 (Declaring Attributes of Functions): Extended
776 description of short_call and long_call attributes for ARC and
779 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
780 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
781 Joern Rennecke <joern.rennecke@embecosm.com>
782 Muhammad Khurram Riaz <khurram.riaz@arc.com>
783 Brendan Kehoe <brendan@zen.org>
784 Michael Eager <eager@eagercon.com>
785 Simon Cook <simon.cook@embecosm.com>
786 Jeremy Bennett <jeremy.bennett@embecosm.com>
788 * config/arc, common/config/arc: New directories.
790 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
791 Brendan Kehoe <brendan@zen.org>
792 Simon Cook <simon.cook@embecosm.com>
794 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
796 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
798 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
799 * tree-ssa-coalesce.h: New. Move prototype to here.
800 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
801 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
802 (gimple_can_coalesce_p): Move to...
803 * gimple.c (gimple_can_coalesce_p): Here.
805 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
807 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
808 (dump_decl_set): Move to gimple.c.
809 * gimple.h: Don't include tree-ssa-operands.h.
810 (dump_decl_set): Add prototype.
811 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
812 Move to gimple-ssa.h.
813 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
814 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
815 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
816 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
817 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
819 (dump_decl_set): Relocate here.
820 * gimple-ssa.h: New file.
821 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
822 Relocate from gimple.h.
823 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
824 * tree-ssa-operands.c (swap_ssa_operands): Rename from
825 swap_tree_operands and remove non-ssa path.
826 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
827 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
829 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
830 vect_is_simple_reduction_1): Use swap_ssa_operands.
831 * tree-flow.h: Move various prototypes to tree-phinodes.h.
832 (enum need_phi_state): Move to tree-into-ssa.c.
833 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
834 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
835 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
836 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
837 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
838 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
839 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
840 num_imm_uses): Move to ssa-iterators.h.
841 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
842 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
844 (op_iter_done, op_iter_next_def, op_iter_next_tree,
845 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
846 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
847 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
848 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
849 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
850 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
851 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
852 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
853 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
854 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
855 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
856 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
857 (set_phi_nodes): Move to tree-phinodes.h.
858 * tree-ssa-operands.h (enum ssa_op_iter_type,
859 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
860 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
861 (dump_decl_set): Remove prototype.
862 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
863 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
864 (set_phi_nodes): Relocate from tree-flow-inline.h.
865 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
867 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
868 include list. Temporarily add gimple.h to include list.
869 * ssa-iterators.h: New file.
870 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
871 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
872 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
873 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
874 Relocate from tree-ssa-operands.h.
875 (delink_imm_use, link_imm_use_to_list, link_imm_use,
876 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
877 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
878 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
879 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
880 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
881 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
882 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
883 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
884 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
885 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
886 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
887 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
888 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
889 Relocate from tree-flow-inline.h.
890 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
892 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
895 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
896 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
897 Insert '\t' to output template.
898 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
899 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
900 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
901 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
903 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
905 * doc/install.texi (Host/target specific installation notes for GCC):
906 Put @anchor before @heading.
907 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
908 Use @email for email addresses.
910 2013-10-01 Jeff Law <law@redhat.com>
912 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
913 a vec. Only delete the path if we create one without successfully
914 registering a jump thread.
915 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
917 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
918 (paths): New vector of jump threading paths.
919 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
920 (THREAD_PATH): New accessor macro for the entire thread path.
921 (lookup_redirection_data): Get intermediate and final outgoing edge
922 from the thread path.
923 (create_edge_and_update_destination_phis): Copy the threading path.
924 (ssa_fix_duplicate_block_edges): Get edges and block types from the
926 (ssa_redirect_edges): Get edges and block types from the jump threading
927 path. Free the path vector.
928 (thread_block): Get edges from the jump threading path. Look at the
929 entire path to see if we thread to a loop exit. If we cancel a jump
930 thread request, then free the path vector.
931 (thread_single_edge): Get edges and block types from the jump threading
932 path. Free the path vector.
933 (thread_through_loop_header): Get edges and block types from the jump
934 threading path. Free the path vector.
935 (mark_threaded_blocks): Iterate over the vector of paths and store
936 the path on the appropriate edge. Get edges and block types from the
938 (mark_threaded_blocks): Get edges and block types from the jump
939 threading path. Free the path vector.
940 (thread_through_all_blocks): Use the vector of paths rather than
941 a vector of 3-edge sets.
942 (register_jump_thread): Accept pointer to a path vector rather
943 than the path vector itself. Store the path vector for later use.
946 2013-10-01 Jakub Jelinek <jakub@redhat.com>
947 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
950 * config/s390/s390.c (s390_split_branches): Modify check for table
952 (s390_chunkify_start): Rearrange table jump insn check in order to
953 deal with compare and branch insns correctly.
955 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
959 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
960 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
961 define_insn_and_split.
962 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
963 (shiftsi3_compare): New pattern.
965 * config/arm/unspecs.md (UNSPEC_RRX): New.
967 2013-10-01 Alan Modra <amodra@gmail.com>
969 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
970 casing inout operands.
972 2013-10-01 Richard Biener <rguenther@suse.de>
974 PR tree-optimization/58553
975 * tree-loop-distribution.c (struct partition_s): Add niter member.
976 (classify_partition): Populate niter member for the partition
977 and properly identify whether the relevant store happens before
978 or after the loop exit.
979 (generate_memset_builtin): Use niter member from the partition.
980 (generate_memcpy_builtin): Likewise.
982 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
984 * vec.h (vec_prefix, vec): Prefix member names with "m_".
985 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
987 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
989 * basic-block.h (edge_list): Prefix member names with "m_".
990 * context.h (context): Likewise.
991 * domwalk.h (dom_walker): Likewise.
992 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
993 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
994 * hash-table.h (hash_table): Likewise.
995 * machmode.h (bit_field_mode_iterator): Likewise.
996 * pass_manager.h (pass_list): Likewise.
997 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
998 * tree-pass.h (pass_data): Likewise.
999 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
1000 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
1001 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
1002 * asan.c (pass_data_asan): Update accordingly.
1003 * cfganal.c (control_dependences::find_control_dependence): Likewise.
1004 (control_dependences::control_dependences): Likewise.
1005 (control_dependences::~control_dependences): Likewise.
1006 (control_dependences::~control_dependences): Likewise.
1007 (control_dependences::get_edges_dependent_on): Likewise.
1008 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
1009 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
1010 * context.c (gcc::context::context): Likewise.
1011 * cprop.c (pass_rtl_cprop::clone): Likewise.
1012 * domwalk.c (dom_walker::walk): Likewise.
1013 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
1014 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
1015 * mode-switching.c (pass_mode_switching::clone): Likewise.
1016 * passes.c (opt_pass::opt_pass): Likewise.
1017 (pass_manager::pass_manager): Likewise.
1018 * predict.c (pass_strip_predict_hints::clone): Likewise.
1019 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
1020 (pass_split_all_insns::clone): Likewise.
1021 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
1023 (bit_field_mode_iterator::next_mode): Likewise.
1024 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
1025 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
1026 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
1027 * tree-complex.c (pass_lower_complex::clone): Likewise.
1028 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
1029 * tree-object-size.c (pass_object_sizes::clone): Likewise.
1030 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
1031 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
1032 (pass_fold_builtins::clone): Likewise.
1033 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
1034 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
1035 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
1036 (pass_cd_dce::clone): Likewise.
1037 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
1038 (pass_phi_only_cprop::clone): Likewise.
1039 * tree-ssa-dse.c (pass_dse::clone): Likewise.
1040 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
1041 * tree-ssa-loop.c (pass_lim::clone): Likewise.
1042 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
1043 * tree-ssa-pre.c (pass_fre::clone): Likewise.
1044 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
1045 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
1046 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
1047 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
1048 * tree-vrp.c (pass_vrp::clone): Likewise.
1049 * tsan.c (pass_tsan::clone): Likewise.
1051 2013-09-30 Jakub Jelinek <jakub@redhat.com>
1054 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
1055 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
1058 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
1059 optimization, punt if sign_bit_p looked through any zero extension.
1061 2013-09-30 Teresa Johnson <tejohnson@google.com>
1063 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
1064 Update redirected out edge count in joiner case.
1065 (ssa_redirect_edges): Common the joiner and non-joiner cases
1066 so that joiner case gets profile updates.
1068 2013-09-30 Richard Biener <rguenther@suse.de>
1070 PR tree-optimization/58554
1071 * tree-loop-distribution.c (classify_partition): Require
1072 unconditionally executed stores for memcpy and memset recognition.
1073 (tree_loop_distribution): Calculate dominance info.
1075 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
1077 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
1078 (NO_PROFILE_COUNTERS): Likewise.
1079 (PROFILE_HOOK): Likewise.
1080 (FUNCTION_PROFILER): Likewise.
1081 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
1083 2013-09-30 Iain Sandoe <iain@codesourcery.com>
1085 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
1086 calls and defines in TARGET_MACHO conditional.
1087 (load_macho_picbase_di): Likewise.
1088 (reload_macho_picbase): Likewise.
1089 (reload_macho_picbase_si): Likewise.
1090 (reload_macho_picbase_di): Likewise.
1091 (nonlocal_goto_receiver): Likewise.
1093 2013-09-30 Nick Clifton <nickc@redhat.com>
1095 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
1096 that use the MSP430X ISA.
1097 (msp430_option_override): Scan -mmcu command line option for any
1098 MCU name that supports the MSP430X ISA.
1099 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
1100 -mmcu options which enable the MSP430X ISA.
1102 2013-09-30 Richard Biener <rguenther@suse.de>
1105 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
1106 before looking for setjmp-like calls.
1108 2013-09-29 Iain Sandoe <iain@codesourcery.com>
1111 * config/darwin-protos.h (machopic_get_function_picbase): New.
1112 * config/darwin.c (machopic_get_function_picbase): New.
1113 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
1114 label for a new func. (load_macho_picbase_di): Likewise.
1115 (reload_macho_picbase): New expand.
1116 (reload_macho_picbase_si): New insn.
1117 (reload_macho_picbase_di): New insn.
1118 (nonlocal_goto_receiver): New define and split.
1119 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
1120 (unspecv enum): Add UNSPECV_NLGR.
1122 2013-09-29 Iain Sandoe <iain@codesourcery.com>
1124 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
1125 that altivec registers are correctly sized on Darwin.
1127 2013-09-29 Iain Sandoe <iain@codesourcery.com>
1129 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
1130 darwin-driver.o): Use COMPILE and POSTCOMPILE.
1131 * config/x-darwin (host-darwin.o): Likewise.
1132 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
1133 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
1134 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
1136 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
1138 * doc/invoke.texi: Fix usage of @tie{} command.
1140 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
1142 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
1144 2013-09-28 Jan Hubicka <jh@suse.cz>
1146 * config/i386/x86-tune.def: Add documentation for each of the options;
1149 2013-09-28 Jan Hubicka <jh@suse.cz>
1151 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
1153 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
1154 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
1155 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
1156 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
1157 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
1159 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
1161 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
1162 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
1163 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
1164 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
1165 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
1166 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
1167 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
1168 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
1169 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
1170 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
1171 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
1172 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
1173 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
1174 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
1175 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
1176 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
1177 gimple.h, godump.c, graphite-clast-to-gimple.c,
1178 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
1179 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
1180 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
1181 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
1182 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
1183 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
1184 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
1185 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
1186 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
1187 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
1188 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
1189 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
1190 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
1191 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
1192 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
1193 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
1194 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
1195 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
1196 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
1197 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
1198 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
1199 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
1200 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
1201 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
1202 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1203 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1204 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
1205 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
1206 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
1207 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1208 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1209 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
1210 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
1211 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
1212 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
1213 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
1214 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
1215 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
1216 whitespace before "(".
1218 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
1220 * expr.h (extract_bit_field): Remove packedp parameter.
1221 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
1222 from forward declaration.
1223 (store_split_bit_field): Remove packedp arg from calls to
1224 extract_fixed_bit_field.
1225 (extract_bit_field_1): Remove packedp parameter and packedp
1226 argument from recursive calls and calls to extract_fixed_bit_field.
1227 (extract_bit_field): Remove packedp parameter and corresponding
1228 arg to extract_bit_field_1.
1229 (extract_fixed_bit_field): Remove packedp parameter. Remove code
1231 (extract_split_bit_field): Remove packedp arg from call to
1232 extract_fixed_bit_field.
1233 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
1234 (copy_blkmode_from_reg): Likewise.
1235 (copy_blkmode_to_reg): Likewise.
1236 (read_complex_part): Likewise.
1237 (store_field): Likewise.
1238 (expand_expr_real_1): Likewise.
1239 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
1240 to extract_bit_field.
1241 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
1242 call to extract_bit_field.
1243 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
1244 call to extract_bit_field.
1245 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
1246 and special packedp behavior from -fstrict-volatile-bitfields
1249 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1251 * lra-eliminations.c (init_elim_table): Guard value_p.
1253 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
1255 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
1256 DFmode, DImode, and SFmode in the upper VSX registers based on the
1257 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
1258 if -mpower8-vector. Combine -mvsx-timode handling with the rest
1259 of the VSX register handling.
1261 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
1263 (zero_extendsidi2_lfiwzx): Add support for loading into the
1264 Altivec registers with -mpower8-vector. Use wu/wv constraints to
1265 only do VSX memory options on Altivec registers.
1266 (extendsidi2_lfiwax): Likewise.
1267 (extendsfdf2_fpr): Likewise.
1268 (mov<mode>_hardfloat, SF/SD modes): Likewise.
1269 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
1270 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
1271 (movdi_internal64): Likewise.
1273 2013-09-27 Xinliang David Li <davidxl@google.com>
1275 * opts.c (finish_options): Adjust parameters
1276 according to vect cost model.
1277 (common_handle_option): Set dynamic vect cost
1279 targhooks.c (default_add_stmt_cost): Compute stmt cost
1281 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
1282 Use helper function.
1283 * tree-vectorizer.h (unlimited_cost_model): New function.
1284 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
1285 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
1287 (vect_enhance_data_refs_alignment): Ditto.
1288 * flag-types.h: New enum.
1289 * common/config/i386/i386-common.c (ix86_option_init_struct):
1290 No need to initialize vect_cost_model flag.
1291 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
1294 2013-09-27 Diego Novillo <dnovillo@google.com>
1296 * gimple.h (enum ssa_mode): Remove.
1298 2013-09-27 Paulo Matos <pmatos@broadcom.com>
1301 * gcc.dg/pr58463.c: New test.
1303 2013-09-27 Paulo Matos <pmatos@broadcom.com>
1305 * cfgloop.h (number_of_loops): Fix typo in check for null.
1307 2013-09-27 Jakub Jelinek <jakub@redhat.com>
1310 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
1311 are children of outermost saved_cfun's loop, and set it up to
1312 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
1313 if loop != loop0 and SESE region contains bbs that belong to loop0.
1315 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
1317 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
1318 (binary_scale_code_p, get_base_term, get_index_term): New functions.
1319 (set_address_segment, set_address_base, set_address_index)
1320 (set_address_disp): Accept the argument unconditionally.
1321 (baseness): Remove must_be_base_p and must_be_index_p checks.
1322 (decompose_normal_address): Classify as much as possible in the
1325 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
1327 * cse.c (count_reg_usage): Handle INT_LIST.
1328 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
1329 * reginfo.c (reg_scan_mark_refs): Likewise.
1330 * reload1.c (eliminate_regs_1): Likewise.
1332 2013-09-27 Iain Sandoe <iain@codesourcery.com>
1335 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
1338 2013-09-27 Richard Biener <rguenther@suse.de>
1340 PR tree-optimization/58459
1341 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
1342 restriction not propagating into loops.
1344 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
1346 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
1347 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
1348 * doc/tree-ssa.texi (Walking use-def chains): Delete.
1350 2013-09-26 Richard Biener <rguenther@suse.de>
1352 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
1354 2013-09-26 Richard Biener <rguenther@suse.de>
1356 * alias.h (component_uses_parent_alias_set): Rename to ...
1357 (component_uses_parent_alias_set_from): ... this.
1358 * alias.c (component_uses_parent_alias_set): Rename to ...
1359 (component_uses_parent_alias_set_from): ... this and return
1361 (reference_alias_ptr_type_1): Use the result from
1362 component_uses_parent_alias_set_from instead of stripping
1363 components one at a time.
1364 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
1366 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
1368 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
1369 Move prototypes to...
1370 * tree-ssa-ter.h: New File. Move prototypes here.
1371 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
1372 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
1373 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
1374 from is_replaceable_p.
1375 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
1376 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
1377 (process_replaceable): Use ter_is_replaceable_p.
1378 (find_replaceable_in_bb): Use ter_is_replaceable_p.
1379 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
1380 newly refactored ssa_is_replaceable_p.
1381 * cfgexpand.c: Include tree-outof-ssa.h.
1382 * ssaexpand.h: Delete.
1384 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
1386 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
1387 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
1389 (create_gimple_tmp): Delete.
1390 (get_expr_type, build_assign, build_type_cast): Move to...
1391 * gimple-builder.c: New File.
1392 (get_expr_type): Relocate from gimple.c.
1393 (build_assign, build_type_cast): Change to only create ssanames.
1394 * gimple.h: Move prototypes to...
1395 * gimple-builder.h: New File. Here.
1396 * tree-ssa.h: And here.
1397 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
1398 count_uses_and_derefs): Relocate from gimple.c.
1399 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
1400 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
1401 * tree-ssa-math-opts (execute_cse_reciprocals): Use
1402 gimple_replace_ssa_lhs.
1403 * asan.c: Include gimple-builder.h.
1404 * Makefile.in: Add gimple-builder.o.
1406 2013-09-26 Richard Biener <rguenther@suse.de>
1408 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
1409 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
1410 (loe_visit_block): Use gcc_checking_assert.
1411 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
1412 gimple_assign_ssa_name_copy_p.
1413 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
1416 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
1418 * config/rs6000/t-rs6000 (rs6000.o): Remove.
1419 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
1421 2013-09-26 Richard Biener <rguenther@suse.de>
1423 PR tree-optimization/58539
1424 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
1425 the fact that debug statements are not taking part in loop-closed
1428 2013-09-26 Nick Clifton <nickc@redhat.com>
1430 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
1431 time warning message.
1432 (msp430_print_operand_raw): Delete unused letter parameter.
1433 (TARGET_PRINT_OPERAND_ADDRESS): Define.
1434 (msp430_print_operand_address): New function.
1435 (msp430_print_operand): Move address printing code from here to
1437 * config/msp430/msp430.md (movsipsi2): Add comment in generated
1439 (zero_extendpsisi2): Likewise.
1440 (extendpsisi2): New pattern.
1441 (andneghi3): New pattern.
1443 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
1445 * config/aarch64/aarch64.opt (mlra): New option.
1446 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
1447 (TARGET_LRA_P): Define.
1449 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
1451 * expr.c (expand_assignment): Remove obsolete comment.
1453 2013-09-25 Jeff Law <law@redhat.com>
1455 * tree-flow.h (thread_through_all_blocks): Prototype moved into
1456 tree-ssa-threadupdate.h.
1457 (register_jump_thread): Similarly.
1458 * tree-ssa-threadupdate.h: New header file.
1459 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
1460 * tree-vrp.c: Likewise.
1461 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
1462 (thread_around_empty_blocks): Change type of path vector argument to
1463 an edge,type pair from just an edge. Initialize both elements when
1464 appending to a jump threading path. Tweak references to elements
1466 (thread_across_edge): Similarly. Release memory for the elements
1468 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
1469 (dump_jump_thread_path): New function broken out from
1470 register_jump_thread.
1471 (register_jump_thread): Use dump_jump_thread_path. Change type of
1472 path vector entries. Search the path for NULL edges and dump
1473 the path if one is found. Tweak the conversion of path to 3-edge
1474 form to use the block copy type information embedded in the path.
1476 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
1478 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
1480 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
1481 Vladimir Makarov <vmakarov@redhat.com>
1483 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
1484 from the least significant bit.
1485 (strip_address_mutations): Add bitfield operations handling.
1486 (must_be_index_p): Add shifting and rotate operations handling.
1487 (set_address_base): Use must_be_base_p predicate.
1488 (set_address_index): Use must_be_index_p predicate.
1490 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
1491 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1492 Sergey Lega <sergey.s.lega@intel.com>
1493 Anna Tikhonova <anna.tikhonova@intel.com>
1494 Ilya Tocar <ilya.tocar@intel.com>
1495 Andrey Turetskiy <andrey.turetskiy@intel.com>
1496 Ilya Verbin <ilya.verbin@intel.com>
1497 Kirill Yukhin <kirill.yukhin@intel.com>
1498 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1500 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
1502 (ix86_expand_vector_move_misalign): Support new unaligned load and
1503 stores and use new names.
1504 (CODE_FOR_sse2_storedqu): Rename to ...
1505 (CODE_FOR_sse2_storedquv16qi): ... this.
1506 (CODE_FOR_sse2_loaddqu): Rename to ...
1507 (CODE_FOR_sse2_loaddquv16qi): ... this.
1508 (CODE_FOR_avx_loaddqu256): Rename to ...
1509 (CODE_FOR_avx_loaddquv32qi): ... this.
1510 (CODE_FOR_avx_storedqu256): Rename to ...
1511 (CODE_FOR_avx_storedquv32qi): ... this.
1512 * config/i386/i386.md (fpint_logic): New.
1513 * config/i386/sse.md (VMOVE): Extend for AVX512.
1517 (VI_UNALIGNED_LOADSTORE): Ditto.
1518 (sse2_avx_avx512f): Ditto.
1519 (sse2_avx2): Extend for AVX512.
1520 (sse4_1_avx2): Ditto.
1521 (avx2_avx512f): New.
1522 (sse): Extend for AVX512.
1525 (avxsizesuffix): Ditto.
1526 (sseintvecmode): Ditto.
1528 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
1529 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
1530 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
1531 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
1532 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
1533 (<sse2_avx_avx512f>_storedqu<mode): ... this.
1534 (<sse>_movnt<mode>): Replace constraint "x" with "v".
1535 (STORENT_MODE): Extend for AVX512.
1536 (*absneg<mode>2): Replace constraint "x" with "v".
1537 (*mul<mode>3): Ditto.
1538 (*ieee_smin<mode>3): Ditto.
1539 (*ieee_smax<mode>3): Ditto.
1540 (avx_cmp<mode>3): Replace VF with VF_128_256.
1541 (*<sse>_maskcmp<mode>3_comm): Ditto.
1542 (<sse>_maskcmp<mode>3): Ditto.
1543 (<sse>_andnot<mode>3): Extend for AVX512.
1544 (<code><mode>3, anylogic): Replace VF with VF_128_256.
1545 (<code><mode>3, fpint_logic): New.
1546 (*<code><mode>3): Extend for AVX512.
1547 (avx512flogicsuff): New.
1548 (avx512f_<logic><mode>): Ditto.
1549 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
1551 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
1552 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
1553 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
1554 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
1555 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
1556 (xop_vpermil2<mode>3): Ditto.
1557 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
1558 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
1559 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
1560 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
1562 2013-09-25 Tom Tromey <tromey@redhat.com>
1564 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
1565 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
1566 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
1567 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
1568 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
1569 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
1570 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
1571 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
1572 (GRAPHITE_HTAB_H): Remove.
1574 2013-09-25 Tom Tromey <tromey@redhat.com>
1576 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
1578 2013-09-25 Tom Tromey <tromey@redhat.com>
1580 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
1582 2013-09-25 Tom Tromey <tromey@redhat.com>
1584 * config/i386/t-i386 (i386.o): Remove.
1585 (i386-c.o): Use COMPILE and POSTCOMPILE.
1587 2013-09-25 Tom Tromey <tromey@redhat.com>
1589 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
1591 2013-09-25 Tom Tromey <tromey@redhat.com>
1593 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
1594 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
1595 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
1596 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
1597 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
1598 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
1599 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
1600 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
1601 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
1602 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
1603 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
1604 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
1605 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
1606 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
1607 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
1608 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
1609 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
1610 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
1611 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
1612 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
1613 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
1614 (tree-ssa-pre.o, tree-ssa-sccvn.o)
1615 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
1616 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
1617 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
1618 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
1619 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
1620 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
1621 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
1622 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
1623 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
1624 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
1625 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
1626 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
1627 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
1628 (graphite-clast-to-gimple.o, graphite-dependences.o)
1629 (graphite-interchange.o, graphite-poly.o)
1630 (graphite-scop-detection.o, graphite-sese-to-poly.o)
1631 (graphite-optimize-isl.o, tree-vect-loop.o)
1632 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
1633 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
1634 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
1635 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
1636 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
1637 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
1638 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
1639 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
1640 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
1641 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
1642 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
1643 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
1644 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
1645 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
1646 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
1647 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
1648 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
1649 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
1650 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
1651 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
1652 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
1653 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
1654 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
1655 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
1656 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
1657 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
1658 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
1659 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
1660 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
1661 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
1662 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
1663 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
1664 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
1665 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
1666 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
1667 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
1668 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
1669 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
1670 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
1671 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
1672 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
1673 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
1674 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
1675 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
1676 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
1677 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
1678 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
1679 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
1680 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
1681 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
1682 (gcov-dump.o): Remove.
1683 (default-c.o): Use COMPILE and POSTCOMPILE.
1684 (CFLAGS-gcc.o): New variable.
1685 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
1687 2013-09-25 Tom Tromey <tromey@redhat.com>
1689 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
1690 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
1691 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
1692 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
1693 (c-family/c-pragma.o, c-family/c-pretty-print.o)
1694 (c-family/c-semantics.o, c-family/c-ada-spec.o)
1695 (c-family/array-notation-common.o, c-family/stub-objc.o)
1696 (c-family/c-ubsan.o): Remove.
1698 2013-09-25 Tom Tromey <tromey@redhat.com>
1700 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
1702 2013-09-25 Tom Tromey <tromey@redhat.com>
1704 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
1705 to add -DENABLE_SHARED_LIBGCC.
1706 (gcc.o): Don't use subshell.
1708 2013-09-25 Tom Tromey <tromey@redhat.com>
1710 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
1711 * configure.ac: Don't invoke AM_PROG_CC_C_O.
1712 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
1713 * configure, config.in: Rebuild.
1715 2013-09-25 Tom Tromey <tromey@redhat.com>
1717 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
1718 (COMPILE, POSTCOMPILE): New variables.
1719 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
1720 (DEPFILES): New variable.
1721 Include ".Po" files.
1722 * configure.ac: Add checks for dependency checking.
1723 * configure, aclocal.m4: Regenerate.
1725 2013-09-25 Tom Tromey <tromey@redhat.com>
1727 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
1728 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
1730 2013-09-25 Tom Tromey <tromey@redhat.com>
1732 * Makefile.in (generated_files): Add options.h,
1733 target-hooks-def.h, insn-opinit.h,
1734 common/common-target-hooks-def.h, pass-instances.def,
1735 c-family/c-target-hooks-def.h.
1737 2013-09-25 Jeff Law <law@redhat.com>
1739 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
1740 than foo[foo.length () - 1] to access last member in a vec.
1741 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
1743 2013-09-25 Richard Biener <rguenther@suse.de>
1746 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
1748 2013-09-25 Jan Hubicka <jh@suse.cz>
1750 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
1753 2013-09-25 Marek Polacek <polacek@redhat.com>
1756 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
1757 when determining the type name.
1759 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
1761 * config/sh/sh.md: Fix formatting.
1763 2013-09-24 Xinliang David Li <davidxl@google.com>
1765 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
1766 max peel iterations parameter.
1767 * param.def: New parameter.
1768 * doc/invoke.texi: Document New parameter.
1770 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
1772 * gimple-pretty-print.c: Various whitespace tweaks.
1773 * tree-core.h: Likewise.
1774 * tree-pretty-print.c: Likewise.
1775 * tree-ssa-alias.c: Likewise.
1776 * tree-ssa-copy.c: Likewise.
1777 * tree-ssanames.c: Likewise.
1778 * tree-ssanames.h: Likewise.
1779 * tree-vrp.c: Likewise.
1781 2013-09-24 Alan Modra <amodra@gmail.com>
1785 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
1786 for output operands that disallow regs. Don't use EXPAND_WRITE on
1789 2013-09-24 Richard Biener <rguenther@suse.de>
1792 * tree.c (reference_alias_ptr_type): Move ...
1793 * alias.c (reference_alias_ptr_type): ... here and implement
1794 in terms of the new reference_alias_ptr_type_1.
1795 (ref_all_alias_ptr_type_p): New helper.
1796 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
1797 use ref_all_alias_ptr_type_p.
1798 (get_deref_alias_set): Add flag_strict_aliasing check here.
1799 (reference_alias_ptr_type_1): New function, split out from ...
1800 (get_alias_set): ... here.
1801 (alias_ptr_types_compatible_p): New function.
1802 * alias.h (reference_alias_ptr_type): Declare.
1803 (alias_ptr_types_compatible_p): Likewise.
1804 * tree.h (reference_alias_ptr_type): Remove.
1805 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
1806 to compare MEM_REF alias types.
1808 2013-09-24 Richard Biener <rguenther@suse.de>
1810 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
1812 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
1814 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
1815 reload helper function arrays into a single array reg_addr.
1816 (reload_fpr_gpr): Likewise.
1817 (reload_gpr_vsx): Likewise.
1818 (reload_vsx_gpr): Likewise.
1819 (struct rs6000_reg_addr): Likewise.
1820 (reg_addr): Likewise.
1821 (rs6000_debug_reg_global): Change rs6000_vector_reload,
1822 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
1823 (rs6000_init_hard_regno_mode_ok): Likewise.
1824 (rs6000_secondary_reload_direct_move): Likewise.
1825 (rs6000_secondary_reload): Likewise.
1827 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
1828 constraints: wu, ww, and wy. Repurpose wv constraint added during
1829 power8 changes. Put wg constraint in alphabetical order.
1831 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
1832 for future work to add ISA 2.07 VSX single precision support.
1833 (-mvsx-scalar-double): Change default from -1 to 1, update
1834 documentation comment.
1835 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
1836 (-mupper-regs-df): New debug switch to control whether DF values
1837 can go in the traditional Altivec registers.
1838 (-mupper-regs-sf): New debug switch to control whether SF values
1839 can go in the traditional Altivec registers.
1841 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
1843 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
1844 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
1845 Add new constraints, wu/ww/wy. Repurpose wv constraint.
1846 (rs6000_debug_legitimate_address_p): Print if we are running
1847 before, during, or after reload.
1848 (rs6000_secondary_reload): Add a comment.
1849 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
1851 * config/rs6000/constraints.md (wa constraint): Sort w<x>
1852 constraints. Update documentation string.
1853 (wd constraint): Likewise.
1854 (wf constraint): Likewise.
1855 (wg constraint): Likewise.
1856 (wn constraint): Likewise.
1857 (ws constraint): Likewise.
1858 (wt constraint): Likewise.
1859 (wx constraint): Likewise.
1860 (wz constraint): Likewise.
1861 (wu constraint): New constraint for ISA 2.07 SFmode scalar
1863 (ww constraint): Likewise.
1864 (wy constraint): Likewise.
1865 (wv constraint): Repurpose ISA 2.07 constraint that we not used in
1866 the previous submissions.
1867 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
1869 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
1871 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
1872 (REG_BR_PROB): Say that the probability is stored in an int_list.
1873 * reg-notes.def: Update commentary to mention INT_LIST.
1874 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
1875 (INT_LIST): New rtx.
1876 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
1877 * rtlanal.c (int_reg_note_p): New function.
1878 (alloc_reg_note): Assert that the note does not have an int argument.
1879 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
1880 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
1881 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
1882 rather than an INSN_LIST. Handle INT_LIST.
1883 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
1884 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
1885 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
1886 Manipulate them as ints rather than rtxes.
1887 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
1888 * regmove.c (copy_src_to_dest): Likewise.
1889 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
1891 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
1892 into the cases that need it.
1893 * config/arm/arm.c (arm_unwind_emit): Likewise.
1895 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
1896 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
1897 * loop-doloop.c (add_test, doloop_modify): Likewise.
1898 * loop-unswitch.c (compare_and_jump_seq): Likewise.
1899 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
1900 * predict.c (combine_predictions_for_insn): Likewise.
1901 * print-rtl.c (print_rtx): Handle INT_LIST.
1902 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
1903 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
1904 * config/arm/arm.c (emit_unlikely_jump): Likewise.
1905 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
1906 (ix86_split_fp_branch, predict_jump): Likewise.
1907 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
1908 * config/sh/sh.c (expand_cbranchsi4): Likewise.
1909 * config/spu/spu.c (ea_load_store_inline): Likewise.
1911 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
1912 value of a REG_BR_PROB note.
1913 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
1914 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
1915 * emit-rtl.c (try_split): Likewise.
1916 * predict.c (br_prob_note_reliable_p): Likewise.
1917 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
1918 * reorg.c (mostly_true_jump): Likewise.
1919 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
1920 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
1921 * config/i386/i386.c (ix86_print_operand): Likewise.
1922 * config/ia64/ia64.c (ia64_print_operand): Likewise.
1923 * config/mmix/mmix.c (mmix_print_operand): Likewise.
1924 * config/rs6000/rs6000.c (output_cbranch): Likewise.
1925 * config/s390/s390.c (s390_expand_tbegin): Likewise.
1926 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
1927 * config/sparc/sparc.c (output_cbranch): Likewise.
1928 * config/spu/spu.c (get_branch_target): Likewise.
1929 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
1930 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
1932 2013-09-23 Jan Hubicka <jh@suse.cz>
1934 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
1936 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
1937 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
1939 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
1940 (gimple_fold_call): Dump inconsistent devirtualizations; add
1941 sanity check for type based devirtualizations.
1942 * ipa-prop.c: Include ipa-utils.h
1943 (ipa_intraprocedural_devirtualization): Add sanity check.
1944 (try_make_edge_direct_virtual_call): Likewise.
1946 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
1948 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
1949 assignment statements.
1951 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
1953 * gimple-pretty-print.c (dump_ssaname_info): New function.
1954 (dump_gimple_phi): Call it.
1955 (pp_gimple_stmt_1): Likewise.
1956 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
1957 (range_info_def): Declare.
1958 * tree-pretty-print.c (pp_double_int): New function.
1959 (dump_generic_node): Call it.
1960 * tree-pretty-print.h (pp_double_int): Declare.
1961 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
1962 * tree-ssanames.h (range_info_def): New structure.
1963 (value_range_type): Move definition here.
1964 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
1966 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
1968 (set_range_info): New function.
1969 (get_range_info): Likewise.
1970 (duplicate_ssa_name_range_info): Likewise.
1971 (duplicate_ssa_name_fn): Check pointer type and call
1972 duplicate_ssa_name_range_info.
1973 * tree-ssa-copy.c (fini_copy_prop): Likewise.
1974 * tree-vrp.c (value_range_type): Remove definition, now in
1976 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
1977 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
1978 (SSA_NAME_RANGE_INFO): New macro.
1980 2013-09-23 Richard Biener <rguenther@suse.de>
1982 PR tree-optimization/58464
1983 * tree-ssa-pre.c (phi_trans_lookup): Remove.
1984 (phi_trans_add): Change to add conditionally on being not present.
1985 (phi_translate_1): Remove recursion detection here.
1986 (phi_translate): Pre-seed the cache with NULL to catch
1987 recursion here in a more generic way.
1988 (bitmap_find_leader): Adjust comment.
1989 (get_representative_for): Dump value-numbers.
1990 (create_expression_by_pieces): Likewise.
1991 (insert_into_preds_of_block): Likewise.
1993 2013-09-23 Christian Bruel <christian.bruel@st.com>
1996 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
1997 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
1999 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
2002 2013-09-20 Renlin Li <renlin.li@arm.com>
2004 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
2006 (aarch64_expand_epilogue): Likewise.
2007 (aarch64_legitimize_reload_address): Likewise.
2009 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
2011 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
2012 NULL_TREE before pushing them onto the vector. Likewise for labels.
2014 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
2016 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
2017 comparison operators when -fno-trapping-math is in effect.
2018 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
2019 comparison operators in TFmode and assert that unsupported operators
2021 (ia64_print_operand): Likewise.
2023 2013-09-21 Jan Hubicka <jh@suse.cz>
2025 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
2026 (sse_typeless_stores): Enable for core
2027 (sse_load0_by_pxor): Likewise.
2028 (four_jump_limit): Disable for core.
2029 (pad_returns): Likewise.
2030 (avoid_vector_decode): Likewise.
2031 (fuse_cmp_and_branch): Enable for cores.
2032 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
2034 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
2037 * config/pa/pa.c (pa_option_override): Disable auto increment and
2038 decrement instructions until reload is completed.
2040 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
2041 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
2042 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
2044 2013-09-20 DJ Delorie <dj@redhat.com>
2045 Nick Clifton <nickc@redhat.com>
2047 * config/rl78/rl78.c: Various whitespace and comment tweaks.
2048 (need_to_save): Save bank 0 on interrupts.
2049 (characterize_address): Strip far address wrappers.
2050 (rl78_as_legitimate_address): Likewise.
2051 (transcode_memory_rtx): Likewise.
2052 (rl78_peep_movhi_p): Disable this peephole after devirt.
2053 (rl78_propogate_register_origins): Forget all origins when a
2055 * config/rl78/rl78-virt.md: Various whitespace tweaks.
2056 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
2058 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
2059 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
2060 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
2062 * config/rl78/constraints.md: Various whitespace and paren tweaks.
2064 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
2066 * config/pa/pa.md: In "scc" insn patterns, change output template to
2067 handle const0_rtx in reg_or_0_operand operands.
2069 2013-09-20 Martin Husemann <martin@NetBSD.org>
2072 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
2073 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
2075 2013-09-20 Richard Biener <rguenther@suse.de>
2078 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
2079 remembering SSA name version and block index.
2080 (new_scev_info_str): Adjust.
2081 (hash_scev_info): Likewise. Also hash the block index.
2082 (eq_scev_info): Adjust.
2083 (find_var_scev_info): Likewise.
2084 (struct instantiate_cache_entry): Remove.
2085 (struct instantiate_cache_type): Use a htab to map name, block
2087 (instantiate_cache_type::~instantiate_cache_type): Adjust.
2088 (get_instantiated_value_entry): Likewise.
2089 (hash_idx_scev_info, eq_idx_scev_info): New functions.
2090 (instantiate_scev_name): Adjust.
2092 2013-09-20 Jeff Law <law@redhat.com>
2094 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
2096 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
2098 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
2099 Call aarch64_simd_expand_args to update op[argc].
2101 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
2103 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
2106 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
2108 * gengtype.c (file_rules): Added rule for *.cc files.
2109 (get_output_file_with_visibility): Give fatal message when no
2112 2013-09-20 Renlin Li <renlin.li@arm.com>
2114 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
2115 (aarch64_expand_epilogue): Likewise.
2116 (aarch64_legitimize_reload_address): Likewise.
2118 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
2121 * expr.c (expand_assignment): Remove misalignp code path.
2123 2013-09-20 Marek Polacek <polacek@redhat.com>
2126 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
2127 TYPE_PRECISION. Add asserts.
2129 2013-09-20 Richard Biener <rguenther@suse.de>
2131 PR tree-optimization/58453
2132 * tree-loop-distribution.c (distribute_loop): Apply the cost
2133 model for -ftree-loop-distribute-patterns, too.
2135 2013-09-20 Richard Biener <rguenther@suse.de>
2138 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
2139 make type comparison less strict.
2141 2013-09-20 Alan Modra <amodra@gmail.com>
2143 * configure: Regenerate.
2144 * aclocal.m4: Regenerate.
2146 2013-09-20 Marek Polacek <polacek@redhat.com>
2149 * doc/extend.texi: Document that attribute used is meant to be used
2150 on variables with static storage duration.
2152 2013-09-19 Jakub Jelinek <jakub@redhat.com>
2154 PR tree-optimization/58472
2155 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
2156 simd_lane_access set inv_p = false.
2157 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
2158 the simduid magic VAR_DECL.
2160 2013-09-19 Jan Hubicka <jh@suse.cz>
2162 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
2164 2013-09-17 Jeff Law <law@redhat.com>
2166 * tree-ssa-dom.c (record_temporary_equivalences): New function
2167 split out of dom_opt_dom_walker::after_dom_children.
2168 (dom_opt_dom_walker::thread_across_edge): Move common code
2169 in here from dom_opt_dom_walker::after_dom_children.
2170 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
2172 2013-09-19 Jan Hubicka <jh@suse.cz>
2174 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
2175 (TARGET_GENERIC): Use PROCESOR_GENERIC
2176 (enum processor_type): Unify generic32 and 64.
2177 * i386.md (cpu): Likewise.
2178 * x86-tune.def (use_leave): Enable for generic32.
2179 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
2180 * athlon.md: Change generic64 to generic in all occurences.
2181 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
2182 (ix86_target_macros_internal): Likewise.
2183 * driver-i386.c (host_detect_local_cpu): Likewise.
2184 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
2186 (generic_memcpy, generic_memset, generic_cost): This one.
2187 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
2188 (m_GENERIC32, m_GENERIC64): Remove.
2189 (m_GENERIC): Turn into one flag.
2190 (processor_target): Unify generic tunnings.
2191 (ix86_option_override_internal): Replace generic32/64 by generic.
2192 (ix86_issue_rate): Likewise.
2193 (ix86_adjust_cost): Likewise.
2195 2013-09-19 Jan Hubicka <jh@suse.cz>
2197 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
2198 of speculative flag.
2200 2013-09-19 Jakub Jelinek <jakub@redhat.com>
2202 * omp-low.c (expand_omp_sections): Always pass len - 1 to
2203 GOMP_sections_start, even if !exit_reachable.
2205 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
2207 * lra-constraints.c (need_for_all_save_p): Use macro
2208 HARD_REGNO_CALL_PART_CLOBBERED.
2209 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
2210 set up pseudo conflict hard regs.
2212 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
2215 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
2218 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
2220 PR rtl-optimization/58438
2221 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
2222 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
2223 from previous subpasses.
2225 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
2227 * arm.c (arm_get_frame_offsets): Validate architecture supports
2228 LDRD/STRD before accepting the tuning preference.
2229 (arm_expand_prologue): Likewise.
2230 (arm_expand_epilogue): Likewise.
2232 2013-09-18 Richard Biener <rguenther@suse.de>
2234 PR tree-optimization/58417
2235 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
2236 have chrecs with symbols defined in the loop as operands.
2237 (chrec_fold_multiply): Likewise.
2238 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
2239 parameters before folding binary operations.
2240 (struct instantiate_cache_entry_hasher): Remove.
2241 (struct instantiate_cache_type): Use a pointer-map.
2242 (instantiate_cache_type::instantiate_cache_type): New function.
2243 (instantiate_cache_type::get): Likewise.
2244 (instantiate_cache_type::set): Likewise.
2245 (instantiate_cache_type::~instantiate_cache_type): Adjust.
2246 (get_instantiated_value_entry): Likewise.
2247 (global_cache): New global.
2248 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
2249 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
2250 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
2251 (instantiate_scev_name): Adjust.
2252 (instantiate_scev): Construct global instead of local cache.
2253 (resolve_mixers): Likewise.
2255 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
2256 Paolo Carlini <paolo.carlini@oracle.com>
2259 * doc/implement-cxx.texi: Fix references to the C++ standards.
2261 2013-09-18 Jakub Jelinek <jakub@redhat.com>
2263 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
2264 * tree-vect-data-refs.c (vect_analyze_data_refs): For
2265 simd_lane_access drs, update also DR_ALIGNED_TO.
2267 2013-09-18 Marek Polacek <polacek@redhat.com>
2270 * doc/extend.texi: Document no_sanitize_undefined attribute.
2271 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
2272 no_sanitize_undefined attribute.
2274 2013-09-18 Nick Clifton <nickc@redhat.com>
2276 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
2277 (ASM_DECLARE_FUNCTION_NAME): Define.
2279 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
2281 * compare-elim.c (find_comparison_dom_walker): New class
2282 (find_comparisons_in_bb): Rename to
2283 find_comparison_dom_walker::before_dom_children
2284 (find_comparisons): Adjust
2285 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
2287 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
2288 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
2289 (init_walk_dominator_tree): Remove declaration.
2290 (fini_walk_dominator_tree): Remove declaration.
2291 * fwprop.c (single_def_use_dom_walker): New class
2292 (single_def_use_enter_block): Convert to
2293 single_def_use_dom_walker::before_dom_children.
2294 (single_def_use_leave_block): Convert to
2295 single_def_use_dom_walker::after_dom_children.
2296 (build_single_def_use_links): Adjust.
2297 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
2299 (find_candidates_in_block): Convert to
2300 find_candidates_dom_walker::before_dom_children.
2301 (execute_strength_reduction): Adjust.
2302 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
2303 (sese_dom_walker): New class.
2304 (sese_dom_walker::sese_dom_walker): New constructor.
2305 (sese_dom_walker::~sese_dom_walker): New destructor.
2306 (build_sese_conditions_before): Convert to
2307 sese_dom_walker::before_dom_children.
2308 (build_sese_conditions_after): Convert to
2309 sese_dom_walker::after_dom_children.
2310 (build_poly_scop): Adjust
2311 * tree-into-ssa.c (rewrite_dom_walker): New class
2312 (rewrite_enter_block): Convert to
2313 rewrite_dom_walker::before_dom_children.
2314 (rewrite_leave_block): Convert to
2315 rewrite_dom_walker::after_dom_children.
2316 (rewrite_update_dom_walker): New class.
2317 (rewrite_update_enter_block): Convert to
2318 rewrite_update_dom_walker::before_dom_children.
2319 (rewrite_update_leave_block): Convert to
2320 rewrite_update_dom_walker::after_dom_children.
2321 (rewrite_blocks, rewrite_into_ssa): Adjust.
2322 (mark_def_dom_walker): New class.
2323 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
2324 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
2325 (mark_def_sites_blocks): Convert to
2326 mark_def_dom_walker::before_dom_children.
2327 (mark_def_site_blocks): Remove.
2328 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
2329 (tree_ssa_dominator_optimize): Adjust.
2330 (dom_thread_across_edge): Convert to method
2331 dom_opt_dom_walker::thread_across_edge.
2332 (dom_opt_enter_block): Convert to member function
2333 dom_opt_dom_walker::before_dom_children.
2334 (dom_opt_leave_block): Convert to member function
2335 dom_opt_dom_walker::after_dom_children.
2336 * tree-ssa-dse.c (dse_dom_walker): New class.
2337 (dse_enter_block): Convert to member function
2338 dse_dom_walker::before_dom_children.
2339 (tree_ssa_dse): Adjust.
2340 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
2341 (determine_invariantness_stmt): Convert to method
2342 invariantness_dom_walker::before_dom_children.
2343 (determine_invariantness): Remove
2344 (move_computations_dom_walker): New class.
2345 (move_computations_stmt): Convert to method
2346 move_computations_dom_walker::before_dom_children.
2347 (move_computations, tree_ssa_lim): Adjust.
2348 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
2349 (nt_init_block): Make method
2350 notrappping_dom_walker::before_dom_children.
2351 (nt_fini_block): Make
2352 method nontrapping_dom_walker::after_dom_children.
2353 (get_non_trapping): Adjust.
2354 * tree-ssa-pre.c (eliminate_dom_walker): New class.
2355 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
2356 (eliminate_leave_block): Make method.
2357 eliminate_dom_walker::after_dom_children.
2359 * tree-ssa-strlen.c (strlen_dom_walker): New class.
2360 (strlen_enter_block): Make method
2361 strlen_dom_walker::before_dom_children.
2362 (strlen_leave_block): Make
2363 method strlen_dom_walker::after_dom_children.
2364 (tree_ssa_strlen): Adjust.
2365 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
2366 (tree_ssa_uncprop): Adjust.
2367 (uncprop_leave_block): Make method
2368 uncprop_dom_walker::after_dom_children.
2369 (uncprop_leave_block): Make method
2370 uncprop_dom_walker::before_dom_children.
2372 2013-09-18 Bin Cheng <bin.cheng@arm.com>
2374 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
2375 branch in same basic block. Check both src and dest of the move insn.
2377 2013-09-17 Nick Clifton <nickc@redhat.com>
2379 * config/rl78/rl78-real.md (bf): New pattern.
2381 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
2382 (rl78_print_operand): Do not put a # before a %B.
2383 * config/rl78/rl78.opt: Tweak doc strings.
2385 2013-09-17 DJ Delorie <dj@redhat.com>
2387 * config/rl78/constraints.md (Wcv): Allow up to $r31.
2388 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
2389 (rl78_option_override): Likewise, if -mallregs.
2390 (is_virtual_register): Likewise.
2391 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
2392 (REGNO_OK_FOR_BASE_P): Likewise.
2393 * config/rl78/rl78.opt (-mallregs): New.
2395 2013-09-17 Nick Clifton <nickc@redhat.com>
2397 * config/rl78/rl78.c (need_to_save): Change return type to bool.
2398 For interrupt functions: save all call clobbered registers if the
2399 interrupt handler is not a leaf function.
2400 (rl78_expand_prologue): Always recompute the frame information.
2401 For interrupt functions: only select bank 0 if one of the bank 0
2402 registers is going to be psuhed.
2404 2013-09-17 DJ Delorie <dj@redhat.com>
2406 * config/rl78/constraints.md: For each W* constraint, rename to C*
2407 and create a W* constraint that checks for an optional ES: prefix
2409 * config/rl78/rl78.md (UNS_ES_ADDR): New.
2410 (es_addr): New. Used to wrap far addresses.
2411 * config/rl78/rl78-protos.h (rl78_es_addr): New.
2412 (rl78_es_base): New.
2413 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
2414 wrapped far addresses.
2415 (rl78_print_operand_1): Unwrap far addresses before processing.
2416 (rl78_lo16): Wrap far addresses in unspecs.
2417 (rl78_es_addr): New.
2418 (rl78_es_base): New.
2419 (insn_ok_now): Check for not-yet-wrapped far addresses.
2420 (transcode_memory_rtx): Properly re-wrap far addresses.
2422 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
2424 * config/sparc/t-rtems: Add leon3 multilibs.
2426 2013-09-17 Cong Hou <congh@google.com>
2428 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
2429 when checking the dot production pattern. The type of rhs operand
2430 of multiply is now checked correctly.
2432 2013-09-17 Jeff Law <law@redhat.com>
2434 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
2435 edge implied equivalences into successor phis.
2436 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
2437 here from tree-ssa-threadedge.c.
2438 (mark_threaded_blocks): When threading through a joiner, if both
2439 successors of the joiner's clone reach the same block, verify the
2440 PHI arguments are equal. If not, cancel the jump threading request.
2441 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
2442 tree-ssa-threadupdate.c
2443 (thread_across_edge): Don't check PHI argument equality when
2444 threading through joiner block here.
2446 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
2448 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
2449 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
2450 (warn_uninit, warn_uninitialized_vars,
2451 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
2452 Move to tree-ssa-uninit.c.
2453 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
2454 (has_undefined_value_p): New. Pass dependant parts of
2455 ssa_undefined_value_p.
2456 (uninit_undefined_value_p): Use has_undefined_value_p.
2457 (warn_uninit, warn_uninitialized_vars,
2458 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
2459 Move from tree-ssa.c.
2460 * tree-ssa.h: Adjust prototypes.
2462 2013-09-17 Jan Hubicka <jh@suse.cz>
2465 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
2466 * ipa-inline.c (can_inline_edge_p): Do not downgrade
2467 FUNCTION_NOT_OPTIMIZED.
2468 * ipa-inline-analysis.c (compute_inline_parameters): Function
2469 not optimized is not inlinable unless it is alwaysinline.
2470 (inline_analyze_function): Force calls in not optimized
2471 function not inlinable.
2473 2013-09-17 Jan Hubicka <jh@suse.cz>
2476 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
2478 * ipa.c (function_and_variable_visibility): Likewise.
2479 * ipa-profile.c (ipa_profile): Likewise.
2481 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
2484 * cgraph.c (cgraph_function_body_availability): Check for ifunc
2485 attribute, and don't inline the resolver in this case.
2487 2013-09-17 Teresa Johnson <tejohnson@google.com>
2489 * coverage.c (get_coverage_counts): Add missing newline.
2491 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2493 PR tree-optimization/58088
2494 * fold-const.c (mask_with_trailing_zeros): New function.
2495 (fold_binary_loc): Make sure we don't recurse infinitely
2496 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
2497 Use mask_with_trailing_zeros where appropriate.
2499 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
2501 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
2502 of 'prev' var to get better distance estimation.
2504 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
2506 * tree-inline.h (struct copy_body_data): Add transform_parameter.
2507 * tree-inline.c (is_parameter_of): New predicate.
2508 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
2509 a parameter has been remapped.
2510 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
2511 (optimize_inline_calls): Initialize transform_parameter.
2512 (copy_gimple_seq_and_replace_locals): Likewise.
2513 (tree_function_versioning): Likewise.
2514 (maybe_inline_call_in_expr): Likewise.
2516 2013-09-17 Nick Clifton <nickc@redhat.com>
2518 * config/msp430/msp430-protos.h: Add prototypes for new functions.
2519 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
2521 (is_attr_func): New function.
2522 (msp430_is_interrupt_func): New function.
2523 (is_naked_func): New function.
2524 (is_reentrant_func): New function.
2525 (is_critical_func): New function.
2526 (msp430_start_function): Add annotations for function attributes.
2527 (msp430_attr): New function.
2528 (msp430_attribute_table): New.
2529 (msp430_function_section): New function.
2530 (TARGET_ASM_FUNCTION_SECTION): Define.
2531 (msp430_builtin): New enum.
2532 (msp430_init_builtins): New function.
2533 (msp430_builtin_devl): New function.
2534 (msp430_expand_builtin): New function.
2535 (TARGET_INIT_BUILTINS): Define.
2536 (TARGET_EXPAND_BUILTINS): Define.
2537 (TARGET_BUILTIN_DECL): Define.
2538 (msp430_expand_prologue): Add support for naked, interrupt,
2539 critical and reentrant functions.
2540 (msp430_expand_epilogue): Likewise.
2541 (msp430_print_operand): Handle 'O' character.
2542 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
2544 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
2545 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
2546 (pushm): Use a 'n' rather than an 'i' constraint.
2547 (msp_return): Add generation of the interrupt return instruction.
2548 (disable_interrupts): New pattern.
2549 (enable_interrupts): New pattern.
2550 (push_intr_state): New pattern.
2551 (pop_intr_state): New pattern.
2552 (bic_SR): New pattern.
2553 (bis_SR): New pattern.
2554 * doc/extend.texi: Document MSP430 function attributes and builtin
2557 2013-09-17 Richard Biener <rguenther@suse.de>
2559 PR tree-optimization/58432
2560 * tree-loop-distribution.c (tree_loop_distribution): Also
2561 scan PHIs for outside loop uses and seed a partition from them.
2563 2013-09-17 Bin Cheng <bin.cheng@arm.com>
2565 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
2566 (restructure_reference): Call backtrace_base_for_ref.
2568 2013-09-17 Alan Modra <amodra@gmail.com>
2571 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
2574 2013-09-16 DJ Delorie <dj@redhat.com>
2576 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
2577 vregs location for RL78/G10.
2578 (rl78_expand_prologue): Avoid SEL on G10.
2579 (rl78_expand_epilogue): Likewise.
2580 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
2581 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
2582 __RL78_G10__ when appropriate.
2583 (ASM_SPEC): Pass -mg10 along to the assembler.
2584 * config/rl78/rl78.md (sel_rb): Disable for G10.
2585 * config/rl78/rl78.opt: Add -mg10 option.
2586 * config/rl78/t-rl78: Add -mg10 multilib.
2588 2013-09-16 Xinliang David Li <davidxl@google.com>
2590 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
2591 * omp-low.c (omp_max_vf): Ditto.
2592 (expand_omp_simd): Ditto.
2593 * tree-vectorizer.c (vectorize_loops): Ditto.
2594 (gate_vect_slp): Ditto.
2595 (gate_increase_alignment): Ditto.
2596 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
2597 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
2598 (gate_tree_vectorize): Name change.
2599 (tree_vectorize): Ditto.
2600 (pass_vectorize::gate): Call new function.
2601 (pass_vectorize::execute): Ditto.
2602 * opts.c: O3 default setting change.
2603 (finish_options): Check new flag.
2604 * doc/invoke.texi: Document new flags.
2605 * common.opt: New flags.
2607 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
2609 * doc/tm.texi.in (Cond Exec Macros): Remove node.
2610 (Condition Code): Don't reference it.
2611 * doc/tm.texi: Regenerate.
2613 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
2616 * lra-constraints.c (undo_optional_reloads): Consider all optional
2617 reload even if it did not get a hard reg.
2619 2013-09-16 Teresa Johnson <tejohnson@google.com>
2621 * dumpfile.c (dump_loc): Remove newline emission.
2622 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
2623 emission to dump_printf_loc calls where missing.
2624 (vect_mark_for_runtime_alias_test): Ditto.
2625 (vect_analyze_data_ref_dependence): Ditto.
2626 (vect_analyze_data_ref_dependences): Ditto.
2627 (vect_slp_analyze_data_ref_dependence): Ditto.
2628 (vect_slp_analyze_data_ref_dependences): Ditto.
2629 (vect_compute_data_ref_alignment): Ditto.
2630 (vect_update_misalignment_for_peel): Ditto.
2631 (vect_verify_datarefs_alignment): Ditto.
2632 (vector_alignment_reachable_p): Ditto.
2633 (vect_get_data_access_cost): Ditto.
2634 (vect_enhance_data_refs_alignment): Ditto.
2635 (vect_find_same_alignment_drs): Ditto.
2636 (vect_analyze_data_refs_alignment): Ditto.
2637 (vect_analyze_group_access): Ditto.
2638 (vect_analyze_data_ref_access): Ditto.
2639 (vect_analyze_data_ref_accesses): Ditto.
2640 (vect_prune_runtime_alias_test_list): Ditto.
2641 (vect_analyze_data_refs): Ditto.
2642 (vect_create_addr_base_for_vector_ref): Ditto.
2643 (vect_create_data_ref_ptr): Ditto.
2644 (vect_grouped_store_supported): Ditto.
2645 (vect_grouped_load_supported): Ditto.
2646 * value-prof.c (check_counter): Ditto.
2647 (check_ic_target): Ditto.
2648 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
2649 (vect_recog_widen_mult_pattern): Ditto.
2650 (vect_recog_widen_sum_pattern): Ditto.
2651 (vect_recog_over_widening_pattern): Ditto.
2652 (vect_recog_widen_shift_pattern): Ditto.
2653 (vect_recog_rotate_pattern): Ditto.
2654 (vect_recog_vector_vector_shift_pattern): Ditto.
2655 (vect_recog_divmod_pattern): Ditto.
2656 (vect_recog_mixed_size_cond_pattern): Ditto.
2657 (vect_recog_bool_pattern): Ditto.
2658 (vect_pattern_recog_1): Ditto.
2659 (vect_pattern_recog): Ditto.
2660 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
2661 (vect_is_simple_iv_evolution): Ditto.
2662 (vect_analyze_scalar_cycles_1): Ditto.
2663 (vect_get_loop_niters): Ditto.
2664 (vect_analyze_loop_1): Ditto.
2665 (vect_analyze_loop_form): Ditto.
2666 (vect_analyze_loop_operations): Ditto.
2667 (vect_analyze_loop_2): Ditto.
2668 (vect_analyze_loop): Ditto.
2669 (report_vect_op): Ditto.
2670 (vect_is_slp_reduction): Ditto.
2671 (vect_is_simple_reduction_1): Ditto.
2672 (vect_get_known_peeling_cost): Ditto.
2673 (vect_estimate_min_profitable_iters): Ditto.
2674 (vect_model_reduction_cost): Ditto.
2675 (vect_model_induction_cost): Ditto.
2676 (get_initial_def_for_induction): Ditto.
2677 (vect_create_epilog_for_reduction): Ditto.
2678 (vectorizable_reduction): Ditto.
2679 (vectorizable_induction): Ditto.
2680 (vectorizable_live_operation): Ditto.
2681 (vect_loop_kill_debug_uses): Ditto.
2682 (vect_transform_loop): Ditto.
2683 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
2684 (vect_stmt_relevant_p): Ditto.
2685 (process_use): Ditto.
2686 (vect_mark_stmts_to_be_vectorized): Ditto.
2687 (vect_model_simple_cost): Ditto.
2688 (vect_model_promotion_demotion_cost): Ditto.
2689 (vect_model_store_cost): Ditto.
2690 (vect_get_store_cost): Ditto.
2691 (vect_model_load_cost): Ditto.
2692 (vect_get_load_cost): Ditto.
2693 (vect_init_vector_1): Ditto.
2694 (vect_get_vec_def_for_operand): Ditto.
2695 (vect_finish_stmt_generation): Ditto.
2696 (vectorizable_call): Ditto.
2697 (vectorizable_conversion): Ditto.
2698 (vectorizable_assignment): Ditto.
2699 (vectorizable_shift): Ditto.
2700 (vectorizable_operation): Ditto.
2701 (vectorizable_store): Ditto.
2702 (vectorizable_load): Ditto.
2703 (vectorizable_condition): Ditto.
2704 (vect_analyze_stmt): Ditto.
2705 (vect_transform_stmt): Ditto.
2706 (vect_is_simple_use): Ditto.
2707 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
2708 (vect_can_advance_ivs_p): Ditto.
2709 (vect_update_ivs_after_vectorizer): Ditto.
2710 (vect_do_peeling_for_loop_bound): Ditto.
2711 (vect_gen_niters_for_prolog_loop): Ditto.
2712 (vect_update_inits_of_drs): Ditto.
2713 (vect_create_cond_for_alias_checks): Ditto.
2714 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
2715 (vect_build_slp_tree_1): Ditto.
2716 (vect_supported_load_permutation_p): Ditto.
2717 (vect_analyze_slp_instance): Ditto.
2718 (vect_analyze_slp): Ditto.
2719 (vect_make_slp_decision): Ditto.
2720 (vect_detect_hybrid_slp): Ditto.
2721 (vect_bb_vectorization_profitable_p): Ditto.
2722 (vect_slp_analyze_bb_1): Ditto.
2723 (vect_update_slp_costs_according_to_vf): Ditto.
2724 (vect_get_mask_element): Ditto.
2725 (vect_transform_slp_perm_load): Ditto.
2726 (vect_schedule_slp_instance): Ditto.
2727 (vect_schedule_slp): Ditto.
2728 (vect_slp_transform_bb): Ditto.
2729 * profile.c (read_profile_edge_counts): Ditto.
2730 (compute_branch_probabilities): Ditto.
2731 * coverage.c (get_coverage_counts): Ditto.
2733 2013-09-16 Diego Novillo <dnovillo@google.com>
2735 * tree-core.h: Add missing comment lines from refactoring of tree.h.
2737 2013-09-16 Jan Hubicka <jh@suse.cz>
2739 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
2740 abstract functions; for static functions check the presence of body.
2742 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
2744 * config/aarch64/aarch64-simd-builtins.def (fma): New.
2745 * config/aarch64/aarch64-simd.md
2746 (aarch64_mla_elt<mode>): New.
2747 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
2748 (aarch64_mls_elt<mode>): Likewise.
2749 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
2750 (aarch64_fma4_elt<mode>): Likewise.
2751 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
2752 (aarch64_fma4_elt_to_128v2df): Likewise.
2753 (aarch64_fma4_elt_to_64df): Likewise.
2754 (fnma<mode>4): Likewise.
2755 (aarch64_fnma4_elt<mode>): Likewise.
2756 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
2757 (aarch64_fnma4_elt_to_128v2df): Likewise.
2758 (aarch64_fnma4_elt_to_64df): Likewise.
2759 * config/aarch64/iterators.md (VDQSF): New.
2760 * config/aarch64/arm_neon.h
2761 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
2762 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
2764 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
2766 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
2767 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
2768 (aarch64_mul3_elt_to_128df): Likewise.
2769 (aarch64_mul3_elt_to_64v2df): Likewise.
2770 * config/aarch64/iterators.md (VEL): Also handle DFmode.
2772 (VMUL_CHANGE_NLANES) Likewise.
2775 * config/aarch64/arm_neon.h
2776 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
2778 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
2780 * config/aarch64/arm_neon.h
2781 (vcvtx_high_f32_f64): Fix parameters.
2783 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2784 Uros Bizjak <ubizjak@gmail.com>
2786 * config/alpha.c: Include tree-ssanames.h.
2788 2013-09-16 Richard Biener <rguenther@suse.de>
2790 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
2791 (dot_rdg_1): Handle control_dd.
2792 (create_edge_for_control_dependence): New function.
2793 (create_rdg_edges): Add control dependences if asked for.
2794 (build_rdg): Likewise.
2795 (generate_loops_for_partition): If there are not necessary
2796 control stmts remove all their dependencies.
2797 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
2798 (distribute_loop): Pass on control dependences.
2799 (tree_loop_distribution): Compute control dependences and remove
2800 restriction on number of loop nodes.
2802 2013-09-16 Jakub Jelinek <jakub@redhat.com>
2804 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
2807 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
2809 * cse.c (try_const_anchors): Punt on CC modes.
2811 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2813 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
2815 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
2818 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
2821 2013-09-14 Iain Sandoe <iain@codesourcery.com>
2824 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
2826 (darwin_objc1_section): Likewise.
2827 (darwin_file_end): Emit Image Info section when required.
2829 2013-09-14 Jan Hubicka <jh@suse.cz>
2831 * tree-into-ssa.c (gate_into_ssa): New.
2832 (pass_data_build_ssa): Use it.
2833 * cgraph.h (expand_thunk): Update prototype.
2834 * cgraphunit.c (analyze_function): Expand thunks early.
2835 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
2836 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
2837 set lowered flag; do not add new function.
2838 (assemble_thunks_and_aliases): Update.
2839 * tree-ssa.c (gate_init_datastructures): New gate.
2840 (pass_data_init_datastructures): Use it.
2842 2013-09-14 Iain Sandoe <iain@codesourcery.com>
2845 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
2846 xmm register set described in the psABI.
2848 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
2850 * dwarf2out.c (should_emit_struct_debug): Add check
2851 for type_decl variable is not NULL.
2853 2013-09-13 Jacek Caban <jacek@codeweavers.com>
2855 * config.gcc: Use new winnt-c.c target hooks
2856 * config/t-winnt: New file
2857 * config/winnt-c.c: New file
2858 * doc/tm.texi.in: Document new hook
2859 * doc/tm.texi: Regenerated
2861 2013-09-13 Jan Hubicka <jh@suse.cz>
2864 * ipa-inline.c (check_callers): New function.
2865 (check_caller_edge): Remove.
2866 (want_inline_function_to_all_callers_p): Also permit alises that are
2868 (inline_to_all_callers): Terminate the walk when devirtualization
2869 introduce new calls.
2871 2013-09-13 Jan Hubicka <jh@suse.cz>
2873 * ipa-inline-analysis.c (struct growth_data): Add node.
2874 (do_estimate_growth_1): Fix detection of recursion.
2876 2013-09-13 Jakub Jelinek <jakub@redhat.com>
2878 PR tree-optimization/58392
2879 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
2880 to avoid shadowing of outer loop variable. If
2881 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
2882 replace_by_duplicate_decl simduid of loops that have it set and
2883 set dest_cfun->has_simduid_loops and/or
2884 dest_cfun->has_force_vect_loops.
2885 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
2886 instead of maybe_lookup_decl.
2887 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
2888 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
2889 call. Copy over force_vect and copy and remap simduid. Set
2890 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
2891 (copy_cfg_body): Remove blocks_to_copy argument. Use
2892 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
2893 caller. Don't set cfun->has_simduid_loops and/or
2894 cfun->has_force_vect_loops here.
2895 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
2897 (expand_call_inline, tree_function_versioning): Adjust copy_body
2900 2013-09-13 Martin Jambor <mjambor@suse.cz>
2903 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
2904 the assert if the edge was a speculative one.
2906 2013-09-13 Richard Biener <rguenther@suse.de>
2908 * tree-data-ref.h (known_dependences_p): Move here ...
2909 * tree-loop-distribution.c (known_dependences_p): ... from here.
2910 (dump_rdg_component, debug_rdg_component): Remove.
2912 (generate_loops_for_partition): Use gimple_uid instead of
2913 relying on matching stmt visit order.
2914 (rdg_build_partitions): Take starting stmt vector.
2915 (ldist_gen): Merge into ...
2916 (distribute_loop): ... this function. Do not compute starting
2918 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
2920 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2922 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
2923 Set type attribute correctly. Set predicable_short_it attribute.
2924 (cmpsi_shiftsi): Remove %? from output template.
2926 2013-09-13 Richard Biener <rguenther@suse.de>
2928 * tree-loop-distribution.c (struct rdg_component,
2929 rdg_defs_used_in_other_loops_p, free_rdg_components,
2930 rdg_build_components): Remove.
2931 (stmts_from_loop): Do not record virtual PHIs.
2932 (generate_loops_for_partition): Skip virtual PHIs.
2933 (build_rdg_partition_for_component): Rename to ...
2934 (build_rdg_partition_for_vertex): ... this and adjust.
2935 (rdg_build_partitions): Take a vector of starting vertices
2936 instead of components. Remove unnecessary leftover handling.
2937 (ldist_gen): Do not build components or record other stores.
2938 (distribute_loop): Do not distribute loops containing stmts
2941 2013-09-13 Christian Bruel <christian.bruel@st.com>
2944 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
2946 2013-09-13 Kai Tietz <ktietz@redhat.com>
2948 * config.gcc: Separate cases for mingw and cygwin targets,
2949 and add 64-bit cygwin target case.
2951 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
2952 dll-export inline-functions.
2953 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
2956 2013-09-13 Jeff Law <law@redhat.com>
2960 2013-09-06 Jeff Law <law@redhat.com>
2962 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
2963 edge implied equivalences into successor phis.
2965 2013-09-12 DJ Delorie <dj@redhat.com>
2967 * config/rl78/rl78-virt.md: Change from | to \; for asm line
2970 2013-09-12 Brooks Moses <bmoses@google.com>
2973 * Makefile.in: Do not install driver binaries in $(target)/bin.
2975 2013-09-12 DJ Delorie <dj@redhat.com>
2977 * config/rl78/rl78.opt (mrelax): New.
2978 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
2979 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
2981 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
2983 (rl78_expand_epilogue): Likewise.
2985 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
2988 * lra-eliminations.c (remove_reg_equal_offset_note): New.
2989 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
2990 pointer elimination with using remove_reg_equal_offset_note.
2992 2013-09-12 DJ Delorie <dj@redhat.com>
2994 * config/msp430/: New port.
2995 * config.gcc (msp430): Added.
2996 * doc/invoke.texi: Document MSP430 options.
2997 * doc/install.texi: Document msp430-elf
2998 * doc/md.texi: Document msp430-elf
2999 * doc/contrib.texi: Document msp430-elf
3001 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
3004 2013-09-12 Martin Jambor <mjambor@suse.cz>
3007 * ipa-prop.c (remove_described_reference): Give up if the edge in the
3008 reference descriptor is NULL.
3009 (ipa_edge_removal_hook): If owning a reference descriptor, set its
3012 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
3014 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
3015 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
3016 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
3017 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
3018 make_temp_ssa_name): move to tree-ssanames.h
3019 * tree-ssa-alias.h: Move prototype.
3020 * tree-ssa.h: Include tree-ssanames.h.
3021 * tree-ssanames.c (FREE_SSANAMES): Move to here.
3022 * tree-ssanames.h: New. Move items from tree-flow*.h
3023 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
3025 2013-09-12 Richard Biener <rguenther@suse.de>
3027 PR tree-optimization/58404
3028 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
3029 propagate non-invariant addresses into dereferences wrapped
3030 in component references.
3032 2013-09-12 Richard Biener <rguenther@suse.de>
3034 PR tree-optimization/58402
3035 * passes.def: Move pass_late_warn_uninitialized later.
3037 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
3039 * tree-ssa.h: New. Move content from tree-flow.h and
3041 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
3042 Move prototypes belonging to tree-ssa.c.
3043 * tree-flow-inline.h (redirect_edge_var_map_def,
3044 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
3046 * gimple.h: Adjust prototypes.
3047 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
3049 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
3050 * tree.h: Move prototype to tree-ssa.h.
3051 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
3052 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
3053 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
3054 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
3055 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
3056 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
3057 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
3058 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
3059 graphite-blocking.c, graphite-clast-to-gimple.c,
3060 graphite-dependences.c, graphite-interchange.c,
3061 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
3062 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
3063 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
3064 ipa-reference.c, ipa-split.c, ipa-utils.c,
3065 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
3066 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
3067 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
3068 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
3069 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
3070 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
3071 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
3072 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
3073 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
3074 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
3075 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
3076 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
3077 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
3078 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
3079 value-prof.c, var-tracking.c,
3080 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
3082 2013-09-12 Richard Biener <rguenther@suse.de>
3084 PR tree-optimization/58396
3085 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
3086 (build_rdg): Take a loop-nest parameter, fix memleaks.
3087 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
3089 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
3091 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
3094 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
3096 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
3097 The FMA instruction names should have a 'v' prefix.
3099 2013-09-12 Richard Biener <rguenther@suse.de>
3101 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
3102 (dot_rdg): Use popen instead of system in optional code.
3103 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
3104 (already_processed_vertex_p): Adjust.
3105 (has_anti_or_output_dependence, predecessor_has_mem_write,
3106 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
3107 rdg_flag_uses): Remove.
3108 (rdg_flag_vertex): Simplify.
3109 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
3111 (build_rdg_partition_for_component): Process the first vertex
3112 of a component only.
3113 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
3115 2013-09-12 Alan Modra <amodra@gmail.com>
3117 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
3119 2013-09-11 DJ Delorie <dj@redhat.com>
3120 Nick Clifton <nickc@redhat.com>
3122 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
3123 (rl78_stack_based_mem): New.
3124 * config/rl78/constraints.md (Iv08): New.
3133 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
3135 (movsi): Change from expand to insn-and-split.
3136 (ashrsi3): Clobber AX.
3140 * config/rl78/rl78.md (CC_REG): Fix.
3141 (addsi3): Allow memory and immediate operands.
3142 (addsi3_internal): Split into...
3143 (addsi3_internal_virt): ...new, and ...
3144 (addsi3_internal_real): ...new.
3146 (subsi3_internal_virt): New.
3147 (subsi3_internal_real): New.
3148 (mulsi3): Add memory operand.
3149 (mulsi3_rl78): Likewise.
3150 (mulsi3_g13): Likewise.
3151 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
3152 (cbranchqi4_real): Add more constraint options.
3153 (cbranchhi4_real): Expand pattern.
3154 (cbranchhi4_real_signed): New.
3155 (cbranchhi4_real_inverted): New.
3156 (cbranchsi4_real_lt): New.
3157 (cbranchsi4_real_ge): New.
3158 (cbranchsi4_real_signed): New.
3159 (cbranchsi4_real): New.
3161 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
3163 (lshrsi3_virt): Likewise.
3164 (ashlsi3_virt): Likewise.
3165 (cbranchqi4_virt_signed): New.
3166 (cbranchhi4_virt_signed): New.
3167 (cbranchsi4_virt): New.
3168 * config/rl78/rl78.c: Whitespace fixes throughout.
3169 (move_elim_pass): New.
3170 (pass_data_rl78_move_elim): New.
3171 (pass_rl78_move_elim): New.
3172 (make_pass_rl78_move_elim): New.
3173 (rl78_devirt_info): Run devirt earlier.
3174 (rl78_move_elim_info): New.
3175 (rl78_asm_file_start): Register it.
3176 (rl78_split_movsi): New.
3177 (rl78_as_legitimate_address): Allow virtual base registers when
3179 (rl78_addr_space_convert): Remove spurious debug stuff.
3180 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
3181 (rl78_print_operand): More cases for not printing '#'.
3182 (rl78_expand_compare): Remove most of the logic.
3183 (content_memory): New.
3184 (clear_content_memory): New.
3185 (get_content_index): New.
3186 (get_content_name): New.
3187 (display_content_memory): New.
3188 (update_content): New.
3189 (record_content): New.
3190 (already_contains): New.
3191 (insn_ok_now): Re-recog insns with virtual registers.
3192 (add_postponed_content_update): New.
3193 (process_postponed_content_update): New.
3194 (gen_and_emit_move): New.
3195 (transcode_memory_rtx): Record new location content.
3196 Use gen_and_emit_move.
3197 (force_into_acc): New.
3198 (move_to_acc): Use gen_and_emit_move.
3199 (move_from_acc): Likewise.
3200 (move_acc_to_reg): Likewise.
3201 (move_to_x): Likewise.
3202 (move_to_hl): Likewise.
3203 (move_to_de): Likewise.
3204 (rl78_alloc_physical_registers_op1): Record location content.
3205 (has_constraint): New.
3206 (rl78_alloc_physical_registers_op2): Record location content.
3208 (rl78_alloc_physical_registers_ro1): Likewise.
3209 (rl78_alloc_physical_registers_cmp): Likewise.
3210 (rl78_alloc_physical_registers_umul): Likewise.
3211 (rl78_alloc_address_registers_macax): New.
3212 (rl78_alloc_physical_registers): Initialize and set location
3213 content memory as needed.
3214 (rl78_reorg): Make sure split2 is called.
3215 (rl78_rtx_costs): New.
3217 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
3219 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
3220 for (not (neg ...)) and (neg (not ...)) cases.
3222 2013-09-11 Richard Biener <rguenther@suse.de>
3225 * passes.def: Split critical edges before late uninit warning passes.
3226 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
3228 2013-09-11 Jakub Jelinek <jakub@redhat.com>
3230 PR tree-optimization/58385
3231 * fold-const.c (build_range_check): If both low and high are NULL,
3232 use omit_one_operand_loc to preserve exp side-effects.
3234 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3236 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
3238 2013-09-11 Richard Biener <rguenther@suse.de>
3240 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
3241 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
3242 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
3243 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
3244 stmts_from_loop, known_dependences_p, build_empty_rdg,
3245 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
3246 * tree-loop-distribution.c: ... here.
3247 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
3248 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
3249 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
3250 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
3251 * tree-loop-distribution.c: ... here.
3252 * tree-loop-distribution.c: Include gimple-pretty-print.h.
3253 (struct partition_s): Add loops member.
3254 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
3255 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
3256 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
3258 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3259 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3260 Sergey Lega <sergey.s.lega@intel.com>
3261 Anna Tikhonova <anna.tikhonova@intel.com>
3262 Ilya Tocar <ilya.tocar@intel.com>
3263 Andrey Turetskiy <andrey.turetskiy@intel.com>
3264 Ilya Verbin <ilya.verbin@intel.com>
3265 Kirill Yukhin <kirill.yukhin@intel.com>
3266 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3268 * config/i386/constraints.md (k): New.
3270 * config/i386/i386.c (const regclass_map): Add new mask registers.
3271 (dbx_register_map): Ditto.
3272 (dbx64_register_map): Ditto.
3273 (svr4_dbx_register_map): Ditto.
3274 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
3276 (ix86_preferred_reload_class): Disable constants for mask registers.
3277 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
3278 (ix86_hard_regno_mode_ok): Support new mask registers.
3279 (x86_order_regs_for_local_alloc): Ditto.
3280 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
3281 (FIXED_REGISTERS): Add new mask registers.
3282 (CALL_USED_REGISTERS): Ditto.
3283 (REG_ALLOC_ORDER): Ditto.
3284 (VALID_MASK_REG_MODE): New.
3285 (FIRST_MASK_REG): Ditto.
3286 (LAST_MASK_REG): Ditto.
3287 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
3288 (MAYBE_MASK_CLASS_P): New.
3289 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
3290 (REG_CLASS_CONTENTS): Ditto.
3291 (MASK_REGNO_P): New.
3292 (ANY_MASK_REG_P): Ditto.
3293 (HI_REGISTER_NAMES): Add new mask registers.
3294 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
3295 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
3297 (attribute "type"): Add mskmov, msklog.
3298 (attribute "length_immediate"): Support them.
3299 (attribute "memory"): Ditto.
3300 (attribute "prefix_0f"): Ditto.
3301 (*movhi_internal): Support new mask registers.
3302 (*movqi_internal): Ditto.
3303 (define_split): Split out clobber pattern is a logic
3304 insn on mask registers.
3305 (*k<logic><mode>): New.
3306 (*andhi_1): Extend to support mask regs.
3307 (*andqi_1): Extend to support mask regs.
3309 (define_split): Split and-not to and and not if operands
3311 (*<code><mode>_1): Separate HI mode to new pattern...
3312 (*<code>hi_1): This.
3313 (*<code>qi_1): Extend to support mask regs.
3315 (kortestzhi): Ditto.
3316 (kortestchi): Ditto.
3318 (*one_cmpl<mode>2_1): Remove HImode and handle it...
3319 (*one_cmplhi2_1): ...Here, now with mask registers support.
3320 (*one_cmplqi2_1): Support new mask registers.
3321 (HI/QImode arithmetics splitter): Don't split if mask registers
3323 (HI/QImode not splitter): Ditto.
3324 * config/i386/predicated.md (mask_reg_operand): New.
3325 (general_reg_operand): Ditto.
3327 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3329 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
3330 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
3332 2013-09-10 Jeff Law <law@redhat.com>
3334 PR tree-optimization/58380
3335 * tree-ssa-threadupdate.c (thread_block): Recognize another case
3336 of threading through a buried loop header.
3338 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
3339 return value for single successor case.
3341 2013-09-10 Jan Hubicka <jh@suse.cz>
3343 * ipa-devirt.c (ipa_devirt): Enable with LTO.
3345 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
3348 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
3349 support conditional execution.
3350 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
3352 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
3354 * lra.c (lra): Clear lra_optional_reload_pseudos before every
3356 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
3357 Check destination too to check move insn.
3358 (undo_optional_reloads): Add check that the original peudo did not
3359 changed its allocation and the optional reload was inherited on last
3360 inheritance pass. Break loop after deciding to keep optional reload.
3361 (lra_undo_inheritance): Add check that inherited pseudo still in
3364 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
3366 * config/aarch64/aarch64.md (generic_sched): New.
3367 * config/aarch64/aarch64-generic.md (load): Make conditional
3368 on generic_sched attribute.
3369 (nonload): Likewise.
3371 2013-09-10 Jan Hubicka <jh@suse.cz>
3373 * lto-cgraph.c: Include ipa-utils.h.
3374 (compute_ltrans_boundary): Also add possible targets into the boundary.
3376 2013-09-10 Jan Hubicka <jh@suse.cz>
3378 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
3379 VAR_DECL of vtable rather than full expression.
3381 2013-09-10 Jan Hubicka <jh@suse.cz>
3382 Paolo Carlini <paolo.carlini@oracle.com>
3384 * cgraphunit.c (analyze_functions): Save input_location, set it
3385 to UNKNOWN_LOCATION and restore it at the end.
3387 2013-09-10 Martin Jambor <mjambor@suse.cz>
3389 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
3390 represented by a thunk.
3392 2013-09-10 Jeff Law <law@redhat.com>
3394 PR tree-optimization/58343
3395 * tree-ssa-threadupdate.c (thread_block): Identify and disable
3396 jump threading requests through loop headers buried in the middle
3397 of a jump threading path.
3399 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
3400 in return value/type.
3402 2013-09-10 Jakub Jelinek <jakub@redhat.com>
3404 PR rtl-optimization/58365
3405 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
3406 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
3409 2013-09-10 Richard Biener <rguenther@suse.de>
3411 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
3412 * tree-data-ref.c (create_rdg_vertices): Collect all data
3413 references, signal failure to the caller, use data-ref API.
3414 (build_rdg): Compute data references only once. Maintain lifetime
3415 of data references and data dependences from within RDG.
3416 (free_rdg): Free dependence relations.
3417 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
3418 inventing extra dependences.
3419 (distribute_loop): Update for RDG API changes.
3421 2013-09-10 Kai Tietz <ktietz@redhat.com>
3423 * doc/invoke.texi (fms-extensions): Document changed
3424 behavior for ms-abi targets.
3425 * config/i386/i386.c (ix86_option_override_internal):
3426 Set default value of option -fms-extension for ms-abi targets.
3428 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
3430 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
3432 2013-09-10 Alan Modra <amodra@gmail.com>
3435 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
3437 2013-09-10 Alan Modra <amodra@gmail.com>
3439 * config/rs6000/predicates.md (add_cint_operand): New.
3440 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
3441 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
3442 using add_cint_operand.
3443 (largetoc_high_plus_aix): Likewise.
3445 2013-09-09 Jakub Jelinek <jakub@redhat.com>
3447 PR tree-optimization/58364
3448 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
3449 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
3450 the current range can't be an unconditional true or false.
3452 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
3454 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
3456 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
3458 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
3460 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
3463 * doc/invoke.texi (-Wdelete-incomplete): Document it.
3465 2013-09-09 Ian Bolton <ian.bolton@arm.com>
3467 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
3468 NO_REGS for immediate that can't be moved directly into FP_REGS.
3470 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3472 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
3473 comparison with negated operand.
3474 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
3477 2013-09-09 Richard Biener <rguenther@suse.de>
3480 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
3481 of a subloop record all its block as affecting loop-closed SSA form.
3483 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
3485 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
3486 of an rtx/bitpos pair.
3487 (store_fixed_bit_field): Update accordingly.
3489 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
3491 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
3493 * builtins.c (expand_errno_check): Likewise.
3494 * dwarf2cfi.c (init_return_column_size): Likewise.
3495 * except.c (sjlj_mark_call_sites): Likewise.
3496 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
3497 * lra-constraints.c (emit_inc): Likewise.
3498 * ree.c (combine_set_extension): Likewise.
3499 * regmove.c (fixup_match_2): Likewise.
3500 * reload1.c (inc_for_reload): Likewise.
3502 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
3504 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
3505 (force_to_mode, simplify_shift_const_1, simplify_comparison):
3506 Use gen_int_mode with the mode of the associated simplify_* call.
3507 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
3508 * expmed.c (expand_shift_1): Likewise.
3509 * function.c (instantiate_virtual_regs_in_insn): Likewise.
3510 * loop-iv.c (iv_number_of_iterations): Likewise.
3511 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
3512 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
3514 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
3516 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
3517 of the associated expand_* call.
3518 (asan_emit_stack_protection): Likewise.
3519 * builtins.c (round_trampoline_addr): Likewise.
3520 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
3521 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
3522 (emit_store_flag): Likewise.
3523 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
3524 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
3526 * function.c (instantiate_virtual_regs_in_insn): Likewise.
3527 * ifcvt.c (noce_try_store_flag_constants): Likewise.
3528 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
3529 * modulo-sched.c (generate_prolog_epilog): Likewise.
3530 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
3531 (expand_ctz, expand_ffs, expand_unop): Likewise.
3533 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
3535 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
3536 of the associated gen_rtx_* call.
3537 * caller-save.c (init_caller_save): Likewise.
3538 * combine.c (find_split_point, make_extraction): Likewise.
3539 (make_compound_operation): Likewise.
3540 * dwarf2out.c (mem_loc_descriptor): Likewise.
3541 * explow.c (plus_constant, probe_stack_range): Likewise.
3542 * expmed.c (expand_mult_const): Likewise.
3543 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
3544 * reload1.c (init_reload): Likewise.
3545 * valtrack.c (cleanup_auto_inc_dec): Likewise.
3546 * var-tracking.c (adjust_mems): Likewise.
3547 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
3548 rather than gen_rtx_fmt_ee.
3550 2013-09-09 Jan Hubicka <jh@suse.cz>
3553 * value-prof.c (gimple_ic): Copy also abnormal edges.
3555 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
3557 * asan.c (asan_shadow_cst): Use gen_int_mode.
3559 2013-09-08 Jan Hubicka <jh@suse.cz>
3561 * ipa-profile.c: Add toplevel comment.
3562 (ipa_propagate_frequency_1): Be more conservative when profile is read.
3563 (contains_hot_call_p): New function.
3564 (ipa_propagate_frequency): Set frequencies based on counts when
3566 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
3567 profile; do not tamper with profile after inlining if it is read.
3569 2013-09-08 Jan Hubicka <jh@suse.cz>
3571 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
3574 2013-09-08 Jan Hubicka <jh@suse.cz>
3576 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
3579 2013-09-08 Jeff Law <law@redhat.com>
3582 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
3585 2013-09-08 Andi Kleen <ak@linux.intel.com>
3587 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
3589 2013-09-08 Jan Hubicka <jh@suse.cz>
3591 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
3593 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
3595 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
3596 for non-debug insns.
3597 * lra.c (new_insn_reg): Take the containing insn as a parameter.
3598 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
3599 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
3602 2013-09-08 Jan Hubicka <jh@suse.cz>
3604 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
3605 targets and devirtualize to BUILT_IN_UNREACHABLE.
3606 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
3607 * ipa.c (walk_polymorphic_call_targets): New function.
3608 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
3609 functions; use the new timevar.
3610 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
3611 was removed from the program.
3612 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
3613 not consider it in the walk when its vtable is dead.
3614 (possible_polymorphic_call_targets_1): Pass anonymous flag to
3616 (devirt_variable_node_removal_hook): New function.
3617 (possible_polymorphic_call_targets): Also register
3618 devirt_variable_node_removal_hook.
3619 (ipa_devirt): Do not do non-speculative devirtualization.
3620 (gate_ipa_devirt): One execute if devirtualizing speculatively.
3622 2013-09-08 Jan Hubicka <jh@suse.cz>
3624 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
3625 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
3626 varpool_remove_variable_insertion_hook): Declare.
3627 * varpool.c (varpool_node_hook_list): New structure.
3628 (first_varpool_node_removal_hook,
3629 first_varpool_variable_insertion_hook): New variables.
3630 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
3631 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
3632 varpool_remove_variable_insertion_hook,
3633 varpool_call_variable_insertion_hooks): New functions.
3634 (varpool_remove_node): Use it.
3636 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
3639 * diagnostic.c (diagnostic_build_prefix): When s.file is
3640 "<built-in>" don't output line and column numbers.
3642 2013-09-06 Jan Hubicka <jh@suse.cz>
3644 * cgraphunit.c (expand_thunk): Get body before touching arguments.
3645 * lto-streamer-out.c: Stream thunks, too.
3646 * lto-streamer-in.c (input_function): Pop cfun here
3647 (lto_read_body): Instead of here.
3649 2013-09-06 Caroline Tice <cmtice@google.com>
3651 * doc/install.texi: Add documentation for the --enable-vtable-verify
3652 and the --disable-libvtv configure options.
3654 2013-09-06 Jeff Law <law@redhat.com>
3656 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
3657 edge implied equivalences into successor phis.
3659 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
3661 * resource.c (mark_referenced_resources): Handle COND_EXEC.
3663 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
3665 * resource.c (mark_target_live_regs): Compute resources taking
3666 into account if a call is predicated or not.
3668 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
3670 * toplev.c (output_stack_usage): Be prepared for suffixes created by
3671 the compiler in the function names.
3673 2013-09-06 Jan Hubicka <jh@suse.cz>
3676 * ipa-inline.c (has_caller_p): New function.
3677 (want_inline_function_to_all_callers_p): Use it.
3678 (sum_callers, inline_to_all_callers): Break out from ...
3679 (ipa_inline): ... here.
3681 2013-09-06 Jan Hubicka <jh@suse.cz>
3683 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
3684 only when AVX is enabled.
3686 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3688 * config/aarch64/aarch64.md
3689 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
3690 is fpsimd_<load/store>2.
3691 (load_pair<mode>): Likewise.
3692 (store_pair<mode>): Likewise.
3694 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3696 * config/arm/types.md (type): Add "mrs" type.
3697 * config/aarch64/aarch64.md
3698 (aarch64_load_tp_hard): Make type "mrs".
3700 (load_tp_hard): Make type "mrs".
3701 * config/arm/cortex-a15.md: Update with new attributes.
3702 * config/arm/cortex-a5.md: Update with new attributes.
3703 * config/arm/cortex-a53.md: Update with new attributes.
3704 * config/arm/cortex-a7.md: Update with new attributes.
3705 * config/arm/cortex-a8.md: Update with new attributes.
3706 * config/arm/cortex-a9.md: Update with new attributes.
3707 * config/arm/cortex-m4.md: Update with new attributes.
3708 * config/arm/cortex-r4.md: Update with new attributes.
3709 * config/arm/fa526.md: Update with new attributes.
3710 * config/arm/fa606te.md: Update with new attributes.
3711 * config/arm/fa626te.md: Update with new attributes.
3712 * config/arm/fa726te.md: Update with new attributes.
3714 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3716 * config/aarch64/aarch64.md
3717 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
3718 (*movtf_aarch64): Likewise.
3720 (thumb1_movdi_insn): Use "multiple" for type where more than one
3721 instruction is used for a move.
3722 (*arm32_movhf): Likewise.
3723 (*thumb_movdf_insn): Likewise.
3725 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3727 * config/arm/types.md (type): Rename fcpys to fmov.
3729 (*arm_movsi_vfp): Rename type fcpys as fmov.
3730 (*thumb2_movsi_vfp): Likewise
3731 (*movhf_vfp_neon): Likewise
3732 (*movhf_vfp): Likewise
3733 (*movsf_vfp): Likewise
3734 (*thumb2_movsf_vfp): Likewise
3735 (*movsfcc_vfp): Likewise
3736 (*thumb2_movsfcc_vfp): Likewise
3737 * config/aarch64/aarch64-simd.md
3738 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
3739 * config/aarch64/aarch64.md
3740 (*movsi_aarch64): Replace type mov_reg with fmovs.
3741 (*movdi_aarch64): Likewise
3742 (*movsf_aarch64): Likewise
3743 (*movdf_aarch64): Likewise
3745 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
3746 * config/arm/iwmmxt.md
3747 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
3748 * config/arm/arm1020e.md: Update with new attributes.
3749 * config/arm/cortex-a15-neon.md: Update with new attributes.
3750 * config/arm/cortex-a5.md: Update with new attributes.
3751 * config/arm/cortex-a53.md: Update with new attributes.
3752 * config/arm/cortex-a7.md: Update with new attributes.
3753 * config/arm/cortex-a8-neon.md: Update with new attributes.
3754 * config/arm/cortex-a9.md: Update with new attributes.
3755 * config/arm/cortex-m4-fpu.md: Update with new attributes.
3756 * config/arm/cortex-r4f.md: Update with new attributes.
3757 * config/arm/marvell-pj4.md: Update with new attributes.
3758 * config/arm/vfp11.md: Update with new attributes.
3760 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3762 * config/aarch64/aarch64.md
3763 (*madd<mode>): Fix type attribute.
3764 (*maddsi_uxtw): Likewise.
3765 (*msub<mode>): Likewise.
3766 (*msubsi_uxtw): Likewise.
3767 (<su_optab>maddsidi4): Likewise.
3768 (<su_optab>msubsidi4): Likewise.
3770 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3772 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
3773 * config/arm/arm.md (core_cycles): Remove fdiv.
3774 * config/arm/vfp.md:
3775 (*sqrtsf2_vfp): Update for attribute changes.
3776 (*sqrtdf2_vfp): Likewise.
3777 * config/aarch64/aarch64.md:
3778 (sqrt<mode>2): Update for attribute changes.
3779 * config/arm/arm1020e.md: Update with new attributes.
3780 * config/arm/cortex-a15-neon.md: Update with new attributes.
3781 * config/arm/cortex-a5.md: Update with new attributes.
3782 * config/arm/cortex-a53.md: Update with new attributes.
3783 * config/arm/cortex-a7.md: Update with new attributes.
3784 * config/arm/cortex-a8-neon.md: Update with new attributes.
3785 * config/arm/cortex-a9.md: Update with new attributes.
3786 * config/arm/cortex-m4-fpu.md: Update with new attributes.
3787 * config/arm/cortex-r4f.md: Update with new attributes.
3788 * config/arm/marvell-pj4.md: Update with new attributes.
3789 * config/arm/vfp11.md: Update with new attributes.
3791 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3793 * config/arm/types.md
3794 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
3795 * config/aarch64/aarch64.md
3796 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
3798 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
3799 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
3800 (float<GPI:mode><GPF:mode>2): Likewise.
3802 (*truncsisf2_vfp): Update with new attributes.
3803 (*truncsidf2_vfp): Likewise.
3804 (fixuns_truncsfsi2): Likewise.
3805 (fixuns_truncdfsi2): Likewise.
3806 (*floatsisf2_vfp): Likewise.
3807 (*floatsidf2_vfp): Likewise.
3808 (floatunssisf2): Likewise.
3809 (floatunssidf2): Likewise.
3810 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
3811 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
3812 * config/arm/arm1020e.md: Update with new attributes.
3813 * config/arm/cortex-a15-neon.md: Update with new attributes.
3814 * config/arm/cortex-a5.md: Update with new attributes.
3815 * config/arm/cortex-a53.md: Update with new attributes.
3816 * config/arm/cortex-a7.md: Update with new attributes.
3817 * config/arm/cortex-a8-neon.md: Update with new attributes.
3818 * config/arm/cortex-a9.md: Update with new attributes.
3819 * config/arm/cortex-m4-fpu.md: Update with new attributes.
3820 * config/arm/cortex-r4f.md: Update with new attributes.
3821 * config/arm/marvell-pj4.md: Update with new attributes.
3822 * config/arm/vfp11.md: Update with new attributes.
3824 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3826 * config/aarch64/arm_neon.h
3827 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
3828 (vqtbx<1,2,3,4><q>_s8): Likewise.
3830 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3832 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
3833 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
3834 (add<mode>3): Add type attribute.
3835 (add<mode>3): Likewise.
3836 (usadd<mode>3): Likewise.
3837 (ssadd<mode>3): Likewise.
3838 (sub<mode>3): Likewise.
3839 (sub<mode>3): Likewise.
3840 (ussub<mode>3): Likewise.
3841 (sssub<mode>3): Likewise.
3842 (ssmulsa3): Likewise.
3843 (usmulusa3): Likewise.
3844 (arm_usatsihi): Likewise.
3846 (*movdi_vfp): Add types for all instructions.
3847 (*movdi_vfp_cortexa8): Likewise.
3848 (*movhf_vfp_neon): Likewise.
3849 (*movhf_vfp): Likewise.
3850 (*movdf_vfp): Likewise.
3851 (*thumb2_movdf_vfp): Likewise.
3852 (*thumb2_movdfcc_vfp): Likewise.
3853 * config/arm/arm.md: Add type attribute to all insn patterns.
3854 (*thumb1_adddi3): Add type attribute.
3855 (*arm_adddi3): Likewise.
3856 (*adddi_sesidi_di): Likewise.
3857 (*adddi_zesidi_di): Likewise.
3858 (*thumb1_addsi3): Likewise.
3859 (addsi3_compare0): Likewise.
3860 (*addsi3_compare0_scratch): Likewise.
3861 (*compare_negsi_si): Likewise.
3862 (cmpsi2_addneg): Likewise.
3863 (*addsi3_carryin_<optab>): Likewise.
3864 (*addsi3_carryin_alt2_<optab>): Likewise.
3865 (*addsi3_carryin_clobercc_<optab>): Likewise.
3866 (*subsi3_carryin): Likewise.
3867 (*subsi3_carryin_const): Likewise.
3868 (*subsi3_carryin_compare): Likewise.
3869 (*subsi3_carryin_compare_const): Likewise.
3870 (*arm_subdi3): Likewise.
3871 (*thumb_subdi3): Likewise.
3872 (*subdi_di_zesidi): Likewise.
3873 (*subdi_di_sesidi): Likewise.
3874 (*subdi_zesidi_di): Likewise.
3875 (*subdi_sesidi_di): Likewise.
3876 (*subdi_zesidi_ze): Likewise.
3877 (thumb1_subsi3_insn): Likewise.
3878 (*arm_subsi3_insn): Likewise.
3879 (*anddi3_insn): Likewise.
3880 (*anddi_zesidi_di): Likewise.
3881 (*anddi_sesdi_di): Likewise.
3882 (*ne_zeroextracts): Likewise.
3883 (*ne_zeroextracts): Likewise.
3884 (*ite_ne_zeroextr): Likewise.
3885 (*ite_ne_zeroextr): Likewise.
3886 (*anddi_notdi_di): Likewise.
3887 (*anddi_notzesidi): Likewise.
3888 (*anddi_notsesidi): Likewise.
3889 (andsi_notsi_si): Likewise.
3890 (thumb1_bicsi3): Likewise.
3891 (*iordi3_insn): Likewise.
3892 (*iordi_zesidi_di): Likewise.
3893 (*iordi_sesidi_di): Likewise.
3894 (*thumb1_iorsi3_insn): Likewise.
3895 (*xordi3_insn): Likewise.
3896 (*xordi_zesidi_di): Likewise.
3897 (*xordi_sesidi_di): Likewise.
3898 (*arm_xorsi3): Likewise.
3899 (*andsi_iorsi3_no): Likewise.
3900 (*smax_0): Likewise.
3901 (*smax_m1): Likewise.
3902 (*arm_smax_insn): Likewise.
3903 (*smin_0): Likewise.
3904 (*arm_smin_insn): Likewise.
3905 (*arm_umaxsi3): Likewise.
3906 (*arm_uminsi3): Likewise.
3907 (*minmax_arithsi): Likewise.
3908 (*minmax_arithsi_): Likewise.
3909 (*satsi_<SAT:code>): Likewise.
3910 (arm_ashldi3_1bit): Likewise.
3911 (arm_ashrdi3_1bit): Likewise.
3912 (arm_lshrdi3_1bit): Likewise.
3913 (*arm_negdi2): Likewise.
3914 (*thumb1_negdi2): Likewise.
3915 (*arm_negsi2): Likewise.
3916 (*thumb1_negsi2): Likewise.
3917 (*negdi_extendsid): Likewise.
3918 (*negdi_zero_extend): Likewise.
3919 (*arm_abssi2): Likewise.
3920 (*thumb1_abssi2): Likewise.
3921 (*arm_neg_abssi2): Likewise.
3922 (*thumb1_neg_abss): Likewise.
3923 (one_cmpldi2): Likewise.
3924 (extend<mode>di2): Likewise.
3925 (*compareqi_eq0): Likewise.
3926 (*arm_extendhisi2addsi): Likewise.
3927 (*arm_movdi): Likewise.
3928 (*thumb1_movdi_insn): Likewise.
3929 (*arm_movt): Likewise.
3930 (*thumb1_movsi_insn): Likewise.
3931 (pic_add_dot_plus_four): Likewise.
3932 (pic_add_dot_plus_eight): Likewise.
3933 (tls_load_dot_plus_eight): Likewise.
3934 (*thumb1_movhi_insn): Likewise.
3935 (*thumb1_movsf_insn): Likewise.
3936 (*movdf_soft_insn): Likewise.
3937 (*thumb_movdf_insn): Likewise.
3938 (cbranchsi4_insn): Likewise.
3939 (cbranchsi4_scratch): Likewise.
3940 (*negated_cbranchsi4): Likewise.
3941 (*tbit_cbranch): Likewise.
3942 (*tlobits_cbranch): Likewise.
3943 (*tstsi3_cbranch): Likewise.
3944 (*cbranchne_decr1): Likewise.
3945 (*addsi3_cbranch): Likewise.
3946 (*addsi3_cbranch_scratch): Likewise.
3947 (*arm_cmpdi_insn): Likewise.
3948 (*arm_cmpdi_unsig): Likewise.
3949 (*arm_cmpdi_zero): Likewise.
3950 (*thumb_cmpdi_zero): Likewise.
3951 (*deleted_compare): Likewise.
3952 (*mov_scc): Likewise.
3953 (*mov_negscc): Likewise.
3954 (*mov_notscc): Likewise.
3955 (*cstoresi_eq0_thumb1_insn): Likewise.
3956 (cstoresi_nltu_thumb1): Likewise.
3957 (cstoresi_ltu_thu): Likewise.
3958 (thumb1_addsi3_addgeu): Likewise.
3959 (*arm_jump): Likewise.
3960 (*thumb_jump): Likewise.
3961 (*check_arch2): Likewise.
3962 (arm_casesi_internal): Likewise.
3963 (thumb1_casesi_dispatch): Likewise.
3964 (*arm_indirect_jump): Likewise.
3965 (*thumb1_indirect_jump): Likewise.
3967 (*and_scc): Likewise.
3968 (*ior_scc): Likewise.
3969 (*compare_scc): Likewise.
3970 (*cond_move): Likewise.
3971 (*cond_arith): Likewise.
3972 (*cond_sub): Likewise.
3973 (*cmp_ite0): Likewise.
3974 (*cmp_ite1): Likewise.
3975 (*cmp_and): Likewise.
3976 (*cmp_ior): Likewise.
3977 (*ior_scc_scc): Likewise.
3978 (*ior_scc_scc_cmp): Likewise.
3979 (*and_scc_scc): Likewise.
3980 (*and_scc_scc_cmp): Likewise.
3981 (*and_scc_scc_nod): Likewise.
3982 (*negscc): Likewise.
3983 (movcond_addsi): Likewise.
3984 (movcond): Likewise.
3985 (*ifcompare_plus_move): Likewise.
3986 (*if_plus_move): Likewise.
3987 (*ifcompare_move_plus): Likewise.
3988 (*if_move_plus): Likewise.
3989 (*ifcompare_arith_arith): Likewise.
3990 (*if_arith_arith): Likewise.
3991 (*ifcompare_arith_move): Likewise.
3992 (*if_arith_move): Likewise.
3993 (*ifcompare_move_arith): Likewise.
3994 (*if_move_arith): Likewise.
3995 (*ifcompare_move_not): Likewise.
3996 (*if_move_not): Likewise.
3997 (*ifcompare_not_move): Likewise.
3998 (*if_not_move): Likewise.
3999 (*ifcompare_shift_move): Likewise.
4000 (*if_shift_move): Likewise.
4001 (*ifcompare_move_shift): Likewise.
4002 (*if_move_shift): Likewise.
4003 (*ifcompare_shift_shift): Likewise.
4004 (*ifcompare_not_arith): Likewise.
4005 (*ifcompare_arith_not): Likewise.
4006 (*if_arith_not): Likewise.
4007 (*ifcompare_neg_move): Likewise.
4008 (*if_neg_move): Likewise.
4009 (*ifcompare_move_neg): Likewise.
4010 (*if_move_neg): Likewise.
4011 (prologue_thumb1_interwork): Likewise.
4012 (*cond_move_not): Likewise.
4013 (*sign_extract_onebit): Likewise.
4014 (*not_signextract_onebit): Likewise.
4015 (stack_tie): Likewise.
4016 (align_4): Likewise.
4017 (align_8): Likewise.
4018 (consttable_end): Likewise.
4019 (consttable_1): Likewise.
4020 (consttable_2): Likewise.
4021 (consttable_4): Likewise.
4022 (consttable_8): Likewise.
4023 (consttable_16): Likewise.
4024 (*thumb1_tablejump): Likewise.
4025 (prefetch): Likewise.
4026 (force_register_use): Likewise.
4027 (thumb_eh_return): Likewise.
4028 (load_tp_hard): Likewise.
4029 (load_tp_soft): Likewise.
4030 (tlscall): Likewise.
4031 (*arm_movtas_ze): Likewise.
4032 (*arm_rev): Likewise.
4033 (*arm_revsh): Likewise.
4034 (*arm_rev16): Likewise.
4035 * config/arm/thumb2.md
4036 (*thumb2_smaxsi3): Likewise.
4037 (*thumb2_sminsi3): Likewise.
4038 (*thumb32_umaxsi3): Likewise.
4039 (*thumb2_uminsi3): Likewise.
4040 (*thumb2_negdi2): Likewise.
4041 (*thumb2_abssi2): Likewise.
4042 (*thumb2_neg_abss): Likewise.
4043 (*thumb2_movsi_insn): Likewise.
4044 (tls_load_dot_plus_four): Likewise.
4045 (*thumb2_movhi_insn): Likewise.
4046 (*thumb2_mov_scc): Likewise.
4047 (*thumb2_mov_negs): Likewise.
4048 (*thumb2_mov_negs): Likewise.
4049 (*thumb2_mov_nots): Likewise.
4050 (*thumb2_mov_nots): Likewise.
4051 (*thumb2_movsicc_): Likewise.
4052 (*thumb2_movsfcc_soft_insn): Likewise.
4053 (*thumb2_indirect_jump): Likewise.
4054 (*thumb2_and_scc): Likewise.
4055 (*thumb2_ior_scc): Likewise.
4056 (*thumb2_ior_scc_strict_it): Likewise.
4057 (*thumb2_cond_move): Likewise.
4058 (*thumb2_cond_arith): Likewise.
4059 (*thumb2_cond_ari): Likewise.
4060 (*thumb2_cond_sub): Likewise.
4061 (*thumb2_negscc): Likewise.
4062 (*thumb2_movcond): Likewise.
4063 (thumb2_casesi_internal): Likewise.
4064 (thumb2_casesi_internal_pic): Likewise.
4065 (*thumb2_alusi3_short): Likewise.
4066 (*thumb2_mov<mode>_shortim): Likewise.
4067 (*thumb2_addsi_short): Likewise.
4068 (*thumb2_subsi_short): Likewise.
4069 (thumb2_addsi3_compare0): Likewise.
4070 (*thumb2_cbz): Likewise.
4071 (*thumb2_cbnz): Likewise.
4072 (*thumb2_one_cmplsi2_short): Likewise.
4073 (*thumb2_negsi2_short): Likewise.
4074 (*orsi_notsi_si): Likewise.
4075 * config/arm/arm1020e.md: Update with new attributes.
4076 * config/arm/arm1026ejs.md: Update with new attributes.
4077 * config/arm/arm1136jfs.md: Update with new attributes.
4078 * config/arm/arm926ejs.md: Update with new attributes.
4079 * config/arm/cortex-a15.md: Update with new attributes.
4080 * config/arm/cortex-a5.md: Update with new attributes.
4081 * config/arm/cortex-a53.md: Update with new attributes.
4082 * config/arm/cortex-a7.md: Update with new attributes.
4083 * config/arm/cortex-a8.md: Update with new attributes.
4084 * config/arm/cortex-a9.md: Update with new attributes.
4085 * config/arm/cortex-m4.md: Update with new attributes.
4086 * config/arm/cortex-r4.md: Update with new attributes.
4087 * config/arm/fa526.md: Update with new attributes.
4088 * config/arm/fa606te.md: Update with new attributes.
4089 * config/arm/fa626te.md: Update with new attributes.
4090 * config/arm/fa726te.md: Update with new attributes.
4092 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4094 * config/aarch64/aarch64-simd.md
4095 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
4096 <vwx> iterator to ensure correct register choice.
4097 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
4098 (aarch64_sqdmull_n<mode>): Likewise.
4099 (aarch64_sqdmull2_n<mode>_internal): Likewise.
4100 * config/aarch64/arm_neon.h
4101 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
4102 (vml<as><q>_n_<su>16): Likewise.
4103 (vml<as>l_high_lane<q>_<su>16): Likewise.
4104 (vml<as>l_high_n_<su>16): Likewise.
4105 (vml<as>l_lane<q>_<su>16): Likewise.
4106 (vml<as>l_n_<su>16): Likewise.
4107 (vmul<q>_lane<q>_<su>16): Likewise.
4108 (vmul<q>_n_<su>16): Likewise.
4109 (vmull_lane<q>_<su>16): Likewise.
4110 (vmull_n_<su>16): Likewise.
4111 (vmull_high_lane<q>_<su>16): Likewise.
4112 (vmull_high_n_<su>16): Likewise.
4113 (vqrdmulh<q>_n_s16): Likewise.
4115 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
4117 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
4118 have the correct lane parameter.
4120 2013-09-06 Richard Biener <rguenther@suse.de>
4122 * cfganal.c (control_dependences::~control_dependences):
4123 Properly free all of the vector.
4125 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
4128 * config/i386/i386.c (ix86_conditional_register_usage):
4129 Proper initialize extended SSE registers.
4131 2013-09-06 Jan Hubicka <jh@suse.cz>
4133 PR tree-optimization/58311
4134 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
4136 2013-09-06 Jan Hubicka <jh@suse.cz>
4138 * Makefile.in (tree-sra.o): Update dependencies.
4139 * tree-sra.c: Include ipa-utils.h
4140 (scan_function): Use recursive_call_p.
4141 (has_caller_p): New function.
4142 (cgraph_for_node_and_aliases): Count also callers of aliases.
4144 2013-09-06 Jan Hubicka <jh@suse.cz>
4147 * cgraph.h (symtab_semantically_equivalent_p): Declare.
4148 * tree-tailcall.c: Include ipa-utils.h.
4149 (find_tail_calls): Use it.
4150 * ipa-pure-const.c (check_call): Likewise.
4151 * ipa-utils.c (recursive_call_p): New function.
4152 * ipa-utils.h (recursive_call_p): Dclare.
4153 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
4154 (symtab_semantically_equivalent_p): New function.
4155 * Makefile.in (tree-tailcall.o): Update dependencies.
4157 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
4159 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
4162 2013-09-06 Richard Biener <rguenther@suse.de>
4164 * lto-streamer.h (lto_global_var_decls): Remove.
4165 * Makefile.in (OBJS): Remove lto-symtab.o.
4166 (lto-symtab.o): Remove.
4167 (GTFILES): Remove lto-symtab.c
4168 * lto-symtab.c: Move to lto/
4170 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4172 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
4173 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
4174 (UNSPEC_FPINT_RINT): New constant definitions.
4175 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
4176 definition with 2 attributes.
4177 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
4178 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
4181 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4183 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
4184 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
4185 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
4186 attribute to "z196_alone".
4187 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
4190 2013-09-06 Richard Biener <rguenther@suse.de>
4192 * basic-block.h (class control_dependences): New.
4193 * tree-ssa-dce.c (control_dependence_map): Remove.
4195 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
4196 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
4197 find_pdom, find_control_dependence, find_all_control_dependences):
4199 (mark_control_dependent_edges_necessary,
4200 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
4201 tree_dce_done, perform_tree_ssa_dce): Adjust.
4202 * cfganal.c (set_control_dependence_map_bit,
4203 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
4204 find_all_control_dependences): Move from tree-ssa-dce.c and
4205 implement as methods of control_dependences class.
4206 (control_dependences::control_dependences): New.
4207 (control_dependences::~control_dependences): Likewise.
4208 (control_dependences::get_edges_dependent_on): Likewise.
4209 (control_dependences::get_edge): Likewise.
4211 2013-09-04 Jan Hubicka <jh@suse.cz>
4213 * tree.c (types_same_for_odr): Drop overactive check.
4214 * ipa-devirt.c (hash_type_name): Likewise.
4216 2013-09-04 Jan Hubicka <jh@suse.cz>
4218 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
4219 (analyze_functions): ... here.
4221 2013-09-04 Jan Hubicka <jh@suse.cz>
4224 * cgraphunit.c (analyze_functions): Clear AUX fields
4225 after processing; initialize assembler name has.
4227 2013-09-05 Jeff Law <law@redhat.com>
4229 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
4230 from thread_around_empty_block. Record threading path into PATH.
4231 Recurse if threading through the initial block is successful.
4232 (thread_across_edge): Corresponding changes to slightly simplify.
4234 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
4236 * config/aarch64/aarch64.md
4237 (type): Remove frecpe, frecps, frecpx.
4238 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
4239 fix to be a TARGET_SIMD instruction.
4240 (aarch64_frecps): Remove.
4241 * config/aarch64/aarch64-simd.md
4242 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
4243 (aarch64_frecps<mode>): Handle all float/vector of float modes.
4245 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
4246 Sofiane Naci <sofiane.naci@arm.com>
4248 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
4249 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
4250 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
4251 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
4252 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
4253 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
4254 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
4255 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
4256 into "clz, "rbit". Rename "shift" to "shift_imm".
4257 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
4258 changes. Update for attribute changes all occurrences of arlo_* and
4260 * config/arm/arm-fixed.md: Update for attribute changes
4261 all occurrences of arlo_* types.
4262 * config/arm/thumb2.md: Update for attribute changes all occurrences
4264 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
4265 (cortexa7_older_only): Likewise.
4266 (cortexa7_younger): Likewise.
4267 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
4268 (1020alu_shift_op): Likewise.
4269 (1020alu_shift_reg_op): Likewise.
4270 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
4271 (alu_shift_op): Likewise.
4272 (alu_shift_reg_op): Likewise.
4273 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
4274 (11_alu_shift_op): Likewise.
4275 (11_alu_shift_reg_op): Likewise.
4276 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
4277 (9_alu_shift_reg_op): Likewise.
4278 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
4280 (cortex_a15_alu_shift): Likewise.
4281 (cortex_a15_alu_shift_reg): Likewise.
4282 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
4284 (cortex_a5_alu_shift): Likewise.
4285 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
4287 (cortex_a53_alu_shift): Likewise.
4288 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
4290 (cortex_a7_alu_reg): Likewise.
4291 (cortex_a7_alu_shift): Likewise.
4292 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
4294 (cortex_a8_alu_shift): Likewise.
4295 (cortex_a8_alu_shift_reg): Likewise.
4296 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
4297 (cortex_a9_dp_shift): Likewise.
4298 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
4300 * config/arm/cortex-r4.md
4301 (cortex_r4_alu): Update for attribute changes.
4302 (cortex_r4_mov): Likewise.
4303 (cortex_r4_alu_shift_reg): Likewise.
4304 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
4305 (526_alu_shift_op): Likewise.
4306 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
4307 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
4308 (626te_alu_shift_op): Likewise.
4309 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
4310 (726te_alu_shift_op): Likewise.
4311 (726te_alu_shift_reg_op): Likewise.
4312 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
4313 (mp626_alu_shift_op): Likewise.
4314 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
4315 (pj4_alu_conds): Likewise.
4316 (pj4_shift): Likewise.
4317 (pj4_shift_conds): Likewise.
4318 (pj4_alu_shift): Likewise.
4319 (pj4_alu_shift_conds): Likewise.
4320 * config/aarch64/aarch64.md: Update for attribute change
4321 all occurrences of arlo_* and shift* types.
4323 2013-09-05 Mike Stump <mikestump@comcast.net>
4325 * tree.h: Move documentation for tree_function_decl to tree-core.h
4326 with the declaration.
4328 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
4331 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
4332 looking for widest mode.
4334 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
4336 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
4338 2013-09-05 Richard Biener <rguenther@suse.de>
4340 PR tree-optimization/58137
4341 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
4342 Do not create vectors of pointers.
4343 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
4344 types for the components of the vector initializer.
4345 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
4346 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
4348 2013-09-05 Martin Jambor <mjambor@suse.cz>
4350 * ipa-prop.c (remove_described_reference): Accept missing references,
4351 return false if that hppens, otherwise return true.
4352 (cgraph_node_for_jfunc): New function.
4353 (try_decrement_rdesc_refcount): Likewise.
4354 (try_make_edge_direct_simple_call): Use them.
4355 (ipa_edge_removal_hook): Remove references from rdescs.
4356 (ipa_edge_duplication_hook): Clone rdescs and their references
4357 when the new edge has the same caller as the old one.
4358 * cgraph.c (cgraph_resolve_speculation): Remove speculative
4359 reference before removing any edges.
4361 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
4363 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
4365 * thumb2.md (thumb2_storewb_parisi): New pattern.
4367 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
4369 * config/aarch64/aarch64-option-extensions.def: Add
4370 AARCH64_OPT_EXTENSION of 'crc'.
4371 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
4372 (AARCH64_ISA_CRC): Ditto.
4373 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
4374 description of the CRC extension.
4376 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
4378 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
4379 * config/rs6000/linux.h: Ditto.
4380 * alpha/linux.h: Ditto.
4381 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
4382 no_c99_libc_has_function.
4383 * config/c6x/uclinux-elf.h: Ditto.
4384 * config/lm32/uclinux-elf.h: Ditto.
4385 * config/m68k/uclinux.h: Ditto.
4386 * config/moxie/uclinux.h: Ditto.
4387 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
4388 (crisv32-*-linux*, cris-*-linux*): Ditto.
4389 * config/bfin/bfin.c: Include "tm_p.h".
4391 2013-09-05 Richard Biener <rguenther@suse.de>
4393 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
4394 check for a definition without a basic-block.
4396 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
4397 Sofiane Naci <sofiane.naci@arm.com>
4399 * config/aarch64/aarch64.md
4400 (*movti_aarch64): Rename r_2_f and f_2_r.
4401 (*movsf_aarch64): Likewise.
4402 (*movdf_aarch64): Likewise.
4403 (*movtf_aarch64): Likewise.
4404 (aarch64_movdi_<mode>low): Likewise.
4405 (aarch64_movdi_<mode>high): Likewise.
4406 (aarch64_mov<mode>high_di): Likewise.
4407 (aarch64_mov<mode>low_di): Likewise.
4408 (aarch64_movtilow_tilow): Likewise.
4409 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
4410 values to config/arm/types.md
4411 (attribute "conds"): Update for attribute change.
4412 (anddi3_insn): Likewise.
4413 (iordi3_insn): Likewise.
4414 (xordi3_insn): Likewise.
4415 (one_cmpldi2): Likewise.
4416 * config/arm/types.md (type): Add Neon types.
4417 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
4418 use "type" attribute.
4419 (movmisalign<mode>_neon_store): Likewise.
4420 (movmisalign<mode>_neon_load): Likewise.
4421 (vec_set<mode>_internal): Likewise.
4422 (vec_setv2di_internal): Likewise.
4423 (vec_extract<mode>): Likewise.
4424 (vec_extractv2di): Likewise.
4425 (add<mode>3_neon): Likewise.
4426 (adddi3_neon): Likewise.
4427 (sub<mode>3_neon): Likewise.
4428 (subdi3_neon): Likewise.
4429 (mul<mode>3_neon): Likewise.
4430 (mul<mode>3add<mode>_neon): Likewise.
4431 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
4432 (fma<VCVTF:mode>4)): Likewise.
4433 (fma<VCVTF:mode>4_intrinsic): Likewise.
4434 (fmsub<VCVTF:mode>4)): Likewise.
4435 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
4436 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
4437 (ior<mode>3): Likewise.
4438 (and<mode>3): Likewise.
4439 (anddi3_neon): Likewise.
4440 (orn<mode>3_neon): Likewise.
4441 (orndi3_neon): Likewise.
4442 (bic<mode>3_neon): Likewise.
4443 (bicdi3_neon): Likewise.
4444 (xor<mode>3): Likewise.
4445 (one_cmpl<mode>2): Likewise.
4446 (abs<mode>2): Likewise.
4447 (neg<mode>2): Likewise.
4448 (umin<mode>3_neon): Likewise.
4449 (umax<mode>3_neon): Likewise.
4450 (smin<mode>3_neon): Likewise.
4451 (smax<mode>3_neon): Likewise.
4452 (vashl<mode>3): Likewise.
4453 (vashr<mode>3_imm): Likewise.
4454 (vlshr<mode>3_imm): Likewise.
4455 (ashl<mode>3_signed): Likewise.
4456 (ashl<mode>3_unsigned): Likewise.
4457 (neon_load_count): Likewise.
4458 (ashldi3_neon_noclobber): Likewise.
4459 (signed_shift_di3_neon): Likewise.
4460 (unsigned_shift_di3_neon): Likewise.
4461 (ashrdi3_neon_imm_noclobber): Likewise.
4462 (lshrdi3_neon_imm_noclobber): Likewise.
4463 (widen_ssum<mode>3): Likewise.
4464 (widen_usum<mode>3): Likewise.
4465 (quad_halves_<code>v4si): Likewise.
4466 (quad_halves_<code>v4sf): Likewise.
4467 (quad_halves_<code>v8hi): Likewise.
4468 (quad_halves_<code>v16qi): Likewise.
4469 (reduc_splus_v2di): Likewise.
4470 (neon_vpadd_internal<mode>): Likewise.
4471 (neon_vpsmin<mode>): Likewise.
4472 (neon_vpsmax<mode>): Likewise.
4473 (neon_vpumin<mode>): Likewise.
4474 (neon_vpumax<mode>): Likewise.
4475 (ss_add<mode>_neon): Likewise.
4476 (us_add<mode>_neon): Likewise.
4477 (ss_sub<mode>_neon): Likewise.
4478 (us_sub<mode>_neon): Likewise.
4479 (neon_vadd<mode>_unspec): Likewise.
4480 (neon_vaddl<mode>): Likewise.
4481 (neon_vaddw<mode>): Likewise.
4482 (neon_vhadd<mode>): Likewise.
4483 (neon_vqadd<mode>): Likewise.
4484 (neon_vaddhn<mode>): Likewise.
4485 (neon_vmul<mode>): Likewise.
4486 (neon_vmla<mode>): Likewise.
4487 (neon_vmlal<mode>): Likewise.
4488 (neon_vmls<mode>): Likewise.
4489 (neon_vmlsl<mode>): Likewise.
4490 (neon_vqdmulh<mode>): Likewise.
4491 (neon_vqdmlal<mode>): Likewise.
4492 (neon_vqdmlsl<mode>): Likewise.
4493 (neon_vmull<mode>): Likewise.
4494 (neon_vqdmull<mode>): Likewise.
4495 (neon_vsub<mode>_unspec): Likewise.
4496 (neon_vsubl<mode>): Likewise.
4497 (neon_vsubw<mode>): Likewise.
4498 (neon_vqsub<mode>): Likewise.
4499 (neon_vhsub<mode>): Likewise.
4500 (neon_vsubhn<mode>): Likewise.
4501 (neon_vceq<mode>): Likewise.
4502 (neon_vcge<mode>): Likewise.
4503 (neon_vcgeu<mode>): Likewise.
4504 (neon_vcgt<mode>): Likewise.
4505 (neon_vcgtu<mode>): Likewise.
4506 (neon_vcle<mode>): Likewise.
4507 (neon_vclt<mode>): Likewise.
4508 (neon_vcage<mode>): Likewise.
4509 (neon_vcagt<mode>): Likewise.
4510 (neon_vtst<mode>): Likewise.
4511 (neon_vabd<mode>): Likewise.
4512 (neon_vabdl<mode>): Likewise.
4513 (neon_vaba<mode>): Likewise.
4514 (neon_vabal<mode>): Likewise.
4515 (neon_vmax<mode>): Likewise.
4516 (neon_vmin<mode>): Likewise.
4517 (neon_vpaddl<mode>): Likewise.
4518 (neon_vpadal<mode>): Likewise.
4519 (neon_vpmax<mode>): Likewise.
4520 (neon_vpmin<mode>): Likewise.
4521 (neon_vrecps<mode>): Likewise.
4522 (neon_vrsqrts<mode>): Likewise.
4523 (neon_vqabs<mode>): Likewise.
4524 (neon_vqneg<mode>): Likewise.
4525 (neon_vcls<mode>): Likewise.
4526 (clz<mode>2): Likewise.
4527 (popcount<mode>2): Likewise.
4528 (neon_vrecpe): Likewise.
4529 (neon_vrsqrte): Likewise.
4530 (neon_vget_lane<mode>_sext_internal): Likewise.
4531 (neon_vget_lane<mode>_zext_internal): Likewise.
4532 (neon_vdup_n<mode>): Likewise.
4533 (neon_vdup_nv2di): Likewise.
4534 (neon_vdpu_lane<mode>_internal): Likewise.
4535 (neon_vswp<mode>): Likewise.
4536 (float<mode><V_cvtto>2): Likewise.
4537 (floatuns<mode><V_cvtto>2): Likewise.
4538 (fix_trunc<mode><V_cvtto>)2): Likewise
4539 (fixuns_trunc<mode><V_cvtto)2): Likewise.
4540 (neon_vcvt<mode>): Likewise.
4541 (neon_vcvtv4sfv4hf): Likewise.
4542 (neon_vcvtv4hfv4sf): Likewise.
4543 (neon_vcvt_n<mode>): Likewise.
4544 (neon_vmovn<mode>): Likewise.
4545 (neon_vqmovn<mode>): Likewise.
4546 (neon_vqmovun<mode>): Likewise.
4547 (neon_vmovl<mode>): Likewise.
4548 (neon_vmul_lane<mode>): Likewise.
4549 (neon_vmull_lane<mode>): Likewise.
4550 (neon_vqdmull_lane<mode>): Likewise.
4551 (neon_vqdmulh_lane<mode>): Likewise.
4552 (neon_vmla_lane<mode>): Likewise.
4553 (neon_vmlal_lane<mode>): Likewise.
4554 (neon_vqdmlal_lane<mode>): Likewise.
4555 (neon_vmls_lane<mode>): Likewise.
4556 (neon_vmlsl_lane<mode>): Likewise.
4557 (neon_vqdmlsl_lane<mode>): Likewise.
4558 (neon_vext<mode>): Likewise.
4559 (neon_vrev64<mode>): Likewise.
4560 (neon_vrev32<mode>): Likewise.
4561 (neon_vrev16<mode>): Likewise.
4562 (neon_vbsl<mode>_internal): Likewise.
4563 (neon_vshl<mode>): Likewise.
4564 (neon_vqshl<mode>): Likewise.
4565 (neon_vshr_n<mode>): Likewise.
4566 (neon_vshrn_n<mode>): Likewise.
4567 (neon_vqshrn_n<mode>): Likewise.
4568 (neon_vqshrun_n<mode>): Likewise.
4569 (neon_vshl_n<mode>): Likewise.
4570 (neon_vqshl_n<mode>): Likewise.
4571 (neon_vqshlu_n<mode>): Likewise.
4572 (neon_vshll_n<mode>): Likewise.
4573 (neon_vsra_n<mode>): Likewise.
4574 (neon_vsri_n<mode>): Likewise.
4575 (neon_vsli_n<mode>): Likewise.
4576 (neon_vtbl1v8qi): Likewise.
4577 (neon_vtbl2v8qi): Likewise.
4578 (neon_vtbl3v8qi): Likewise.
4579 (neon_vtbl4v8qi): Likewise.
4580 (neon_vtbx1v8qi): Likewise.
4581 (neon_vtbx2v8qi): Likewise.
4582 (neon_vtbx3v8qi): Likewise.
4583 (neon_vtbx4v8qi): Likewise.
4584 (neon_vtrn<mode>_internal): Likewise.
4585 (neon_vzip<mode>_internal): Likewise.
4586 (neon_vuzp<mode>_internal): Likewise.
4587 (neon_vld1<mode>): Likewise.
4588 (neon_vld1_lane<mode>): Likewise.
4589 (neon_vld1_dup<mode>): Likewise.
4590 (neon_vld1_dupv2di): Likewise.
4591 (neon_vst1<mode>): Likewise.
4592 (neon_vst1_lane<mode>): Likewise.
4593 (neon_vld2<mode>): Likewise.
4594 (neon_vld2_lane<mode>): Likewise.
4595 (neon_vld2_dup<mode>): Likewise.
4596 (neon_vst2<mode>): Likewise.
4597 (neon_vst2_lane<mode>): Likewise.
4598 (neon_vld3<mode>): Likewise.
4599 (neon_vld3qa<mode>): Likewise.
4600 (neon_vld3qb<mode>): Likewise.
4601 (neon_vld3_lane<mode>): Likewise.
4602 (neon_vld3_dup<mode>): Likewise.
4603 (neon_vst3<mode>): Likewise.
4604 (neon_vst3qa<mode>): Likewise.
4605 (neon_vst3qb<mode>): Likewise.
4606 (neon_vst3_lane<mode>): Likewise.
4607 (neon_vld4<mode>): Likewise.
4608 (neon_vld4qa<mode>): Likewise.
4609 (neon_vld4qb<mode>): Likewise.
4610 (neon_vld4_lane<mode>): Likewise.
4611 (neon_vld4_dup<mode>): Likewise.
4612 (neon_vst4<mode>): Likewise.
4613 (neon_vst4qa<mode>): Likewise.
4614 (neon_vst4qb<mode>): Likewise.
4615 (neon_vst4_lane<mode>): Likewise.
4616 (neon_vec_unpack<US>_lo_<mode>): Likewise.
4617 (neon_vec_unpack<US>_hi_<mode>): Likewise.
4618 (neon_vec_<US>mult_lo_<mode>): Likewise.
4619 (neon_vec_<US>mult_hi_<mode>): Likewise.
4620 (neon_vec_<US>shiftl_<mode>): Likewise.
4621 (neon_unpack<US>_<mode>): Likewise.
4622 (neon_vec_<US>mult_<mode>): Likewise.
4623 (vec_pack_trunc_<mode>): Likewise.
4624 (neon_vec_pack_trunk_<mode>): Likewise.
4625 (neon_vabd<mode>_2): Likewise.
4626 (neon_vabd<mode>_3): Likewise.
4627 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
4628 (thumb2_movsi_vfp): Likewise.
4629 (movdi_vfp): Likewise.
4630 (movdi_vfp_cortexa8): Likewise.
4631 (movhf_vfp_neon): Likewise.
4632 (movhf_vfp): Likewiwse.
4633 (movsf_vfp): Likewiwse.
4634 (thumb2_movsf_vfp): Likewiwse.
4635 (movdf_vfp): Likewise.
4636 (thumb2_movdf_vfp): Likewise.
4637 (movsfcc_vfp): Likewise.
4638 (thumb2_movsfcc_vfp): Likewise.
4639 (movdfcc_vfp): Likewise.
4640 (thumb2_movdfcc_vfp): Likewise.
4641 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
4642 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
4643 (v10_v2c): Likewise.
4644 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
4646 (cortex_a15_neon_int_2): Likewise.
4647 (cortex_a15_neon_int_3): Likewise.
4648 (cortex_a15_neon_int_4): Likewise.
4649 (cortex_a15_neon_int_5): Likewise.
4650 (cortex_a15_neon_vqneg_vqabs): Likewise.
4651 (cortex_a15_neon_vmov): Likewise.
4652 (cortex_a15_neon_vaba): Likewise.
4653 (cortex_a15_neon_vaba_qqq): Likewise.
4654 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4655 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4656 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
4657 scalar_64_32_long_scalar): Likewise.
4658 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4659 (cortex_a15_neon_mla_qqq_8_16): Likewise.
4660 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
4661 lotype_qdd_64_32_long): Likewise.
4662 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4663 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4664 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
4665 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4666 (cortex_a15_neon_shift_1): Likewise.
4667 (cortex_a15_neon_shift_2): Likewise.
4668 (cortex_a15_neon_shift_3): Likewise.
4669 (cortex_a15_neon_vshl_ddd): Likewise.
4670 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4671 (cortex_a15_neon_vsra_vrsra): Likewise.
4672 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
4673 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
4674 (cortex_a15_neon_fp_vmul_ddd): Likewise.
4675 (cortex_a15_neon_fp_vmul_qqd): Likewise.
4676 (cortex_a15_neon_fp_vmla_ddd): Likewise.
4677 (cortex_a15_neon_fp_vmla_qqq): Likewise.
4678 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
4679 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
4680 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4681 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4682 (cortex_a15_neon_bp_simple): Likewise.
4683 (cortex_a15_neon_bp_2cycle): Likewise.
4684 (cortex_a15_neon_bp_3cycle): Likewise.
4685 (cortex_a15_neon_vld1_1_2_regs): Likewise.
4686 (cortex_a15_neon_vld1_3_4_regs): Likewise.
4687 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4688 (cortex_a15_neon_vld2_4_regs): Likewise.
4689 (cortex_a15_neon_vld3_vld4): Likewise.
4690 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4691 (cortex_a15_neon_vst1_3_4_regs): Likewise.
4692 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
4693 (cortex_a15_neon_vst3_vst4): Likewise.
4694 (cortex_a15_neon_vld1_vld2_lane): Likewise.
4695 (cortex_a15_neon_vld3_vld4_lane" 10
4696 (cortex_a15_neon_vst1_vst2_lane): Likewise.
4697 (cortex_a15_neon_vst3_vst4_lane): Likewise.
4698 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
4699 (cortex_a15_neon_ldm_2): Likewise.0
4700 (cortex_a15_neon_stm_2): Likewise.
4701 (cortex_a15_neon_mcr): Likewise.
4702 (cortex_a15_neon_mcr_2_mcrr): Likewise.
4703 (cortex_a15_neon_mrc): Likewise.
4704 (cortex_a15_neon_mrrc): Likewise.
4705 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
4707 (cortex_a15_alu_shift): Likewise.
4708 (cortex_a15_alu_shift_reg): Likewise.
4709 (cortex_a15_mult32): Likewise.
4710 (cortex_a15_mult64): Likewise.
4711 (cortex_a15_block): Likewise.
4712 (cortex_a15_branch): Likewise.
4713 (cortex_a15_load1): Likewise.
4714 (cortex_a15_load3): Likewise.
4715 (cortex_a15_store1): Likewise.
4716 (cortex_a15_store3): Likewise.
4717 (cortex_a15_call): Likewise.
4718 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
4719 (cortex_a5_f2r): Likewise.
4720 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
4722 (cortex_a53_f2r): Likewise.
4723 * config/arm/cortex-a7.md
4724 (cortex_a7_branch): Update for attribute change.
4725 (cortex_a7_call): Likewise.
4726 (cortex_a7_alu_imm): Likewise.
4727 (cortex_a7_alu_reg): Likewise.
4728 (cortex_a7_alu_shift): Likewise.
4729 (cortex_a7_mul): Likewise.
4730 (cortex_a7_load1): Likewise.
4731 (cortex_a7_store1): Likewise.
4732 (cortex_a7_load2): Likewise.
4733 (cortex_a7_store2): Likewise.
4734 (cortex_a7_load3): Likewise.
4735 (cortex_a7_store3): Likewise.
4736 (cortex_a7_load4): Likewise.
4737 (cortex_a7_store4): Likewise.
4738 (cortex_a7_fpalu): Likewise.
4739 (cortex_a7_fconst): Likewise.
4740 (cortex_a7_fpmuls): Likewise.
4741 (cortex_a7_neon_mul): Likewise.
4742 (cortex_a7_fpmacs): Likewise.
4743 (cortex_a7_neon_mla: Likewise.
4744 (cortex_a7_fpmuld: Likewise.
4745 (cortex_a7_fpmacd: Likewise.
4746 (cortex_a7_fpfmad: Likewise.
4747 (cortex_a7_fdivs: Likewise.
4748 (cortex_a7_fdivd: Likewise.
4749 (cortex_a7_r2f: Likewise.
4750 (cortex_a7_f2r: Likewise.
4751 (cortex_a7_f_flags: Likewise.
4752 (cortex_a7_f_loads: Likewise.
4753 (cortex_a7_f_loadd: Likewise.
4754 (cortex_a7_f_stores: Likewise.
4755 (cortex_a7_f_stored: Likewise.
4756 (cortex_a7_neon): Likewise.
4757 * config/arm/cortex-a8-neon.md
4758 (cortex_a8_neon_mrc): Update for attribute change.
4759 (cortex_a8_neon_mrrc): Likewise.
4760 (cortex_a8_neon_int_1): Likewise.
4761 (cortex_a8_neon_int_2): Likewise.
4762 (cortex_a8_neon_int_3): Likewise.
4763 (cortex_a8_neon_int_4): Likewise.
4764 (cortex_a8_neon_int_5): Likewise.
4765 (cortex_a8_neon_vqneg_vqabs): Likewise.
4766 (cortex_a8_neon_vmov): Likewise.
4767 (cortex_a8_neon_vaba): Likewise.
4768 (cortex_a8_neon_vaba_qqq): Likewise.
4769 (cortex_a8_neon_vsma): Likewise.
4770 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4771 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4772 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
4773 long_scalar): Likewise.
4774 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4775 (cortex_a8_neon_mla_qqq_8_16): Likewise.
4776 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
4777 long_scalar_qdd_64_32_long): Likewise.
4778 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4779 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4780 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
4781 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4782 (cortex_a8_neon_shift_1): Likewise.
4783 (cortex_a8_neon_shift_2): Likewise.
4784 (cortex_a8_neon_shift_3): Likewise.
4785 (cortex_a8_neon_vshl_ddd): Likewise.
4786 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4787 (cortex_a8_neon_vsra_vrsra): Likewise.
4788 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
4789 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
4790 (cortex_a8_neon_fp_vsum): Likewise.
4791 (cortex_a8_neon_fp_vmul_ddd): Likewise.
4792 (cortex_a8_neon_fp_vmul_qqd): Likewise.
4793 (cortex_a8_neon_fp_vmla_ddd): Likewise.
4794 (cortex_a8_neon_fp_vmla_qqq): Likewise.
4795 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
4796 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
4797 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4798 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4799 (cortex_a8_neon_bp_simple): Likewise.
4800 (cortex_a8_neon_bp_2cycle): Likewise.
4801 (cortex_a8_neon_bp_3cycle): Likewise.
4802 (cortex_a8_neon_ldr): Likewise.
4803 (cortex_a8_neon_str): Likewise.
4804 (cortex_a8_neon_vld1_1_2_regs): Likewise.
4805 (cortex_a8_neon_vld1_3_4_regs): Likewise.
4806 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4807 (cortex_a8_neon_vld2_4_regs): Likewise.
4808 (cortex_a8_neon_vld3_vld4): Likewise.
4809 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4810 (cortex_a8_neon_vst1_3_4_regs): Likewise.
4811 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
4812 (cortex_a8_neon_vst3_vst4): Likewise.
4813 (cortex_a8_neon_vld1_vld2_lane): Likewise.
4814 (cortex_a8_neon_vld3_vld4_lane): Likewise.
4815 (cortex_a8_neon_vst1_vst2_lane): Likewise.
4816 (cortex_a8_neon_vst3_vst4_lane): Likewise.
4817 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
4818 (cortex_a8_neon_mcr): Likewise.
4819 (cortex_a8_neon_mcr_2_mcrr): Likewise.
4820 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
4821 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
4823 (ca9_neon_mrrc): Likewise.
4824 (cortex_a9_neon_int_1): Likewise.
4825 (cortex_a9_neon_int_2): Likewise.
4826 (cortex_a9_neon_int_3): Likewise.
4827 (cortex_a9_neon_int_4): Likewise.
4828 (cortex_a9_neon_int_5): Likewise.
4829 (cortex_a9_neon_vqneg_vqabs): Likewise.
4830 (cortex_a9_neon_vmov): Likewise.
4831 (cortex_a9_neon_vaba): Likewise.
4832 (cortex_a9_neon_vaba_qqq): Likewise.
4833 (cortex_a9_neon_vsma): Likewise.
4834 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4835 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4836 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
4837 long_scalar): Likewise.
4838 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4839 (cortex_a9_neon_mla_qqq_8_16): Likewise.
4840 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
4841 long_scalar_qdd_64_32_long): Likewise.
4842 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4843 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4844 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
4845 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4846 (cortex_a9_neon_shift_1): Likewise.
4847 (cortex_a9_neon_shift_2): Likewise.
4848 (cortex_a9_neon_shift_3): Likewise.
4849 (cortex_a9_neon_vshl_ddd): Likewise.
4850 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4851 (cortex_a9_neon_vsra_vrsra): Likewise.
4852 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
4853 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
4854 (cortex_a9_neon_fp_vsum): Likewise.
4855 (cortex_a9_neon_fp_vmul_ddd): Likewise.
4856 (cortex_a9_neon_fp_vmul_qqd): Likewise.
4857 (cortex_a9_neon_fp_vmla_ddd): Likewise.
4858 (cortex_a9_neon_fp_vmla_qqq): Likewise.
4859 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
4860 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
4861 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4862 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4863 (cortex_a9_neon_bp_simple): Likewise.
4864 (cortex_a9_neon_bp_2cycle): Likewise.
4865 (cortex_a9_neon_bp_3cycle): Likewise.
4866 (cortex_a9_neon_ldr): Likewise.
4867 (cortex_a9_neon_str): Likewise.
4868 (cortex_a9_neon_vld1_1_2_regs): Likewise.
4869 (cortex_a9_neon_vld1_3_4_regs): Likewise.
4870 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4871 (cortex_a9_neon_vld2_4_regs): Likewise.
4872 (cortex_a9_neon_vld3_vld4): Likewise.
4873 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4874 (cortex_a9_neon_vst1_3_4_regs): Likewise.
4875 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
4876 (cortex_a9_neon_vst3_vst4): Likewise.
4877 (cortex_a9_neon_vld1_vld2_lane): Likewise.
4878 (cortex_a9_neon_vld3_vld4_lane): Likewise.
4879 (cortex_a9_neon_vst1_vst2_lane): Likewise.
4880 (cortex_a9_neon_vst3_vst4_lane): Likewise.
4881 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
4882 (cortex_a9_neon_mcr): Likewise.
4883 (cortex_a9_neon_mcr_2_mcrr): Likewise.
4884 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
4885 (cortex_a9_fps): Likewise.
4886 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
4888 (cortex_m4_fmuls): Likewise.
4889 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
4891 (cortex_r4_mrc): Likewise.
4892 * config/arm/iterators.md: Update comment referring to neon_type.
4893 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
4894 (iwmmxt_movsi_insn): Likewise.
4895 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
4897 (pj4_core_to_vfp): Likewise.
4898 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
4900 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
4901 (vfp_fstore): Likewise.
4902 * doc/md.texi: Change references to neon_type to refer to type.
4904 2013-09-04 Dodji Seketeli <dodji@redhat.com>
4906 * tree.h (DECL_BUILT_IN): Fix typo in comment.
4908 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
4910 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
4913 2013-09-04 Easwaran Raman <eraman@google.com>
4916 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
4917 (build_and_add_sum): Use it.
4918 (appears_later_in_bb): Simplify code.
4920 2013-09-04 Teresa Johnson <tejohnson@google.com>
4922 * dumpfile.c (dump_finish): Don't close stderr/stdout.
4924 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
4926 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
4928 2013-09-04 Jan Hubicka <jh@suse.cz>
4930 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
4931 * ipa-devirt.c: Include diganostic.h
4932 (odr_type_d): Add types and types_set.
4933 (hash_type_name): Work for types with vtables during LTO.
4934 (odr_hasher::remove): Fix comment; destroy types_set.
4935 (add_type_duplicate): New function,
4936 (get_odr_type): Use it.
4937 (dump_type_inheritance_graph): Dump type duplicates.
4938 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
4940 * tree.c (types_same_for_odr): Give exact answers on types with
4943 2013-09-04 Dodji Seketeli <dodji@redhat.com>
4945 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
4946 explaining their differences.
4948 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
4950 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
4952 2013-09-03 Jeff Law <law@redhat.com>
4954 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
4955 when not threading through a joiner block. Pass joiner/no joiner
4956 state to register_jump_thread.
4957 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
4958 state from argument rather than implying on path length.
4959 Dump the entire jump thread path into debugging dump.
4960 * tree-flow.h (register_jump_thread): Update prototype.
4962 2013-08-29 Xinliang David Li <davidxl@google.com>
4964 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
4965 Remove a trivial gcc_assert.
4967 2013-08-29 Xinliang David Li <davidxl@google.com>
4969 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
4970 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
4971 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
4972 Delay base decl alignment adjustment.
4973 * tree-vectorizer.c (vect_destroy_datarefs): New function.
4974 * tree-vectorizer.h: New data structure.
4975 (set_dr_misalignment): New function.
4976 (dr_misalignment): Ditto.
4977 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
4978 (vectorizable_load): Ditto.
4979 (ensure_base_align): New function.
4980 (vectorize_loops): Add dbg_cnt support.
4981 (execute_vect_slp): Ditto.
4982 * dbgcnt.def: New debug counter.
4983 * Makefile: New dependency.
4985 2013-09-03 Meador Inge <meadori@codesourcery.com>
4989 2013-08-30 Meador Inge <meadori@codesourcery.com>
4991 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
4993 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
4995 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
4996 function descriptor.
4998 2013-09-03 Richard Biener <rguenther@suse.de>
5000 * tree-affine.c (add_elt_to_tree): Fix association issue,
5001 avoid useless converts and make sure to always return a
5002 properly typed result.
5004 2013-09-03 Richard Biener <rguenther@suse.de>
5007 * fold-const.c (negate_expr_p): Fix division case.
5008 (negate_expr): Likewise.
5010 2013-09-03 Richard Biener <rguenther@suse.de>
5013 * tree-streamer-out.c: Include tm.h.
5014 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
5016 2013-09-03 Jan Hubicka <jh@suse.cz>
5018 * tree-profile.c (tree_profiling): Cleanup CFG when done.
5020 2013-09-03 Alan Modra <amodra@gmail.com>
5022 * config.gcc (powerpc*-*-linux*): Add support for little-endian
5023 multilibs to big-endian target and vice versa.
5024 * config/rs6000/t-linux64: Use := assignment on all vars.
5025 (MULTILIB_EXTRA_OPTS): Remove fPIC.
5026 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
5027 * config/rs6000/t-linux64le: New file.
5028 * config/rs6000/t-linux64bele: New file.
5029 * config/rs6000/t-linux64lebe: New file.
5031 2013-09-02 Jan Hubicka <jh@suse.cz>
5033 * ipa-inline-transform.c (inline_transform): Do not
5034 optimize_inline_calls when not optimizing.
5036 2013-09-02 Jan Hubicka <jh@suse.cz>
5038 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
5039 duplicated nodes for assembler names.
5040 * symtab.c (symtab_unregister_node): Do not attempt to unlink
5041 hard registers from assembler name hash.
5043 2013-09-02 Jan Hubicka <jh@suse.cz>
5045 * ipa-split.c (execute_split_functions): Split externally visible
5046 functions called once.
5048 2013-09-02 Martin Jambor <mjambor@suse.cz>
5051 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
5052 linked list. When finding the correct duplicate, also consider also
5053 the caller in additon to its inlined_to node.
5055 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
5057 * config/aarch64/aarch64-simd-builtins.def
5058 (dup_lane_scalar): Remove.
5059 * config/aarch64/aarch64-simd.md
5060 (aarch64_simd_dup): Add 'w->w' alternative.
5061 (aarch64_dup_lane<mode>): Allow for VALL.
5062 (aarch64_dup_lane_scalar<mode>): Remove.
5063 (aarch64_dup_lane_<vswap_width_name><mode>): New.
5064 (aarch64_get_lane_signed<mode>): Add w->w altenative.
5065 (aarch64_get_lane_unsigned<mode>): Likewise.
5066 (aarch64_get_lane<mode>): Likewise.
5067 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
5068 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
5069 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
5070 (VCON): Change container of V2SF.
5071 (vswap_width_name): Likewise.
5072 * config/aarch64/arm_neon.h
5073 (__aarch64_vdup_lane_any): New.
5074 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
5075 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
5076 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
5078 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
5081 * expr.c (emit_move_complex): Do not move complex FP values as parts if
5082 the source or the destination is a single hard register.
5084 2013-09-02 Richard Biener <rguenther@suse.de>
5087 * tree-scalar-evolution.c (instantiate_scev_name): Allow
5090 2013-09-02 Richard Biener <rguenther@suse.de>
5092 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
5093 arithmetic to sizetype.
5095 2013-09-02 Bin Cheng <bin.cheng@arm.com>
5097 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
5098 Find auto-increment use both before and after candidate.
5100 2013-09-02 Marek Polacek <polacek@redhat.com>
5102 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
5104 2013-09-01 Jan Hubicka <jh@suse.cz>
5106 * Makefile.in: Add ipa-profile.o
5107 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
5108 * cgraph.c (struct cgraph_propagate_frequency_data,
5109 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
5110 ipa-profile.c; replace cgraph_ by ipa_ prefix.
5111 * cgraph.h (cgraph_propagate_frequency): Remove.
5112 * ipa-inline-analysis.c: Include ipa-utils.h;
5113 drop duplicated cfgloop.h.
5114 (inline_update_callee_summaries): Update.
5115 * ipa-profile.c: New file.
5116 * ipa-utils.h (ipa_propagate_frequency): Declare.
5117 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
5118 data-streamer.h, value-prof.h.
5119 (symtab_remove_unreachable_nodes): Update profile.
5120 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
5121 account_time_size, cmp_counts, dump_histogram,
5122 ipa_profile_generate_summary, ipa_profile_write_summary,
5123 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
5124 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
5125 Move to ipa-profile.c.
5127 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
5129 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
5131 2013-09-01 Jan Hubicka <jh@suse.cz>
5133 * common.opt (fdevirtualize-speculatively): New function.
5134 * invoke.texi (fdevirtualize-speculatively): Document.
5135 * ipa-devirt.c: Include ipa-inline.h
5136 (likely_target_p): New function.
5137 (ipa_devirt): New function.
5138 (gate_ipa_devirt): New function.
5139 (pass_data_ipa_devirt): New static var.
5140 (pass_ipa_devirt): Likewise.
5141 (make_pass_ipa_devirt): New function.
5142 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
5143 (common_handle_option): Disable devirtualization when
5144 value range profiling is available.
5145 * passes.def (pass_ipa_devirt): Add.
5146 * timever.def (TV_IPA_DEVIRT): New timevar.
5147 * tree-pass.h (make_pass_ipa_devirt):
5149 2013-09-01 Iain Sandoe <iain@codesourcery.com>
5151 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
5152 include sanitize(undefined).
5154 2013-08-31 Diego Novillo <dnovillo@google.com>
5156 * Makefile.in (TREE_CORE_H): Define.
5158 (GTFILES): Add tree-core.h.
5159 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
5161 * tree-core.h: New file.
5162 Move all data structures, enum, typedefs, global
5163 declarations and constants from ...
5166 2013-08-31 Jan Hubicka <jh@suse.cz>
5168 * bulitins.c (expand_builtin): Do not early exit for gcov
5169 instrumented functions.
5171 2013-08-31 Marek Polacek <polacek@redhat.com>
5173 * ubsan.c: Include tm_p.h.
5175 2013-08-31 Jan Hubicka <jh@suse.cz>
5177 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
5180 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
5181 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
5182 * tree-cfg.c (verify_gimple_label): ... here.
5183 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
5184 (ipa_merge_profiles): New function.
5185 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
5186 (lto_input_function_body): Likewise.
5187 * ipa-utils.h (ipa_merge_profiles): Declare.
5188 * lto-streamer.h (lto_input_function_body): Update prototype.
5189 (emit_label_in_global_context_p): Remove.
5190 * lto-symtab.c: Include ipa-utils.h
5191 (lto_cgraph_replace_node): Use ipa_merge_profiles.
5193 2013-08-31 Jan Hubicka <jh@suse.cz>
5195 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
5197 2013-08-31 Jan Hubicka <jh@suse.cz>
5199 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
5201 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
5203 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
5204 "cmp" RTX before signed_comparison_operator check to account
5207 2013-08-30 Jan Hubicka <jh@suse.cz>
5209 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
5210 (detect_type_change_1): Rename to ...
5211 (detect_type_change): ... this one; early return on non-polymorphic
5213 (detect_type_change_ssa): Add comp_type parameter; update
5214 use of detect_type_change.
5215 (compute_complex_assign_jump_func): Add param_type parameter;
5216 update use of detect_type_change_ssa.
5217 (compute_complex_ancestor_jump_func): Likewise.
5218 (ipa_get_callee_param_type): New function.
5219 (ipa_compute_jump_functions_for_edge): Compute parameter type;
5220 update calls to the jump function computation functions.
5222 2013-08-30 Teresa Johnson <tejohnson@google.com>
5223 Steven Bosscher <steven@gcc.gnu.org>
5225 * cfgrtl.c (fixup_new_cold_bb): New routine.
5226 (commit_edge_insertions): Invoke fixup_partitions.
5227 (find_partition_fixes): New routine.
5228 (fixup_partitions): Ditto.
5229 (verify_hot_cold_block_grouping): Update comments.
5230 (rtl_verify_edges): Invoke find_partition_fixes.
5231 (rtl_verify_bb_pointers): Update comments.
5232 (rtl_verify_bb_layout): Ditto.
5233 * basic-block.h (probably_never_executed_edge_p): Declare.
5234 (fixup_partitions): Ditto.
5235 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
5236 * bb-reorder.c (sanitize_hot_paths): New function.
5237 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
5239 * predict.c (probably_never_executed_edge_p): New routine.
5240 * cfg.c (check_bb_profile): Add partition insanity warnings.
5242 2013-08-30 Meador Inge <meadori@codesourcery.com>
5244 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
5246 2013-08-30 Marek Polacek <polacek@redhat.com>
5248 * Makefile.in (ubsan.o): Add.
5249 (c-family/c-ubsan.o): Add.
5250 (builtins.o): Add ubsan.h dependency.
5251 * ubsan.h: New file.
5252 * ubsan.c: New file.
5253 * common.opt: Add -fsanitize=undefined option.
5254 (flag_sanitize): Add variable.
5255 (fsanitize=): Add option. Add Driver.
5256 (fsanitize=thread): Remove option.
5257 (fsanitize=address): Likewise.
5258 (static-libubsan): New option.
5259 * doc/invoke.texi: Document the new flag and -static-libubsan.
5260 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
5261 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
5262 * builtin-attrs.def (ATTR_COLD): Define.
5263 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
5264 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
5265 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
5266 * flag-types.h (sanitize_code): New enum.
5267 * opts.c (common_handle_option): Parse command line arguments
5268 of -fsanitize=. Add -fsanitize=unreachable option.
5269 * varasm.c (get_variable_section): Adjust.
5270 (assemble_noswitch_variable): Likewise.
5271 (assemble_variable): Likewise.
5272 (output_constant_def_contents): Likewise.
5273 (categorize_decl_for_section): Likewise.
5274 (place_block_symbol): Likewise.
5275 (output_object_block): Likewise.
5276 * builtins.def: Likewise.
5277 * toplev.c (compile_file): Likewise.
5278 (process_options): Likewise.
5279 * cppbuiltin.c: Likewise.
5280 * tsan.c (tsan_pass): Likewise.
5281 (tsan_gate): Likewise.
5282 (tsan_gate_O0): Likewise.
5283 * cfgexpand.c (partition_stack_vars): Likewise.
5284 (expand_stack_vars): Likewise.
5285 (defer_stack_allocation): Likewise.
5286 (expand_used_vars): Likewise.
5287 * cfgcleanup.c (old_insns_match_p): Likewise.
5288 * asan.c (asan_finish_file): Likewise.
5289 (asan_instrument): Likewise.
5290 (gate_asan): Likewise.
5291 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
5292 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
5293 (asan_global_struct): Use pointer_sized_int_node instead
5294 calling build_nonstandard_integer_type.
5295 (initialize_sanitizer_builtins): Likewise.
5296 (asan_finish_file): Likewise.
5297 * gcc.c: Document %{%:function(args):X}.
5298 (static_spec_functions): Add sanitize.
5299 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
5300 store funcval != NULL there.
5301 (do_spec_1): Adjust handle_spec_function caller.
5302 (handle_braces): Allow %:function(args) as condition.
5303 (sanitize_spec_function): New function.
5304 (ADD_STATIC_LIBUBSAN_LIBS): Define.
5305 (LIBUBSAN_SPEC): Likewise.
5306 (LIBUBSAN_EARLY_SPEC): Likewise.
5307 (SANITIZER_SPEC): Handle libubsan.
5308 (SANITIZER_EARLY_SPEC): Likewise.
5309 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
5310 instead of fsanitize=address.
5311 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
5312 instead of fsanitize=address*.
5313 * builtins.c: Include ubsan.h.
5314 (fold_builtin_0): Instrument __builtin_unreachable.
5315 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
5316 instead of flag_asan.
5317 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
5318 (pointer_sized_int_node): Define.
5319 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
5321 2013-08-30 Mike Stump <mikestump@comcast.net>
5323 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
5326 2013-08-29 Jan Hubicka <jh@suse.cz>
5328 * cgraph.c (cgraph_function_body_availability): Handle weakref
5330 * passes.def: Remove pass_fixup_cfg.
5331 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
5332 track when we need to remove functions.
5333 (gate_ipa_inline): Execute inlining always; add comment why.
5334 (pass_data_ipa_inline): Remove TODO_remove_functions.
5335 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
5336 do not produce summaries.
5337 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
5338 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
5339 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
5340 constant pool and vtable.
5342 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
5344 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
5345 New arm_neon.h's internal macros to specify 64-bit constants.
5346 Avoid using stdint.h's macros.
5348 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
5350 * recog.c (verify_changes): Verify that changes[i].old is non-zero
5351 before applying REG_P.
5353 2013-08-30 Jakub Jelinek <jakub@redhat.com>
5355 PR tree-optimization/58277
5356 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
5357 after seeing too many stmts with vdef in between dombb and current
5358 bb, invalidate everything.
5360 2013-08-30 Richard Biener <rguenther@suse.de>
5362 * fold-const.c (fold_single_bit_test): Fix overflow test.
5364 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
5366 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
5367 and which can live in a register, always retrieve the value on entry.
5368 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
5369 passed by invisible reference specially.
5370 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
5371 (vt_add_function_parameter): Correctly deal with a parameter passed by
5372 invisible reference.
5374 2013-08-30 Jan Hubicka <jh@suse.cz>
5376 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
5378 2013-08-30 Richard Biener <rguenther@suse.de>
5380 PR tree-optimization/58228
5381 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
5382 allow invariant loads in nested loop vectorization.
5384 2013-08-30 Richard Biener <rguenther@suse.de>
5386 PR tree-optimization/58223
5387 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
5388 (has_anti_or_output_dependence): ... this and adjust to also
5389 look for output dependences.
5390 (mark_nodes_having_upstream_mem_writes): Adjust.
5391 (rdg_flag_uses): Likewise.
5393 2013-08-30 Richard Biener <rguenther@suse.de>
5395 PR tree-optimization/58010
5396 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
5397 assert that we have a loop-closed PHI.
5399 2013-08-29 Jan Hubicka <jh@suse.cz>
5401 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
5402 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
5403 * lto-section-in.c (lto_free_function_in_decl_state): New function.
5404 (lto_free_function_in_decl_state_for_node): New function.
5406 2013-08-29 Xinliang David Li <davidxl@google.com>
5408 * loop-unroll.c (report_unroll_peel): Minor message change.
5409 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
5410 Emit alignment peeling message with default -fopt-info.
5411 (vect_loop_versioning): Emit loop version info message.
5412 * tree-vectorizer.c (vectorize_loops): Minor message change.
5413 (execute_vect_slp): Ditto.
5415 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
5417 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
5418 of the clone from the DECL_NAME of the original function.
5420 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
5422 * passes.c (register_pass): Add overload.
5423 * tree-pass.h (register_pass): Forward declare it. Add comment.
5425 2013-08-29 Jan Hubicka <jh@suse.cz>
5427 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
5428 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
5429 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
5430 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
5431 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
5432 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
5433 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
5434 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
5436 2013-08-29 Teresa Johnson <tejohnson@google.com>
5438 * dumpfile.c (dump_loc): Output column number.
5439 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
5440 * doc/invoke.texi: Document optall -fopt-info flag.
5441 * profile.c (read_profile_edge_counts): Use new dump framework.
5442 (compute_branch_probabilities): Ditto.
5443 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
5444 when pass not in any opt group.
5445 * pass_manager.h (pass_manager::get_pass_profile): New method.
5446 * value-prof.c (check_counter): Use new dump framework.
5447 (check_ic_target): Ditto.
5448 * coverage.c (get_coverage_counts): Ditto.
5449 (coverage_init): Setup new dump framework.
5451 2013-08-29 Richard Biener <rguenther@suse.de>
5453 PR tree-optimization/58246
5454 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
5455 handle the dominance check inside a basic-block.
5457 2013-08-29 Richard Biener <rguenther@suse.de>
5460 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
5461 of default defs that appear in abnormal PHI nodes.
5463 2013-08-29 Richard Biener <rguenther@suse.de>
5465 PR tree-optimization/57685
5466 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
5467 single-use operands to avoid exponential complexity.
5469 2013-08-28 Dehao Chen <dehao@google.com>
5471 * ipa-inline.c (edge_badness): Fix integer underflow.
5473 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
5475 * gtm-builtins.def (_ITM_free): Declare leaf.
5477 2013-08-28 Jakub Jelinek <jakub@redhat.com>
5480 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
5481 (*tls_local_dynamic_base_64_largepic): Likewise.
5482 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
5483 Remove predicate from call operand.
5484 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
5485 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
5487 2013-08-28 Jeff Law <law@redhat.com>
5489 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
5490 checks for the number of predecessors and successors allowed.
5491 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
5492 which require copying a joiner block if there is a request which
5493 is a subpath that requires no joiner block copying.
5495 2013-08-28 Jan Hubicka <jh@suse.cz>
5497 * lto-streamer-out.c (DFS_write_tree_body): Drop
5498 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
5499 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
5500 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
5501 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
5502 Do not read DECL_ERROR_ISSUED.
5503 (unpack_ts_decl_with_vis_value_fields): Do not read
5505 (lto_input_ts_binfo_tree_pointers): Do not read
5506 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
5507 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
5508 write DECL_ERROR_ISSUED..
5509 (pack_ts_decl_with_vis_value_fields): Do not write
5511 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
5512 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
5513 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
5514 * tree.h (tree_decl_common): Update comment.
5515 (DECL_ERROR_ISSUED): Remove.
5517 2013-08-28 Jakub Jelinek <jakub@redhat.com>
5520 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
5522 2013-08-28 Jan Hubicka <jh@suse.cz>
5524 * builtins.def (free): Declare leaf.
5526 2013-08-27 David Malcolm <dmalcolm@redhat.com>
5529 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
5530 * configure: Regenerate.
5532 2013-08-27 Martin Jambor <mjambor@suse.cz>
5534 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
5535 (ipa_ancestor_jf_data): Likewise.
5536 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
5537 (ipa_get_jf_pass_through_type_preserved): New function.
5538 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
5539 (ipa_get_jf_ancestor_type_preserved): New function.
5540 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
5541 (ipa_get_jf_ancestor_result): Likewise.
5542 (propagate_vals_accross_pass_through): Use
5543 ipa_get_jf_pass_through_result to do all the value mappings.
5544 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
5545 type_preserved flag.
5546 (ipa_set_jf_cst_copy): New function.
5547 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
5548 (ipa_set_jf_arith_pass_through): Likewise.
5549 (ipa_set_ancestor_jf): Likewise.
5550 (compute_complex_assign_jump_func): Set type_preserved instead of
5552 (ipa_compute_jump_functions_for_edge): Likewise.
5553 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
5554 (update_jump_functions_after_inlining): Update type_preserved.
5555 Explicitely create jump functions when combining one with pass_through.
5556 (ipa_write_jump_function): Stream the type_preserved flags.
5557 (ipa_read_jump_function): Likewise.
5559 2013-08-27 Jakub Jelinek <jakub@redhat.com>
5560 Aldy Hernandez <aldyh@redhat.com>
5562 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
5563 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
5564 * function.h (struct function): Add has_force_vect_loops and
5566 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
5567 * gimple.c (gimple_build_omp_critical): Add KIND argument and
5569 * gimple.def: Update CLAUSES comments.
5570 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
5571 (gimple_build_omp_for): Add argument to prototype.
5572 (gimple_omp_for_kind): New.
5573 (gimple_omp_for_set_kind): New.
5574 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
5575 GOVD_DATA_SHARE_CLASS.
5576 (enum omp_region_type): Add ORT_SIMD.
5577 (gimple_add_tmp_var): Handle ORT_SIMD.
5578 (gimplify_var_or_parm_decl): Same.
5579 (is_gimple_stmt): Same.
5580 (omp_firstprivatize_variable): Same.
5581 (omp_add_variable): Only use splay_tree_insert if lookup failed.
5582 (omp_notice_variable): Handle ORT_SIMD.
5583 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
5584 (omp_check_private): Handle ORT_SIMD.
5585 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
5587 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
5588 Handle OMP_CLAUSE_LASTPRIVATE.
5589 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
5591 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
5592 (gimplify_expr): Handle OMP_SIMD.
5593 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
5594 (expand_GOMP_SIMD_VF): New.
5595 (expand_GOMP_SIMD_LAST_LANE): New.
5596 * internal-fn.def (GOMP_SIMD_LANE): New.
5597 (GOMP_SIMD_VF): New.
5598 (GOMP_SIMD_LAST_LANE): New.
5599 * omp-low.c: Include target.h.
5600 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
5602 (check_omp_nesting_restrictions): Same.
5604 (lower_rec_simd_input_clauses): New.
5605 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
5607 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
5608 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
5609 (expand_omp_build_assign): New.
5610 (expand_omp_for_init_counts): New.
5611 (expand_omp_for_init_vars): New.
5612 (extract_omp_for_update_vars): New.
5613 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
5614 and rewrite accordingly.
5615 (expand_omp_simd): New.
5616 (expand_omp_for): Use expand_omp_simd.
5617 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
5618 (lower_omp_for): Do not lower the body.
5619 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
5621 * tree-flow.h (find_omp_clause): Remove prototype.
5622 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
5623 forcing vectorization of the loop, or if flag_tree_vectorize.
5624 (gate_tree_if_conversion): Similarly.
5625 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
5626 gimple_build_omp_for.
5627 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
5628 * tree-parloops (create_parallel_loop): Pass kind argument to
5629 gimple_build_omp_for.
5630 * tree-pretty-print.c (dump_omp_clause): Add cases for
5631 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
5632 OMP_CLAUSE__SIMDUID_.
5633 (dump_generic_node): Handle OMP_SIMD.
5634 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
5635 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
5636 unroll OMP_SIMD loops here.
5637 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
5638 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
5640 (vect_analyze_data_refs): Handle simd loops.
5641 * tree-vect-loop.c (vectorizable_live_operation): Handle
5643 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
5644 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
5645 (vectorizable_load): Same.
5646 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
5647 (struct simduid_to_vf): New.
5648 (simduid_to_vf::hash): New.
5649 (simduid_to-vf::equal): New.
5650 (struct simd_array_to_simduid): New.
5651 (simd_array_to_simduid::hash): New.
5652 (simd_array_to_simduid::equal): New.
5653 (adjust_simduid_builtins): New.
5654 (struct note_simd_array_uses_struct): New.
5655 (note_simd_array_uses_cb): New.
5656 (note_simd_array_uses): New.
5657 (vectorize_loops): Handle simd hints and adjust simd builtins
5659 * tree-vectorizer.h (struct _stmt_vec_info): Add
5660 simd_lane_access_p field.
5661 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
5662 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
5663 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
5664 (omp_clause_code_name): Same.
5665 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
5666 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
5667 * tree.def (OMP_SIMD): New entry.
5668 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
5669 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
5670 (OMP_CLAUSE_DECL): Adjust range for new clauses.
5671 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
5672 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
5673 (OMP_CLAUSE_LINEAR_STEP): New.
5674 (OMP_CLAUSE_SAFELEN_EXPR): New.
5675 (OMP_CLAUSE__SIMDUID__DECL): New.
5676 (find_omp_clause): New prototype.
5678 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
5680 * config/i386/driver-i386.c (host_detect_local_cpu): Update
5681 Haswell processor detection.
5683 2013-08-27 Christian Widmer <shadow@umbrox.de>
5686 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
5687 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
5688 AVX2 capable processors.
5690 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
5692 * config/aarch64/arm_neon.h: Replace all inline asm implementations
5693 of vget_low_* with implementations in terms of other intrinsics.
5695 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
5698 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
5699 values to -1, 0 and 1.
5701 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
5703 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
5704 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
5705 (<optab><mode>3_insn): Remove.
5706 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
5707 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
5708 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
5709 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
5710 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
5711 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
5712 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
5713 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
5714 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
5715 (ror<mode>3_insn): Likewise.
5716 * config/aarch64/predicates.md (aarch64_simd_register): New.
5718 2013-08-27 Richard Biener <rguenther@suse.de>
5720 PR tree-optimization/57521
5721 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
5722 one edge is non-critical.
5723 (find_phi_replacement_condition): Make sure to use a non-critical
5724 edge. Cleanup and remove old bug workarounds.
5725 (bb_postdominates_preds): Remove.
5726 (if_convertible_loop_p_1): Do not compute post-dominators.
5727 (combine_blocks): Do not free post-dominators.
5728 (main_tree_if_conversion): Likewise.
5729 (pass_data_if_conversion): Add TODO_verify_ssa.
5731 2013-08-27 DJ Delorie <dj@redhat.com>
5733 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
5735 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
5737 * function.c (assign_parm_find_data_types): Set passed_mode and
5738 nominal_mode to the TYPE_MODE of nominal_type for the built
5739 pointer type in case of the struct-pass-by-reference.
5741 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
5743 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
5744 (UINT16_TYPE): Change default to "unsigned int".
5746 * config/avr/avr.opt (mfract-convert-truncate): New option.
5747 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
5748 is set, round negative fractional integers according to n1169
5749 when converting to integer types.
5751 2013-08-26 Jan Hubicka <jh@suse.cz>
5753 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
5754 methods can not be called indirectly when their address is not taken.
5756 2013-08-26 Jan Hubicka <jh@suse.cz>
5758 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
5761 2013-08-26 Jan Hubicka <jh@suse.cz>
5763 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
5766 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
5768 * reload.c (find_valid_class): Allow classes that do not include
5769 FIRST_PSEUDO_REGISTER - 1.
5771 2013-08-26 Jan Hubicka <jh@suse.cz>
5773 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
5774 fix edge count/frequency when speculation failed; fix type check
5775 for the direct call.
5777 2013-08-26 Jan Hubicka <jh@suse.cz>
5779 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
5781 2013-08-26 Jan Hubicka <jh@suse.cz>
5783 * ipa-inline-transform.c (inline_transform): Be ready for basic block
5784 to be changed by edge redirection.
5786 2013-08-26 Jan Hubicka <jh@suse.cz>
5788 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
5789 formating; add sanity check.
5790 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
5791 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
5792 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
5793 (resolve_noninline_speculation): Update callee keys, too.
5795 2013-08-26 Jan Hubicka <jh@suse.cz>
5797 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
5798 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
5800 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
5802 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
5805 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
5807 * config/i386/i386.c (ix86_debug_options): Remove prototype.
5808 (x86_64_elf_select_section): Ditto.
5809 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
5811 (ix86_pass_by_reference): Ditto.
5812 (output_set_got): Ditto.
5813 (ix86_unary_operator_ok): Ditto.
5814 (ix86_expand_builtin): Ditto.
5816 2013-08-23 Jan Hubicka <jh@suse.cz>
5818 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
5820 2013-08-23 Jan Hubicka <jh@suse.cz>
5822 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
5823 (tree_decl_with_vis): Add FINAL field.
5825 2013-08-23 Jeff Law <law@redhat.com>
5827 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
5828 the debugging dump when the expression is fully redundant.
5830 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
5832 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
5833 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
5834 * pretty-print.c (pp_formatted_text_data): Likewise.
5835 (pp_write_text_to_stream): Likewise.
5836 (pp_write_text_as_dot_label_to_stream): Likewise.
5837 (pp_append_r): Likewise.
5838 (pp_format): Likewise.
5839 (pp_flush): Likewise.
5840 (pp_clear_output_area): Likewise.
5841 (pp_append_text): Likewise.
5842 (pp_formatted_text): Likewise.
5843 (pp_remaining_character_count_for_line): Likewise.
5844 (pp_newline): Likewise.
5845 (pp_character): Likewise.
5846 (output_buffer::~output_buffer): Define.
5847 (pretty_printer::~pretty_printer): Destruct output buffer.
5848 * pretty-print.h (output_buffer::~output_buffer): Declare.
5849 (pretty_printer::~pretty_printer): Declare virtual.
5851 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
5854 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
5855 HOST_WIDE_INT_M1U): New macros.
5856 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
5857 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
5858 unsigned -1 for lshift.
5859 * cse.c (cse_insn): Likewise.
5860 * double-int.c (rshift_double, lshift_double): Likewise.
5861 * builtins.c (fold_builtin_bitop): Likewise.
5862 * combine.c (force_to_mode): Likewise.
5863 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
5864 * simplify-rtx.c (simplify_const_unary_operation,
5865 simplify_const_binary_operation): Likewise.
5866 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
5867 check_va_list_escapes): Likewise.
5868 * rtlanal.c (nonzero_bits1): Likewise.
5869 * expmed.c (expand_smod_pow2): Likewise.
5870 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
5872 2013-08-23 Jan Hubicka <jh@suse.cz>
5874 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
5875 as having address taken.
5877 2013-08-23 Jan Hubicka <jh@suse.cz>
5879 * ipa-utils.h (method_class_type): Declare.
5880 * ipa-devirt.c (method_class_type): Export.
5882 * cgraphunit.c (analyze_functions): Do basic devirtualization;
5883 do not walk base classes of anonymous types.
5885 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
5887 PR rtl-optimization/58220
5889 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
5890 handle SEQUENCE insns properly.
5892 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
5894 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
5896 (pp_newline_and_indent): Likewise.
5897 (pp_separate_with): Likewise.
5898 * pretty-print.c (pp_newline_and_flush): Define.
5899 (pp_newline_and_indent): Likewise.
5900 (pp_separate_with): Likewise.
5902 2013-08-23 Jakub Jelinek <jakub@redhat.com>
5905 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
5906 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
5908 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
5910 * config/i386/predicates.md (ext_sse_reg_operand): New.
5911 * config/i386/i386.md (*movti_internal): Use
5912 predicate to determine if EVEX is needed.
5913 (*movsi_internal): Ditto.
5914 (*movdf_internal): Ditto.
5915 (*movsf_internal): Ditto.
5916 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
5918 2013-08-23 Jakub Jelinek <jakub@redhat.com>
5920 PR tree-optimization/58209
5921 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
5922 (find_tail_calls): Give up for pointer result types if m is non-NULL.
5923 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
5924 emit POINTER_PLUS_EXPR.
5925 (create_tailcall_accumulator): For pointer result type accumulate in
5928 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
5930 * configure.ac: Add backslashes missing from the last change.
5931 * configure: Regenerate.
5933 2013-08-22 Jan Hubicka <jh@susue.cz>
5935 * ipa.c (function_and_variable_visibility): First remember function
5936 was global and then make it local.
5938 2013-08-22 Julian Brown <julian@codesourcery.com>
5940 * configure.ac: Add aarch64 to list of arches which use "nop" in
5942 * configure: Regenerate.
5944 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5946 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
5947 gnu_libc_has_function.
5948 * config/s390/tpf.h: Likewise.
5950 2013-08-22 Jan Hubicka <jh@susue.cz>
5952 * timevar.c (validate_phases): Add cast.
5954 2013-08-22 Jan Hubicka <jh@susue.cz>
5956 * timevar.c (validate_phases): Use size_t for memory.
5957 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
5959 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
5961 * pretty-print.h (output_buffer::output_buffer): Declare.
5962 (pretty_printer::pretty_printer): Likewise.
5963 (pp_construct): Remove.
5964 * pretty-print.c (output_buffer::output_buffer): Define.
5965 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
5966 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
5967 (print_gimple_expr): Likewise.
5968 (print_gimple_seq): Likewise.
5969 (gimple_dump_bb): Likewise.
5970 * sched-vis.c (dump_value_slim): Likewise.
5971 (dump_insn_slim): Likewise.
5972 (dump_rtl_slim): Likewise.
5973 (str_pattern_slim): Likewise.
5974 * tree-mudflap.c (mf_varname_tree): Likewise.
5975 * graph.c (print_graph_cfg): Likewise.
5976 (start_graph_dump): Likewise.
5977 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
5979 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
5980 pretty printer initialization.
5981 * coretypes.h (diagnostic_context): Remove superflous type alias
5983 (pretty_printer): Likewise. Declare directly as a class.
5984 (pretty_print_info): Remove declaration as class.
5985 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
5986 and pp_clear_output_area.
5987 (asan_add_global): Likewise.
5989 2013-08-22 Jan Hubicka <jh@suse.cz>
5991 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
5992 * ipa-utils.h (update_type_inheritance_graph): Declare.
5993 (possible_polymorphic_call_target_p): Declare.
5994 (possible_polymorphic_call_target_p): New.
5995 * ipa-devirt.c: Update toplevel comments.
5996 (cached_polymorphic_call_targets): Move up.
5997 (odr_type_d): Move ID down.
5998 (polymorphic_type_binfo_p): Update comment.
5999 (odr_hasher::remove): Likewise;
6000 (get_odr_type): Set anonymous_namespace.
6001 (dump_odr_type): Dump it.
6002 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
6003 (maybe_record_node): Record node in cached_polymorphic_call_targets.
6004 (record_binfo): Add comment.
6005 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
6007 (devirt_node_removal_hook): Do not iCE when cache is freed.
6008 (possible_polymorphic_call_target_p): New predicate.
6009 (update_type_inheritance_graph): New function.
6011 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
6012 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6013 Sergey Lega <sergey.s.lega@intel.com>
6014 Anna Tikhonova <anna.tikhonova@intel.com>
6015 Ilya Tocar <ilya.tocar@intel.com>
6016 Andrey Turetskiy <andrey.turetskiy@intel.com>
6017 Ilya Verbin <ilya.verbin@intel.com>
6018 Kirill Yukhin <kirill.yukhin@intel.com>
6019 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6021 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
6022 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
6023 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
6024 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
6025 (OPTION_MASK_ISA_AVX2_UNSET): Update.
6026 (OPTION_MASK_ISA_AVX512F_UNSET): New.
6027 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
6028 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
6029 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
6030 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
6031 OPT_mavx512pf, OPT_mavx512er cases.
6032 * config/i386/constraints.md (v): New constraint.
6033 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
6034 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
6035 (bit_AVX512CD): New.
6036 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
6037 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
6038 * config/i386/i386-c.c (ix86_target_macros_internal):
6039 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
6041 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
6042 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
6043 * config/i386/i386.c (regclass_map, dbx_register_map)
6044 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
6045 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
6046 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
6048 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
6049 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
6050 -mavx512pf options. Fix formatting.
6051 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
6052 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
6053 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
6054 -mavx512cd, -mavx512pf options.
6055 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
6056 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
6057 (ix86_preferred_output_reload_class): Replace SSE_REGS with
6059 (ix86_hard_regno_mode_ok): Support 512-bit registers.
6060 (ix86_set_reg_reg_cost): Ditto.
6061 (x86_order_regs_for_local_alloc): Ditto.
6062 (MAX_VECT_LEN): Extend to 64-byte.
6063 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
6064 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
6065 (TARGET_AVX512ER, TARGET_AVX512CD): New.
6066 (BIGGEST_ALIGNMENT): Extend to 512-bits.
6067 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
6068 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
6069 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
6070 (SSE_REG_MODE_P): Support new modes.
6071 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
6072 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
6073 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
6074 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
6075 (REG_CLASS_CONTENTS): Add new registers.
6076 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
6077 (EXT_REX_SSE_REGNO_P): New.
6078 (HI_REGISTER_NAMES): Add new registers.
6079 * config/i386/i386.md: Define constants for new registers.
6080 (mode): Add new 512-bit modes.
6081 (prefix): Support evex prefix.
6082 (isa): Support avx512f, noavx512f, fma_avx512f.
6083 (ssemodesuffix): Add new 512-bit modes.
6085 (*movxi_internal_avx512f): Ditto.
6086 (*movdi_internal): Replace constraint "x" with the new constraint "v".
6088 (*movsi_internal): Likewise.
6089 (*movdf_internal): Likewise.
6090 (*movsf_internal): Likewise.
6091 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
6092 (<code><mode>3): Likewise.
6093 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
6095 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
6096 with the new constraint "v".
6097 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
6099 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
6100 with the new constraint "v".
6101 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
6102 (<sse2>_storedqu<avxsizesuffix>): Likewise.
6103 (*<plusminus_insn><mode>3): Likewise.
6104 (<sse>_vm<plusminus_insn><mode>3): Likewise.
6105 (*mul<mode>3): Likewise.
6106 (<sse>_vmmul<mode>3): Likewise.
6107 (<sse>_div<mode>3): Likewise.
6108 (<sse>_vmdiv<mode>3): Likewise.
6109 (<sse>_sqrt<mode>2): Likewise.
6110 (<sse>_vmsqrt<mode>2): Likewise.
6111 (*<code><mode>3_finite): Likewise.
6112 (*<code><mode>3) <smaxmin>: Likewise.
6113 (<sse>_vm<code><mode>3): Likewise.
6114 (*<code><mode>3) <any_logic>: Likewise.
6115 (*fma_fmadd_<mode>): Likewise.
6116 (*fma_fmsub_<mode>): Likewise.
6117 (*fma_fnmadd_<mode>): Likewise.
6118 (*fma_fnmsub_<mode>): Likewise.
6119 (*fma_fmaddsub_<mode>): Likewise.
6120 (*fma_fmsubadd_<mode>): Likewise.
6121 (*fmai_fmadd_<mode>): Likewise.
6122 (*fmai_fmsub_<mode>): Likewise.
6123 (*fmai_fnmadd_<mode>): Likewise.
6124 (*fmai_fnmsub_<mode>): Likewise.
6125 (sse_cvtsi2ss): Likewise.
6126 (sse_cvtsi2ssq): Likewise.
6127 (sse_cvtss2si): Likewise.
6128 (sse_cvtss2si_2): Likewise.
6129 (sse_cvtss2siq): Likewise.
6130 (sse_cvtss2siq_2): Likewise.
6131 (sse_cvttss2si): Likewise.
6132 (sse_cvtss2siq_2): Likewise.
6133 (float<sseintvecmodelower><mode>2): Likewise.
6134 (sse2_cvtsd2si_2): Likewise.
6135 (sse2_cvtsd2siq_2): Likewise.
6136 (*<plusminus_insn><mode>3): Likewise.
6137 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
6138 (*<sse4_1_avx2>_mul<mode>3): Likewise.
6139 (ashr<mode>3): Likewise.
6140 (<shift_insn><mode>3): Likewise.
6141 (avx2_<code><mode>3): Likewise.
6142 (*avx2_<code><mode>3): Likewise.
6143 (*andnot<mode>3): Likewise.
6144 (*<code><mode>3) <any_logic>: Likewise.
6145 (abs<mode>2): Likewise.
6146 (avx2_permvar<mode>): Likewise.
6147 (avx2_perm<mode>_1): Likewise.
6148 (*avx_vpermilp<mode>): Likewise.
6149 (avx_vpermilvar<mode>3): Likewise.
6150 (avx2_ashrv<mode>): Likewise.
6151 (avx2_<shift_insn>v<mode>): Likewise.
6152 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
6154 * doc/rtl.texi: Document XImode.
6156 2013-08-21 Jeff Law <law@redhat.com>
6158 * tree-flow.h (register_jump_thread): Pass vector of edges
6159 instead of each important edge.
6160 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
6161 thread path into a vector and pass that to register_jump_thread.
6162 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
6163 passed in edge vector to the current 3-edge form.
6166 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
6168 * dce.c (fini_dce): Call df_analyze again just in case
6169 delete_unmarked_insns removed anything.
6171 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
6173 * reload.h (struct reg_equivs): Rename to ..
6174 (struct reg_equivs_s): .. this.
6176 2013-08-20 Martin Liska <marxin.liska@gmail.com>
6178 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
6180 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6182 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
6184 2013-08-21 Jeff Law <law@redhat.com>
6186 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
6187 simplify assignments too. If the RHS collapses to a singleton
6188 range, then return the value for the range.
6190 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
6192 * config/i386/sse.md (V16): Rename to...
6194 (mov<mode>): Update iterator name.
6195 (*mov<mode>_internal): Ditto.
6196 (push<mode>1): Ditto.
6197 (movmisalign<mode>): Ditto.
6199 2013-08-20 Jan Hubicka <jh@suse.cz>
6202 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
6203 entry for direct edges.
6204 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
6206 2013-08-20 David Malcolm <dmalcolm@redhat.com>
6208 Revert my last two changes, r201865 and r201864:
6211 2013-08-20 David Malcolm <dmalcolm@redhat.com>
6213 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
6214 instances can own GC refs.
6216 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
6217 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
6218 (gcc::context::gt_pch_nx): Likewise.
6219 (gcc::context::gt_pch_nx): Likewise.
6220 * ggc.h (gt_ggc_mx <T>): New.
6221 (gt_pch_nx_with_op <T>): New.
6222 (gt_pch_nx <T>): New.
6223 * passes.c (opt_pass::gt_ggc_mx): New.
6224 (opt_pass::gt_pch_nx): New.
6225 (opt_pass::gt_pch_nx_with_op): New.
6226 (pass_manager::gt_ggc_mx): New.
6227 (pass_manager::gt_pch_nx): New.
6228 (pass_manager::gt_pch_nx_with_op): New.
6229 (pass_manager::operator new): Use
6230 ggc_internal_cleared_alloc_stat rather than xcalloc.
6231 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
6232 (pass_manager::gt_ggc_mx): New.
6233 (pass_manager::gt_pch_nx): New.
6234 (pass_manager::gt_pch_nx_with_op): New.
6235 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
6236 (opt_pass::operator new): New.
6237 (opt_pass::gt_ggc_mx): New.
6238 (opt_pass::gt_pch_nx): New.
6239 (opt_pass::gt_pch_nx_with_op): New.
6242 2013-08-20 David Malcolm <dmalcolm@redhat.com>
6244 * Makefile.in (GTFILES): Add context.h.
6245 * context.c (gcc::context::operator new): New.
6246 (gcc::context::gt_ggc_mx): New.
6247 (gcc::context::gt_pch_nx): New.
6248 (gcc::context::gt_pch_nx): New.
6249 * context.h (gcc::context): Add GTY((user)) marking.
6250 (gcc::context::operator new): New.
6251 (gcc::context::gt_ggc_mx): New.
6252 (gcc::context::gt_pch_nx): New.
6253 (gcc::context::gt_pch_nx): New.
6254 (g): Add GTY marking.
6255 (gt_ggc_mx (gcc::context *)): New.
6256 (gt_pch_nx (gcc::context *)): New.
6257 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
6258 void *cookie)): New.
6259 * gengtype.c (open_base_files) <ifiles>: Add context.h.
6261 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
6263 * dce.c (fini_dce): Call df_analyze again just in case
6264 delete_unmarked_insns removed anything.
6266 2013-08-20 Teresa Johnson <tejohnson@google.com>
6268 PR rtl-optimizations/57451
6269 * final.c (reemit_insn_block_notes): Prevent lexical blocks
6270 from crossing split section boundaries.
6272 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
6274 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
6275 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
6276 with MULTLIB_DEFAULTS.
6278 2013-08-20 Nick Clifton <nickc@redhat.com>
6280 * target.def (narrow_volatile_bitfield): Note that the default
6281 value is false, not !TARGET_STRICT_ALIGN.
6282 * doc/tm.texi: Regenerate.
6284 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
6286 Fix LIB_SPEC for systems without libpthread.
6288 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
6289 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
6291 * config/i386/linux-common.h: Likewise.
6292 * config/mips/linux-common.h: Likewise.
6294 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
6296 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
6299 2013-08-20 David Malcolm <dmalcolm@redhat.com>
6301 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
6302 instances can own GC refs.
6304 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
6305 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
6306 (gcc::context::gt_pch_nx): Likewise.
6307 (gcc::context::gt_pch_nx): Likewise.
6308 * ggc.h (gt_ggc_mx <T>): New.
6309 (gt_pch_nx_with_op <T>): New.
6310 (gt_pch_nx <T>): New.
6311 * passes.c (opt_pass::gt_ggc_mx): New.
6312 (opt_pass::gt_pch_nx): New.
6313 (opt_pass::gt_pch_nx_with_op): New.
6314 (pass_manager::gt_ggc_mx): New.
6315 (pass_manager::gt_pch_nx): New.
6316 (pass_manager::gt_pch_nx_with_op): New.
6317 (pass_manager::operator new): Use
6318 ggc_internal_cleared_alloc_stat rather than xcalloc.
6319 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
6320 (pass_manager::gt_ggc_mx): New.
6321 (pass_manager::gt_pch_nx): New.
6322 (pass_manager::gt_pch_nx_with_op): New.
6323 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
6324 (opt_pass::operator new): New.
6325 (opt_pass::gt_ggc_mx): New.
6326 (opt_pass::gt_pch_nx): New.
6327 (opt_pass::gt_pch_nx_with_op): New.
6329 2013-08-20 David Malcolm <dmalcolm@redhat.com>
6331 * Makefile.in (GTFILES): Add context.h.
6332 * context.c (gcc::context::operator new): New.
6333 (gcc::context::gt_ggc_mx): New.
6334 (gcc::context::gt_pch_nx): New.
6335 (gcc::context::gt_pch_nx): New.
6336 * context.h (gcc::context): Add GTY((user)) marking.
6337 (gcc::context::operator new): New.
6338 (gcc::context::gt_ggc_mx): New.
6339 (gcc::context::gt_pch_nx): New.
6340 (gcc::context::gt_pch_nx): New.
6341 (g): Add GTY marking.
6342 (gt_ggc_mx (gcc::context *)): New.
6343 (gt_pch_nx (gcc::context *)): New.
6344 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
6345 void *cookie)): New.
6346 * gengtype.c (open_base_files) <ifiles>: Add context.h.
6348 2013-08-20 Alan Modra <amodra@gmail.com>
6351 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
6352 (rs6000_emit_epilogue): Likewise.
6354 2013-08-19 Dehao Chen <dehao@google.com>
6356 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
6358 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
6359 Jakub Jelinek <jakub@redhat.com>
6361 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
6362 (BUILT_IN_FABSD64): Likewise.
6363 (BUILT_IN_FABSD128): Likewise.
6364 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
6365 (fold_builtin_1): Likewise.
6366 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
6367 destination and source operands.
6368 (*abstd2_fpr): Likewise.
6369 (*nabstd2_fpr): Likewise.
6371 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
6373 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
6376 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
6378 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
6381 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
6383 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
6384 * builtins.c (default_libc_has_function): New.
6385 (gnu_libc_has_function): Ditto.
6386 (no_c99_libc_has_function): Ditto.
6387 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
6388 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
6389 (fold_builtin_sincos): Likewise.
6390 (fold_builtin_cexp): Likewise.
6391 * builtins.def (DEF_C94_BUILTIN): Likewise.
6392 (DEF_C99_BUILTIN): Likewise.
6393 (DEF_C99_C90RES_BUILTIN): Likewise.
6394 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
6395 definitions to using this define.
6396 * config/darwin-protos.h (darwin_libc_has_function): New.
6397 * config/darwin.c (darwin_libc_has_function): Ditto.
6398 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
6399 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
6400 * config/darwin.h: Ditto.
6401 * config/elfos.h: Ditto.
6402 * config/freebsd.h: Ditto.
6403 * config/i386/cygming.h: Ditto.
6404 * config/i386/djgpp.h: Ditto.
6405 * config/i386/i386-interix.h: Ditto.
6406 * config/microblaze/microblaze.h: Ditto.
6407 * config/mmix/mmix.h: Ditto.
6408 * config/gnu-user.h: Ditto.
6409 * config/ia64/hpux.h: Ditto.
6410 * config/pa/pa-hpux.h: Ditto.
6411 * config/pdp11/pdp11.h: Ditto.
6412 * config/picochip/picochip.h: Ditto.
6413 * config/linux.h: Ditto.
6414 * config/netbsd.h: Ditto.
6415 * config/openbsd.h: Ditto.
6416 * config/rs6000/aix43.h: Ditto.
6417 * config/rs6000/aix51.h: Ditto.
6418 * config/rs6000/aix52.h: Ditto.
6419 * config/rs6000/aix53.h: Ditto.
6420 * config/rs6000/aix61.h: Ditto.
6421 * config/rs6000/darwin.h: Ditto.
6422 * config/rs6000/linux.h: Ditto.
6423 * config/rs6000/linux64.h: Ditto.
6424 * config/s390/tpf.h: Ditto.
6425 * config/sol2-10.h: Ditto.
6426 * config/sol2.h: Ditto.
6427 * config/vms/vms.h: Ditto.
6428 * config/vxworks.h: Ditto.
6429 * config/linux-android.c (linux_android_libc_has_function):
6430 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
6431 * config/linux-protos.h (linux_android_libc_has_function):
6433 * config/i386/i386.c (ix86_libc_has_function): New.
6434 * config/i386/i386-protos.h
6435 (ix86_libc_has_function): New declaration.
6436 * config/i386/i386.md
6437 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
6438 ("isinf<mode>2): Likewise.
6439 * convert.c (convert_to_integer): Using new target hook
6440 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
6441 TARGET_C99_FUNCTIONS.
6442 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
6443 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
6444 * coretypes.h (function_class): New enum for different
6445 classes of functions.
6446 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
6447 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
6448 (TARGET_HAS_SINCOS): Likewise.
6449 (TARGET_LIBC_HAS_FUNCTION): New.
6450 * doc/tm.texi: Regenerated.
6451 * targhooks.h (default_libc_has_function): New declaration.
6452 (no_c99_libc_has_function): Ditto.
6453 (gnu_libc_has_function): Ditto.
6454 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
6455 and TARGET_HAS_SINCOS.
6457 2013-08-18 Jan Hubicka <jh@suse.cz>
6459 * Makeifle-in (ipa-devirt.o): New.
6460 (GTFILES): Add ipa-utils.h and ipa-devirt.c
6461 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
6462 (analyze_functions): Look into possible targets of polymorphic call.
6463 * dumpfile.c (dump_files): Add type-inheritance dump.
6464 * dumpfile.h (TDI_inheritance): New.
6465 * ipa-devirt.c: New file.
6466 * ipa-utils.h (odr_type_d): Forward declare.
6467 (odr_type): New type.
6468 (build_type_inheritance_graph): Declare.
6469 (possible_polymorphic_call_targets): Declare and introduce inline
6470 variant when only edge is pased.
6471 (dump_possible_polymorphic_call_targets): Likewise.
6472 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
6473 * tree.c (type_in_anonymous_namespace_p): Break out from ...
6474 (types_same_for_odr): ... here.
6475 * tree.h (type_in_anonymous_namespace_p): Declare.
6477 2013-08-18 Jakub Jelinek <jakub@redhat.com>
6479 PR tree-optimization/58006
6480 * tree-parloops.c (take_address_of): Don't ICE if get_name
6482 (eliminate_local_variables_stmt): Remove clobber stmts.
6484 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
6486 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
6487 error message is issued for an alias to undefined symbol.
6489 2013-08-18 Jan Hubicka <jh@suse.cz>
6491 * cgraph.c (cgraph_create_indirect_edge): Discover
6492 polymorphic calls and record basic info into indirect_info.
6493 * gimple-fold.c (gimple_fold_call): When doing BINFO based
6494 devirtualization, ignore objc function calls.
6495 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
6496 call with no parm index info.
6497 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
6498 * tree.c (virtual_method_call_p): New function.
6499 * tree.h (virtual_method_call_p): Declare.
6501 2013-08-16 Jan Hubicka <jh@suse.cz>
6504 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
6506 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
6508 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
6511 2013-08-16 David Malcolm <dmalcolm@redhat.com>
6513 * gengtype.c (type_for_name): Add special-case support for
6514 locating types within the "gcc::" namespace.
6515 (open_base_files): Emit a "using namespace gcc" directive.
6517 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
6520 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
6521 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
6523 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
6524 array instead of each individual operand as a separate argument.
6525 (emit_fusion_gpr_load): Likewise.
6526 (expand_fusion_gpr_load): Add new function declaration.
6528 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
6529 signature to have the operands passed as an array, instead of as
6530 separate arguments. Allow ZERO_EXTEND to be in the memory
6531 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
6532 depend on the register live/dead flags when peepholes are run.
6533 (expand_fusion_gpr_load): New function to be called from the
6534 peephole2 pass, to change the register that addis sets to be the
6536 (emit_fusion_gpr_load): Change the calling signature to have the
6537 operands passed as an array, instead of as separate arguments.
6538 Allow ZERO_EXTEND to be in the memory address, and also
6539 SIGN_EXTEND if -mpower8-fusion-sign.
6541 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
6543 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
6544 adjust the register addis loads up in the peephole2 pass. Do not
6545 depend on the register live/dead state when the peephole pass is done.
6547 2013-08-16 David Malcolm <dmalcolm@redhat.com>
6549 * gengtype.c (create_user_defined_type): Ensure that the kind
6550 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
6551 declaration is seen before the GTY((user)) marking.
6553 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
6556 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
6558 2013-08-16 Jan Hubicka <jh@suse.cz>
6560 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
6561 arugment expected_type.
6562 (gimple_fold_call): Use it.
6563 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
6564 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
6565 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
6566 (try_make_edge_direct_virtual_call): Likewise.
6567 * tree.c (obj_type_ref_class): New.
6568 * tree.h (obj_type_ref_class): Use it.
6570 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
6572 * sched-vis.c (rtl_slim_pp_initialized): Remove.
6573 (rtl_slim_pp): Likewise.
6574 (init_rtl_slim_pretty_print): Likewise.
6575 (dump_value_slim): Don't call it. Use local pretty printer.
6576 (dump_insn_slim): Likewise.
6577 (dump_rtl_slim): Likewise.
6578 (str_pattern_slim): Likewise.
6579 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
6582 2013-08-16 Jakub Jelinek <jakub@redhat.com>
6584 PR tree-optimization/58164
6585 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
6586 walk gimple_goto_dest of GIMPLE_GOTO.
6588 PR tree-optimization/58165
6589 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
6590 bi_call must be the last stmt in a bb, don't split_block, instead
6591 use fallthru edge from it and give up if there is none.
6592 Release conds vector when returning early.
6594 2013-08-14 Xinliang David Li <davidxl@google.com>
6596 * config/i386/i386.c (ix86_option_override_internal):
6597 Remove unused variable and field.
6599 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6602 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
6603 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
6604 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
6605 and Linux, correct BLKmode alignment when 128-bit alignment is
6606 required and compatibility flag is not set.
6607 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
6608 for zero-size arguments when compatibility flag is not set.
6610 2013-08-14 Jakub Jelinek <jakub@redhat.com>
6612 PR tree-optimization/58145
6613 * tree-sra.c (build_ref_for_offset): If prev_base has
6614 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
6616 2013-08-14 Xinliang David Li <davidxl@google.com>
6618 * config/i386/i386.c (ix86_option_override_internal):
6619 Fix uninitialized variable error.
6621 2013-08-14 Xinliang David Li <davidxl@google.com>
6623 * config/i386/i386.opt: Define two new options.
6624 * config/i386/x86-tune.def: Add arch selector field in macros.
6625 * config/i386/i386.h: Adjust macro definition.
6626 * config/i386/i386.c (ix86_option_override_internal):
6628 (parse_mtune_ctrl_str): New function.
6629 (set_ix86_tune_features): New function.
6630 (ix86_function_specific_restore): Call the new helper function.
6632 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
6634 PR rtl-optimization/57662
6635 * sel-sched.c (code_motion_process_successors): When the current insn
6636 is removed after the recursive traversal, break from the loop.
6637 Add comments and debug printouts.
6639 2013-08-14 Jakub Jelinek <jakub@redhat.com>
6640 Alexandre Oliva <aoliva@redhat.com>
6643 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
6644 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
6645 there also UNSPEC_PLTOFF.
6647 2013-08-14 Marek Polacek <polacek@redhat.com>
6649 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
6650 NUM_CONDITIONS bit positions.
6652 2013-08-13 Cary Coutant <ccoutant@google.com>
6654 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
6655 (attr_checksum): Hash vector contents instead of pointer.
6656 (attr_checksum_ordered): Likewise.
6658 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
6660 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
6661 when Pmode != word_mode. Add length_address attribute.
6662 (sse3_monitor_<mode>): Merge from sse3_monitor and
6663 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
6664 Pmode != word_mode. Update insn length attribute.
6665 * config/i386/i386.c (ix86_option_override_internal): Update
6666 ix86_gen_monitor selection for merged sse3_monitor insn.
6668 2013-08-13 Julian Brown <julian@codesourcery.com>
6670 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
6671 perform invalid legitimization on greater-than-word-size modes for
6674 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
6676 * ira.c (setup_class_translate_array): Use aclass instead of cl
6677 for classes not fully covered by allocno classes.
6679 2013-08-13 Jakub Jelinek <jakub@redhat.com>
6681 PR tree-optimization/57661
6682 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
6683 * tree-inline.c (tree_function_versioning): Initialize it.
6684 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
6685 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
6686 that is not being copied.
6689 * asan.c (instrument_strlen_call): Fix typo in comment.
6690 Use char * type even for the lhs of POINTER_PLUS_EXPR.
6692 2013-08-13 Steve Ellcey <sellcey@mips.com>
6694 * config/mips/mips.md (prefetch): Use lw instead of ld on
6695 loongson in 32bit mode.
6697 2013-08-13 Nick Clifton <nickc@redhat.com>
6699 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
6701 2013-08-13 Jan Hubicka <jh@suse.cz>
6703 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
6704 introduced edge; fix typo in sanity check.
6705 (cgraph_resolve_speculation): Export; improve diagnostic.
6706 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
6707 speculation at type mismatch.
6708 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
6709 (cgraph_resolve_speculation): Declare.
6710 (symtab_can_be_discarded): New function.
6711 * value-prof.c (gimple_ic_transform): Remove actual transform code.
6712 * ipa-inline-transform.c (speculation_removed): New global var.
6713 (clone_inlined_nodes): See if speculation can be removed.
6714 (inline_call): If speculations was removed, we growths may not match.
6715 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
6716 (speculation_useful_p): New function.
6717 (resolve_noninline_speculation): New function.
6718 (inline_small_functions): Resolve useless speculations.
6719 * ipa-inline.h (speculation_useful_p): Declare
6720 * ipa.c (can_replace_by_local_alias): Simplify.
6721 (ipa_profile): Produce speculative calls in non-lto, too;
6722 add simple cost model; produce local aliases.
6724 2013-08-13 David Malcolm <dmalcolm@redhat.com>
6726 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
6729 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
6731 * config/i386/i386.c (ix86_function_versions): Use error + inform.
6733 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
6735 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
6736 iterator instead of X87MODEF.
6738 2013-08-12 Perez Read <netfirewall@gmail.com>
6741 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
6742 operand 0 for intel asm alternative.
6743 (*movabs<mode>_2): Ditto for operand 1.
6745 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
6747 * config/aarch64/arm_none.h
6748 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
6750 2013-08-12 Nick Clifton <nickc@redhat.com>
6752 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
6754 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
6756 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
6757 expand for QI/HImode operand to produce more effictive code for
6758 unsigned char(short) --> float(double) conversion.
6760 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
6762 * doc/invoke.texi: Mention that -ftls-model does not force the final
6765 2013-08-12 Marek Polacek <polacek@redhat.com>
6766 Marc Glisse <marc.glisse@inria.fr>
6768 PR tree-optimization/57980
6769 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
6770 when creating -1 constant.
6772 2013-08-10 Jan Hubicka <jh@suse.cz>
6774 Workaround binutils PR14342.
6775 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
6776 (gimple_init_edge_profiler): Likewise.
6777 (gimple_gen_ic_func_profiler): Likewise.
6779 2013-08-09 Jan Hubicka <jh@suse.cz>
6781 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
6783 2013-08-09 Xinliang David Li <davidxl@google.com>
6785 * config/i386/stringop.def: New file.
6786 * config/i386/stringop.opt: New file.
6787 * config/i386/i386-opts.h: Include stringopt.def.
6788 * config/i386/i386.opt: Include stringopt.opt.
6789 * config/i386/i386.c (ix86_option_override_internal):
6790 Override default size based stringop inline strategies with options.
6791 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
6794 2013-08-09 Jan Hubicka <jh@suse.cz>
6796 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
6798 2013-08-09 Jan Hubicka <jh@suse.cz>
6800 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
6802 (dump_cgraph_node): Dump profile-id.
6803 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
6804 and common_target_probability.
6805 * lto-cgraph.c (lto_output_edge): Stream common targets.
6806 (lto_output_node): Stream profile ids.
6807 (input_node): Stream profile ids.
6808 (input_edge): Stream common targets.
6809 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
6810 * ipa.c: Include value-prof.h
6811 (ipa_profile_generate_summary): Turn indirect call statement histograms
6812 into common targets.
6813 (ipa_profile): Turn common targets into speculative edges.
6815 2013-08-09 Jan Hubicka <jh@suse.cz>
6817 * cgraph.h (cgraph_node): Add profile_id.
6818 * value-prof.c (cgraph_node_map): Turn into pointer_map.
6819 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
6820 (del_node_map): Update.
6821 (find_func_by_funcdef_no): Replace by ...
6822 (find_func_by_profile_id): ... this one.
6823 (gimple_ic_transform): Do not remove useful histograms when
6824 speculation is not done; dump info when indirect call removal
6826 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
6827 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
6828 (__gcov_indirect_call_profiler_v2): .. this one.
6829 * profile.h (init_node_map): Update.
6830 * coverage.c (coverage_compute_profile_id): New function.
6831 * coverage.h (coverage_compute_profile_id): Declare.
6832 * tree-profile.c (init_ic_make_global_vars): Make
6833 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
6834 (gimple_init_edge_profiler): Update prototype of
6835 __gcov_indirect_call_profiler.
6836 (gimple_gen_ic_func_profiler): Simplify.
6837 (tree_profiling): Use init_node_map
6839 2013-08-09 Jan Hubicka <jh@suse.cz>
6841 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
6842 non-speculative refs.
6843 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
6844 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
6845 (cgraph_set_call_stmt): Likewise.
6846 (cgraph_create_edge_1): Fix release checking compilatoin;
6848 (cgraph_free_edge): Free indirect info.
6849 (cgraph_turn_edge_to_speculative): New function.
6850 (cgraph_speculative_call_info): New function.
6851 (cgraph_make_edge_direct): Return direct edge; handle speculation.
6852 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
6853 (dump_cgraph_node): Dump speculation.
6854 (verify_edge_count_and_frequency): Accept speculative edges.
6855 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
6856 (verify_cgraph_node): Handle speculation.
6857 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
6858 (cgraph_set_call_stmt): Update prototype.
6859 (cgraph_make_edge_direct): Update prototype.
6860 (cgraph_speculative_call_info): Declare.
6861 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
6862 to change; update call of ipa_find_references.
6863 * ipa-ref.c (ipa_record_reference): Fix return value; clear
6864 lto_stmt_uid and speculative flags.
6865 (ipa_dump_references): Dump speculation.
6866 (ipa_clone_references): Clone speculative flag.
6867 (ipa_clone_referring): Likewise.
6868 (ipa_clone_ref): New function.
6869 (ipa_find_reference): Look into lto_stmt_uids
6870 (ipa_clear_stmts_in_references): Do not clear speculative calls.
6871 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
6872 (ipa_find_reference): Update declaration.
6873 (ipa_clone_ref): Declare.
6874 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
6875 stream speculative flag.
6876 (lto_output_ref): Stream statements uids and speculation.
6877 (input_ref): Likewise.
6878 (input_edge): Stream speuclation.
6879 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
6880 (cgraph_set_call_stmt_including_clones): Handle speculation.
6881 * ipa-inline.c (heap_edge_removal_hook): New function.
6882 (inline_small_functions): Register it.
6883 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
6884 also initialize refs.
6885 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
6887 (try_make_edge_direct_simple_call): Likewise.
6888 (try_make_edge_direct_simple_call): Likewise.
6889 (update_indirect_edges_after_inlining): Likewise.
6890 (remove_described_reference): Look proper lto_stmt_uid.
6891 (propagate_controlled_uses): Likewise.
6892 (propagate_controlled_uses): Liekwise.
6893 * tree-inline.c (copy_bb): Copy speculative edges.
6894 (redirect_all_calls): New function.
6895 (copy_cfg_body): Do redirection after loop info is updated.
6896 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
6898 2013-08-09 Jan Hubicka <jh@suse.cz>
6900 * lto-streamer-out.c (output_function): Renumber PHIs.
6901 * lto-streamer-in.c (input_function): Likewise.
6903 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
6905 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
6906 (get_lane_unsigned): Likewise.
6907 (dup_lane_scalar): Likewise.
6908 (get_lane): enable for VALL.
6909 * config/aarch64/aarch64-simd.md
6910 (aarch64_dup_lane_scalar<mode>): Remove.
6911 (aarch64_get_lane_signed<mode>): Likewise.
6912 (aarch64_get_lane_unsigned<mode>): Likewise.
6913 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
6914 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
6915 (aarch64_get_lane<mode>): Enable for all vector modes.
6916 (aarch64_get_lanedi): Remove misleading constraints.
6917 * config/aarch64/arm_neon.h
6918 (__aarch64_vget_lane_any): Define.
6919 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
6920 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
6921 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
6922 * config/aarch64/iterators.md (VDQQH): New.
6926 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
6928 * configure.ac: Add GAS check for LEON instructions on SPARC.
6929 * configure: Regenerate.
6930 * config.in: Likewise.
6931 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
6933 * config/sparc/sparc.opt (LEON, LEON3): New masks.
6934 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
6936 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
6937 (AS_LEON_FLAG): New macro.
6938 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
6939 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
6940 Deal with LEON and LEON3 for the memory model.
6941 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
6942 (atomic_compare_and_swap<mode>_1): Likewise.
6943 (*atomic_compare_and_swap<mode>_1): Likewise.
6945 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6947 * config/arm/neon.md (vcond): Fix floating-point vector
6948 comparisons against 0.
6950 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
6952 * lra-constraints.c (emit_spill_move): Remove assert.
6953 (process_alt_operands): Add more debugging
6954 output. Increase reject for spilling into memory. Decrease
6955 reject for reloading scratch.
6956 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
6958 2013-08-08 Steve Ellcey <sellcey@mips.com>
6960 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
6961 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
6962 micromips incompatible. Add nan2008.
6963 (MULTILIB_DIRNAMES): Add nan2008.
6964 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
6965 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
6966 and micromips incompatible. Add nan2008.
6967 (MULTILIB_DIRNAMES): Add nan2008.
6968 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
6970 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
6972 PR rtl-optimization/58079
6973 * combine.c (combine_simplify_rtx): Avoid using SUBST if
6974 simplify_comparison has widened a comparison with an integer.
6976 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6978 * config/arm/neon.md (movmisalign<mode>): Disable when we
6979 don't allow unaligned accesses.
6980 (*movmisalign<mode>_neon_store): Likewise.
6981 (*movmisalign<mode>_neon_load): Likewise.
6982 (*movmisalign<mode>_neon_store): Likewise.
6983 (*movmisalign<mode>_neon_load): Likewise.
6985 2013-08-08 Jan Hubicka <jh@suse.cz>
6987 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
6988 (make_pass_rebuild_cgraph_edges): Also clear references.
6989 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
6990 * ipa-inline-transform.c (inline_transform): Remove all references
6992 * cgraphunit.c (expand_function): Remove all references after
6994 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
6995 (ipa_find_reference): Rewrite to iterator.
6996 (remove_stmt_references): Likewise.
6997 (ipa_clear_stmts_in_references): New function.
6998 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
6999 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
7001 * ipa-split.c (split_function): Remove references in split function.
7003 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
7006 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
7007 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
7009 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
7012 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
7013 suggested mode for the assignment isn't compatible with the
7016 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
7019 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
7021 2013-08-07 Xinliang David Li <davidxl@google.com>
7023 * config/i386/i386.opt: New option -mtune-ctrl=.
7024 * config/i386/x86-tune.def: New file.
7025 * config/i386/i386.h: include x86-tune.def.
7026 * config/i386/i386.c (ix86_option_override_internal):
7027 Parsing -mtune-ctrl= option and set tune features.
7029 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
7032 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
7033 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
7036 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
7038 * rtl.h (update_alignments): Declare.
7039 * final.c (grow_label_align): New function extracted from...
7040 (shorten_branches): ...here. Call it.
7041 (update_alignments): New function.
7042 * reorg.c (sibling_labels): New variable.
7043 (get_label_before): Add SIBLING parameter. If it is non-zero, push
7044 the new label along with it onto the sibling_labels vector.
7045 (fill_simple_delay_slots): Adjust call to get_label_before.
7046 (fill_slots_from_thread): Likewise.
7047 (relax_delay_slots): Likewise.
7048 (make_return_insns): Likewise.
7049 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
7051 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
7053 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
7054 document its semantics.
7055 (diagnostic_report_diagnostic): Adjust accordingly.
7057 2013-08-07 David Malcolm <dmalcolm@redhat.com>
7059 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
7060 (sparc_option_override): ...and port to new C++ pass API.
7061 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
7063 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
7065 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
7067 2013-08-06 Caroline Tice <cmtice@google.com>
7069 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
7070 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
7071 * tree-pass.h: Add pass_vtable_verify.
7072 * varasm.c (assemble_variable): Add code to properly set the comdat
7073 section and name for the .vtable_map_vars section.
7074 (assemble_vtyv_preinit_initializer): New function.
7075 (default_sectin_type_flags): Make sure .vtable_map_vars section has
7077 * output.h: Add function decl for assemble_vtv_preinit_initializer.
7078 * vtable-verify.c: New file.
7079 * vtable-verify.h: New file.
7080 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
7081 initialiation levels.
7082 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
7083 * passes.def: Insert pass_vtable_verify.
7084 * aclocal.m4: Reorder includes.
7085 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
7086 -fvtv-counts options.
7087 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
7088 as appropriate, if -fvtable-verify=... is used.
7089 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
7090 -fvtable-verify=... is used.
7091 * Makefile.in (OBJS): Add vtable-verify.o to list.
7092 (vtable-verify.o): Add new build rule.
7093 (GTFILES): Add vtable-verify.c to list.
7094 * common.opt (fvtable-verify=): New flag.
7095 (vtv_priority): Values for fvtable-verify= flag.
7096 (fvtv-counts): New flag.
7097 (fvtv-debug): New flag.
7098 * tree.h (save_vtable_map_decl): New extern function decl.
7100 2013-08-07 David Malcolm <dmalcolm@redhat.com>
7102 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
7103 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
7104 (pass_data_rl78_devirt): ...new pass_data instance and...
7105 (make_pass_rl78_devirt): ...new function.
7106 (rl78_asm_file_start): Port pass registration to new C++ API.
7108 2013-08-07 David Malcolm <dmalcolm@redhat.com>
7110 * coretypes.h (rtl_opt_pass): Add.
7111 (gcc::context): Add.
7112 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
7113 (epiphany_init): Port to new C++ pass API.
7114 (epiphany_optimize_mode_switching): Likewise.
7115 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
7116 (pass_manager::get_pass_mode_switching): New.
7117 (pass_manager::get_pass_peephole2): New.
7118 * mode-switching.c (pass_mode_switching): Add clone method.
7119 * recog.c (pass_peephole2): Add clone method.
7120 (pass_split_all_insns): Add clone method.
7122 2013-08-06 David Malcolm <dmalcolm@redhat.com>
7124 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
7125 (mips_option_override): ...here, porting to new C++ API for passes.
7127 2013-08-06 Jan Hubicka <jh@suse.cz>
7129 * cgraph.c (cgraph_get_body): New function based on lto.c
7131 * cgraph.h (cgraph_get_body): Declare.
7132 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
7134 * cgraphunit.c (expand_function): Get body prior expanding.
7135 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
7136 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
7138 * passes.c (do_per_function_toporder): Get body.
7139 * tree-inline.c (expand_call_inline): Get body prior inlining it.
7140 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
7142 2013-08-06 Martin Jambor <mjambor@suse.cz>
7145 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
7146 is not re-finalized. Rename second parameter to no_collect.
7148 2013-08-06 Martin Jambor <mjambor@suse.cz>
7151 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
7152 MEM_REF has proper alignment information.
7154 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
7157 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
7159 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
7160 rtx (*) (rtx, ...) with insn_gen_fn.
7161 * genoutput.c (output_insn_data): Cast gen_? function pointers to
7162 insn_gen_fn::stored_funcptr. Add initializer braces.
7164 2013-08-05 David Malcolm <dmalcolm@redhat.com>
7166 Rewrite how instances of passes are cloned to remove assumptions
7167 about their sizes (thus allowing pass subclasses to have
7168 additional data fields, albeit non-GC-managed ones at this point).
7170 * passes.c (make_pass_instance): Now that passes have clone
7171 methods, rewrite this function to eliminate XNEW and memcpy
7172 calls that used hardcoded sizes. Since this function no longer
7173 creates pass instances, rename it to...
7174 (add_pass_instance): ...this. Document the old way that passes were
7175 numbered and flagged, and rework this function to continue using it.
7176 (next_pass_1): Add an initial_pass argument for use by
7178 (position_pass): When adding multiple instances of a pass, use
7179 the pass's clone method, rather than relying on the XNEW/memcpy
7180 within the former make_pass_instance (now add_pass_instance).
7181 (pass_manager::pass_manager): When invoking next_pass_1, also supply
7182 the initial instance of the current pass within the pass manager.
7184 2013-08-05 David Malcolm <dmalcolm@redhat.com>
7186 This is the automated part of the conversion of passes from C
7187 structs to C++ classes.
7189 Patch autogenerated by refactor_passes.py from
7190 https://github.com/davidmalcolm/gcc-refactoring-scripts
7191 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
7193 * asan.c (pass_asan): Convert from a global struct to a subclass of
7194 gimple_opt_pass along with...
7195 (pass_data_asan): ...new pass_data instance and...
7196 (make_pass_asan): ...new function.
7197 (pass_asan_O0): Convert from a global struct to a subclass of
7198 gimple_opt_pass along with...
7199 (pass_data_asan_O0): ...new pass_data instance and...
7200 (make_pass_asan_O0): ...new function.
7201 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
7202 subclass of rtl_opt_pass along with...
7203 (pass_data_inc_dec): ...new pass_data instance and...
7204 (make_pass_inc_dec): ...new function.
7205 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
7206 a subclass of rtl_opt_pass along with...
7207 (pass_data_reorder_blocks): ...new pass_data instance and...
7208 (make_pass_reorder_blocks): ...new function.
7209 (pass_duplicate_computed_gotos): Convert from a global struct to a
7210 subclass of rtl_opt_pass along with...
7211 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
7212 (make_pass_duplicate_computed_gotos): ...new function.
7213 (pass_partition_blocks): Convert from a global struct to a subclass of
7214 rtl_opt_pass along with...
7215 (pass_data_partition_blocks): ...new pass_data instance and...
7216 (make_pass_partition_blocks): ...new function.
7217 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
7218 struct to a subclass of rtl_opt_pass along with...
7219 (pass_data_branch_target_load_optimize1): ...new pass_data instance
7221 (make_pass_branch_target_load_optimize1): ...new function.
7222 (pass_branch_target_load_optimize2): Convert from a global struct to a
7223 subclass of rtl_opt_pass along with...
7224 (pass_data_branch_target_load_optimize2): ...new pass_data instance
7226 (make_pass_branch_target_load_optimize2): ...new function.
7227 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
7228 of rtl_opt_pass along with...
7229 (pass_data_jump): ...new pass_data instance and...
7230 (make_pass_jump): ...new function.
7231 (pass_jump2): Convert from a global struct to a subclass of
7232 rtl_opt_pass along with...
7233 (pass_data_jump2): ...new pass_data instance and...
7234 (make_pass_jump2): ...new function.
7235 * cfgexpand.c (pass_expand): Convert from a global struct to a
7236 subclass of rtl_opt_pass along with...
7237 (pass_data_expand): ...new pass_data instance and...
7238 (make_pass_expand): ...new function.
7239 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
7240 of rtl_opt_pass along with...
7241 (pass_data_free_cfg): ...new pass_data instance and...
7242 (make_pass_free_cfg): ...new function.
7243 (pass_into_cfg_layout_mode): Convert from a global struct to a
7244 subclass of rtl_opt_pass along with...
7245 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
7246 (make_pass_into_cfg_layout_mode): ...new function.
7247 (pass_outof_cfg_layout_mode): Convert from a global struct to a
7248 subclass of rtl_opt_pass along with...
7249 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
7250 (make_pass_outof_cfg_layout_mode): ...new function.
7251 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
7252 struct to a subclass of gimple_opt_pass along with...
7253 (pass_data_build_cgraph_edges): ...new pass_data instance and...
7254 (make_pass_build_cgraph_edges): ...new function.
7255 (pass_rebuild_cgraph_edges): Convert from a global struct to a
7256 subclass of gimple_opt_pass along with...
7257 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
7258 (make_pass_rebuild_cgraph_edges): ...new function.
7259 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
7260 subclass of gimple_opt_pass along with...
7261 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
7263 (make_pass_remove_cgraph_callee_edges): ...new function.
7264 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
7265 struct to a subclass of rtl_opt_pass along with...
7266 (pass_data_stack_adjustments): ...new pass_data instance and...
7267 (make_pass_stack_adjustments): ...new function.
7268 * combine.c (pass_combine): Convert from a global struct to a subclass
7269 of rtl_opt_pass along with...
7270 (pass_data_combine): ...new pass_data instance and...
7271 (make_pass_combine): ...new function.
7272 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
7273 global struct to a subclass of rtl_opt_pass along with...
7274 (pass_data_compare_elim_after_reload): ...new pass_data instance
7276 (make_pass_compare_elim_after_reload): ...new function.
7277 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
7278 of rtl_opt_pass along with...
7279 (pass_data_rtl_cprop): ...new pass_data instance and...
7280 (make_pass_rtl_cprop): ...new function.
7281 * cse.c (pass_cse): Convert from a global struct to a subclass of
7282 rtl_opt_pass along with...
7283 (pass_data_cse): ...new pass_data instance and...
7284 (make_pass_cse): ...new function.
7285 (pass_cse2): Convert from a global struct to a subclass of
7286 rtl_opt_pass along with...
7287 (pass_data_cse2): ...new pass_data instance and...
7288 (make_pass_cse2): ...new function.
7289 (pass_cse_after_global_opts): Convert from a global struct to a
7290 subclass of rtl_opt_pass along with...
7291 (pass_data_cse_after_global_opts): ...new pass_data instance and...
7292 (make_pass_cse_after_global_opts): ...new function.
7293 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
7294 of rtl_opt_pass along with...
7295 (pass_data_ud_rtl_dce): ...new pass_data instance and...
7296 (make_pass_ud_rtl_dce): ...new function.
7297 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
7298 rtl_opt_pass along with...
7299 (pass_data_fast_rtl_dce): ...new pass_data instance and...
7300 (make_pass_fast_rtl_dce): ...new function.
7301 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
7302 a subclass of rtl_opt_pass along with...
7303 (pass_data_df_initialize_opt): ...new pass_data instance and...
7304 (make_pass_df_initialize_opt): ...new function.
7305 (pass_df_initialize_no_opt): Convert from a global struct to a
7306 subclass of rtl_opt_pass along with...
7307 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
7308 (make_pass_df_initialize_no_opt): ...new function.
7309 (pass_df_finish): Convert from a global struct to a subclass of
7310 rtl_opt_pass along with...
7311 (pass_data_df_finish): ...new pass_data instance and...
7312 (make_pass_df_finish): ...new function.
7313 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
7314 rtl_opt_pass along with...
7315 (pass_data_rtl_dse1): ...new pass_data instance and...
7316 (make_pass_rtl_dse1): ...new function.
7317 (pass_rtl_dse2): Convert from a global struct to a subclass of
7318 rtl_opt_pass along with...
7319 (pass_data_rtl_dse2): ...new pass_data instance and...
7320 (make_pass_rtl_dse2): ...new function.
7321 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
7322 subclass of rtl_opt_pass along with...
7323 (pass_data_dwarf2_frame): ...new pass_data instance and...
7324 (make_pass_dwarf2_frame): ...new function.
7325 * except.c (pass_set_nothrow_function_flags): Convert from a global
7326 struct to a subclass of rtl_opt_pass along with...
7327 (pass_data_set_nothrow_function_flags): ...new pass_data instance
7329 (make_pass_set_nothrow_function_flags): ...new function.
7330 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
7331 subclass of rtl_opt_pass along with...
7332 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
7334 (make_pass_convert_to_eh_region_ranges): ...new function.
7335 * final.c (pass_compute_alignments): Convert from a global struct to a
7336 subclass of rtl_opt_pass along with...
7337 (pass_data_compute_alignments): ...new pass_data instance and...
7338 (make_pass_compute_alignments): ...new function.
7339 (pass_final): Convert from a global struct to a subclass of
7340 rtl_opt_pass along with...
7341 (pass_data_final): ...new pass_data instance and...
7342 (make_pass_final): ...new function.
7343 (pass_shorten_branches): Convert from a global struct to a subclass of
7344 rtl_opt_pass along with...
7345 (pass_data_shorten_branches): ...new pass_data instance and...
7346 (make_pass_shorten_branches): ...new function.
7347 (pass_clean_state): Convert from a global struct to a subclass of
7348 rtl_opt_pass along with...
7349 (pass_data_clean_state): ...new pass_data instance and...
7350 (make_pass_clean_state): ...new function.
7351 * function.c (pass_instantiate_virtual_regs): Convert from a global
7352 struct to a subclass of rtl_opt_pass along with...
7353 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
7354 (make_pass_instantiate_virtual_regs): ...new function.
7355 (pass_leaf_regs): Convert from a global struct to a subclass of
7356 rtl_opt_pass along with...
7357 (pass_data_leaf_regs): ...new pass_data instance and...
7358 (make_pass_leaf_regs): ...new function.
7359 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
7360 subclass of rtl_opt_pass along with...
7361 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
7363 (make_pass_thread_prologue_and_epilogue): ...new function.
7364 (pass_match_asm_constraints): Convert from a global struct to a
7365 subclass of rtl_opt_pass along with...
7366 (pass_data_match_asm_constraints): ...new pass_data instance and...
7367 (make_pass_match_asm_constraints): ...new function.
7368 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
7369 subclass of rtl_opt_pass along with...
7370 (pass_data_rtl_fwprop): ...new pass_data instance and...
7371 (make_pass_rtl_fwprop): ...new function.
7372 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
7373 rtl_opt_pass along with...
7374 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
7375 (make_pass_rtl_fwprop_addr): ...new function.
7376 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
7377 rtl_opt_pass along with...
7378 (pass_data_rtl_pre): ...new pass_data instance and...
7379 (make_pass_rtl_pre): ...new function.
7380 (pass_rtl_hoist): Convert from a global struct to a subclass of
7381 rtl_opt_pass along with...
7382 (pass_data_rtl_hoist): ...new pass_data instance and...
7383 (make_pass_rtl_hoist): ...new function.
7384 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
7385 subclass of gimple_opt_pass along with...
7386 (pass_data_lower_cf): ...new pass_data instance and...
7387 (make_pass_lower_cf): ...new function.
7388 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
7389 from a global struct to a subclass of gimple_opt_pass along with...
7390 (pass_data_strength_reduction): ...new pass_data instance and...
7391 (make_pass_strength_reduction): ...new function.
7392 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
7393 of rtl_opt_pass along with...
7394 (pass_data_rtl_ifcvt): ...new pass_data instance and...
7395 (make_pass_rtl_ifcvt): ...new function.
7396 (pass_if_after_combine): Convert from a global struct to a subclass of
7397 rtl_opt_pass along with...
7398 (pass_data_if_after_combine): ...new pass_data instance and...
7399 (make_pass_if_after_combine): ...new function.
7400 (pass_if_after_reload): Convert from a global struct to a subclass of
7401 rtl_opt_pass along with...
7402 (pass_data_if_after_reload): ...new pass_data instance and...
7403 (make_pass_if_after_reload): ...new function.
7404 * init-regs.c (pass_initialize_regs): Convert from a global struct to
7405 a subclass of rtl_opt_pass along with...
7406 (pass_data_initialize_regs): ...new pass_data instance and...
7407 (make_pass_initialize_regs): ...new function.
7408 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
7409 of ipa_opt_pass_d along with...
7410 (pass_data_ipa_cp): ...new pass_data instance and...
7411 (make_pass_ipa_cp): ...new function.
7412 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
7413 global struct to a subclass of gimple_opt_pass along with...
7414 (pass_data_inline_parameters): ...new pass_data instance and...
7415 (make_pass_inline_parameters): ...new function.
7416 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
7417 subclass of gimple_opt_pass along with...
7418 (pass_data_early_inline): ...new pass_data instance and...
7419 (make_pass_early_inline): ...new function.
7420 (pass_ipa_inline): Convert from a global struct to a subclass of
7421 ipa_opt_pass_d along with...
7422 (pass_data_ipa_inline): ...new pass_data instance and...
7423 (make_pass_ipa_inline): ...new function.
7424 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
7425 struct to a subclass of gimple_opt_pass along with...
7426 (pass_data_local_pure_const): ...new pass_data instance and...
7427 (make_pass_local_pure_const): ...new function.
7428 (pass_ipa_pure_const): Convert from a global struct to a subclass of
7429 ipa_opt_pass_d along with...
7430 (pass_data_ipa_pure_const): ...new pass_data instance and...
7431 (make_pass_ipa_pure_const): ...new function.
7432 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
7433 to a subclass of ipa_opt_pass_d along with...
7434 (pass_data_ipa_reference): ...new pass_data instance and...
7435 (make_pass_ipa_reference): ...new function.
7436 * ipa-split.c (pass_split_functions): Convert from a global struct to
7437 a subclass of gimple_opt_pass along with...
7438 (pass_data_split_functions): ...new pass_data instance and...
7439 (make_pass_split_functions): ...new function.
7440 (pass_feedback_split_functions): Convert from a global struct to a
7441 subclass of gimple_opt_pass along with...
7442 (pass_data_feedback_split_functions): ...new pass_data instance and...
7443 (make_pass_feedback_split_functions): ...new function.
7444 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
7445 global struct to a subclass of simple_ipa_opt_pass along with...
7446 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
7448 (make_pass_ipa_function_and_variable_visibility): ...new function.
7449 (pass_ipa_free_inline_summary): Convert from a global struct to a
7450 subclass of simple_ipa_opt_pass along with...
7451 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
7452 (make_pass_ipa_free_inline_summary): ...new function.
7453 (pass_ipa_whole_program_visibility): Convert from a global struct to a
7454 subclass of ipa_opt_pass_d along with...
7455 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
7457 (make_pass_ipa_whole_program_visibility): ...new function.
7458 (pass_ipa_profile): Convert from a global struct to a subclass of
7459 ipa_opt_pass_d along with...
7460 (pass_data_ipa_profile): ...new pass_data instance and...
7461 (make_pass_ipa_profile): ...new function.
7462 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
7463 ipa_opt_pass_d along with...
7464 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
7465 (make_pass_ipa_cdtor_merge): ...new function.
7466 * ira.c (pass_ira): Convert from a global struct to a subclass of
7467 rtl_opt_pass along with...
7468 (pass_data_ira): ...new pass_data instance and...
7469 (make_pass_ira): ...new function.
7470 (pass_reload): Convert from a global struct to a subclass of
7471 rtl_opt_pass along with...
7472 (pass_data_reload): ...new pass_data instance and...
7473 (make_pass_reload): ...new function.
7474 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
7475 subclass of rtl_opt_pass along with...
7476 (pass_data_cleanup_barriers): ...new pass_data instance and...
7477 (make_pass_cleanup_barriers): ...new function.
7478 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
7479 of rtl_opt_pass along with...
7480 (pass_data_loop2): ...new pass_data instance and...
7481 (make_pass_loop2): ...new function.
7482 (pass_rtl_loop_init): Convert from a global struct to a subclass of
7483 rtl_opt_pass along with...
7484 (pass_data_rtl_loop_init): ...new pass_data instance and...
7485 (make_pass_rtl_loop_init): ...new function.
7486 (pass_rtl_loop_done): Convert from a global struct to a subclass of
7487 rtl_opt_pass along with...
7488 (pass_data_rtl_loop_done): ...new pass_data instance and...
7489 (make_pass_rtl_loop_done): ...new function.
7490 (pass_rtl_move_loop_invariants): Convert from a global struct to a
7491 subclass of rtl_opt_pass along with...
7492 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
7493 (make_pass_rtl_move_loop_invariants): ...new function.
7494 (pass_rtl_unswitch): Convert from a global struct to a subclass of
7495 rtl_opt_pass along with...
7496 (pass_data_rtl_unswitch): ...new pass_data instance and...
7497 (make_pass_rtl_unswitch): ...new function.
7498 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
7499 subclass of rtl_opt_pass along with...
7500 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
7502 (make_pass_rtl_unroll_and_peel_loops): ...new function.
7503 (pass_rtl_doloop): Convert from a global struct to a subclass of
7504 rtl_opt_pass along with...
7505 (pass_data_rtl_doloop): ...new pass_data instance and...
7506 (make_pass_rtl_doloop): ...new function.
7507 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
7508 a subclass of rtl_opt_pass along with...
7509 (pass_data_lower_subreg): ...new pass_data instance and...
7510 (make_pass_lower_subreg): ...new function.
7511 (pass_lower_subreg2): Convert from a global struct to a subclass of
7512 rtl_opt_pass along with...
7513 (pass_data_lower_subreg2): ...new pass_data instance and...
7514 (make_pass_lower_subreg2): ...new function.
7515 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
7516 struct to a subclass of ipa_opt_pass_d along with...
7517 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
7518 (make_pass_ipa_lto_gimple_out): ...new function.
7519 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
7520 of ipa_opt_pass_d along with...
7521 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
7522 (make_pass_ipa_lto_finish_out): ...new function.
7523 * mode-switching.c (pass_mode_switching): Convert from a global struct
7524 to a subclass of rtl_opt_pass along with...
7525 (pass_data_mode_switching): ...new pass_data instance and...
7526 (make_pass_mode_switching): ...new function.
7527 * modulo-sched.c (pass_sms): Convert from a global struct to a
7528 subclass of rtl_opt_pass along with...
7529 (pass_data_sms): ...new pass_data instance and...
7530 (make_pass_sms): ...new function.
7531 * omp-low.c (pass_expand_omp): Convert from a global struct to a
7532 subclass of gimple_opt_pass along with...
7533 (pass_data_expand_omp): ...new pass_data instance and...
7534 (make_pass_expand_omp): ...new function.
7535 (pass_lower_omp): Convert from a global struct to a subclass of
7536 gimple_opt_pass along with...
7537 (pass_data_lower_omp): ...new pass_data instance and...
7538 (make_pass_lower_omp): ...new function.
7539 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
7540 of gimple_opt_pass along with...
7541 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
7542 (make_pass_diagnose_omp_blocks): ...new function.
7543 * passes.c (pass_early_local_passes): Convert from a global struct to
7544 a subclass of simple_ipa_opt_pass along with...
7545 (pass_data_early_local_passes): ...new pass_data instance and...
7546 (make_pass_early_local_passes): ...new function.
7547 (pass_all_early_optimizations): Convert from a global struct to a
7548 subclass of gimple_opt_pass along with...
7549 (pass_data_all_early_optimizations): ...new pass_data instance and...
7550 (make_pass_all_early_optimizations): ...new function.
7551 (pass_all_optimizations): Convert from a global struct to a subclass
7552 of gimple_opt_pass along with...
7553 (pass_data_all_optimizations): ...new pass_data instance and...
7554 (make_pass_all_optimizations): ...new function.
7555 (pass_all_optimizations_g): Convert from a global struct to a subclass
7556 of gimple_opt_pass along with...
7557 (pass_data_all_optimizations_g): ...new pass_data instance and...
7558 (make_pass_all_optimizations_g): ...new function.
7559 (pass_rest_of_compilation): Convert from a global struct to a subclass
7560 of rtl_opt_pass along with...
7561 (pass_data_rest_of_compilation): ...new pass_data instance and...
7562 (make_pass_rest_of_compilation): ...new function.
7563 (pass_postreload): Convert from a global struct to a subclass of
7564 rtl_opt_pass along with...
7565 (pass_data_postreload): ...new pass_data instance and...
7566 (make_pass_postreload): ...new function.
7567 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
7568 subclass of rtl_opt_pass along with...
7569 (pass_data_gcse2): ...new pass_data instance and...
7570 (make_pass_gcse2): ...new function.
7571 * postreload.c (pass_postreload_cse): Convert from a global struct to
7572 a subclass of rtl_opt_pass along with...
7573 (pass_data_postreload_cse): ...new pass_data instance and...
7574 (make_pass_postreload_cse): ...new function.
7575 * predict.c (pass_profile): Convert from a global struct to a subclass
7576 of gimple_opt_pass along with...
7577 (pass_data_profile): ...new pass_data instance and...
7578 (make_pass_profile): ...new function.
7579 (pass_strip_predict_hints): Convert from a global struct to a subclass
7580 of gimple_opt_pass along with...
7581 (pass_data_strip_predict_hints): ...new pass_data instance and...
7582 (make_pass_strip_predict_hints): ...new function.
7583 * recog.c (pass_peephole2): Convert from a global struct to a subclass
7584 of rtl_opt_pass along with...
7585 (pass_data_peephole2): ...new pass_data instance and...
7586 (make_pass_peephole2): ...new function.
7587 (pass_split_all_insns): Convert from a global struct to a subclass of
7588 rtl_opt_pass along with...
7589 (pass_data_split_all_insns): ...new pass_data instance and...
7590 (make_pass_split_all_insns): ...new function.
7591 (pass_split_after_reload): Convert from a global struct to a subclass
7592 of rtl_opt_pass along with...
7593 (pass_data_split_after_reload): ...new pass_data instance and...
7594 (make_pass_split_after_reload): ...new function.
7595 (pass_split_before_regstack): Convert from a global struct to a
7596 subclass of rtl_opt_pass along with...
7597 (pass_data_split_before_regstack): ...new pass_data instance and...
7598 (make_pass_split_before_regstack): ...new function.
7599 (pass_split_before_sched2): Convert from a global struct to a subclass
7600 of rtl_opt_pass along with...
7601 (pass_data_split_before_sched2): ...new pass_data instance and...
7602 (make_pass_split_before_sched2): ...new function.
7603 (pass_split_for_shorten_branches): Convert from a global struct to a
7604 subclass of rtl_opt_pass along with...
7605 (pass_data_split_for_shorten_branches): ...new pass_data instance
7607 (make_pass_split_for_shorten_branches): ...new function.
7608 * ree.c (pass_ree): Convert from a global struct to a subclass of
7609 rtl_opt_pass along with...
7610 (pass_data_ree): ...new pass_data instance and...
7611 (make_pass_ree): ...new function.
7612 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
7613 subclass of rtl_opt_pass along with...
7614 (pass_data_stack_regs): ...new pass_data instance and...
7615 (make_pass_stack_regs): ...new function.
7616 (pass_stack_regs_run): Convert from a global struct to a subclass of
7617 rtl_opt_pass along with...
7618 (pass_data_stack_regs_run): ...new pass_data instance and...
7619 (make_pass_stack_regs_run): ...new function.
7620 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
7621 subclass of rtl_opt_pass along with...
7622 (pass_data_cprop_hardreg): ...new pass_data instance and...
7623 (make_pass_cprop_hardreg): ...new function.
7624 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
7625 subclass of rtl_opt_pass along with...
7626 (pass_data_reginfo_init): ...new pass_data instance and...
7627 (make_pass_reginfo_init): ...new function.
7628 * regmove.c (pass_regmove): Convert from a global struct to a subclass
7629 of rtl_opt_pass along with...
7630 (pass_data_regmove): ...new pass_data instance and...
7631 (make_pass_regmove): ...new function.
7632 * regrename.c (pass_regrename): Convert from a global struct to a
7633 subclass of rtl_opt_pass along with...
7634 (pass_data_regrename): ...new pass_data instance and...
7635 (make_pass_regrename): ...new function.
7636 * reorg.c (pass_delay_slots): Convert from a global struct to a
7637 subclass of rtl_opt_pass along with...
7638 (pass_data_delay_slots): ...new pass_data instance and...
7639 (make_pass_delay_slots): ...new function.
7640 (pass_machine_reorg): Convert from a global struct to a subclass of
7641 rtl_opt_pass along with...
7642 (pass_data_machine_reorg): ...new pass_data instance and...
7643 (make_pass_machine_reorg): ...new function.
7644 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
7645 of rtl_opt_pass along with...
7646 (pass_data_sched): ...new pass_data instance and...
7647 (make_pass_sched): ...new function.
7648 (pass_sched2): Convert from a global struct to a subclass of
7649 rtl_opt_pass along with...
7650 (pass_data_sched2): ...new pass_data instance and...
7651 (make_pass_sched2): ...new function.
7652 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
7653 to a subclass of rtl_opt_pass along with...
7654 (pass_data_stack_ptr_mod): ...new pass_data instance and...
7655 (make_pass_stack_ptr_mod): ...new function.
7656 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
7657 to a subclass of rtl_opt_pass along with...
7658 (pass_data_rtl_store_motion): ...new pass_data instance and...
7659 (make_pass_rtl_store_motion): ...new function.
7660 * tracer.c (pass_tracer): Convert from a global struct to a subclass
7661 of gimple_opt_pass along with...
7662 (pass_data_tracer): ...new pass_data instance and...
7663 (make_pass_tracer): ...new function.
7664 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
7665 to a subclass of gimple_opt_pass along with...
7666 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
7667 (make_pass_diagnose_tm_blocks): ...new function.
7668 (pass_lower_tm): Convert from a global struct to a subclass of
7669 gimple_opt_pass along with...
7670 (pass_data_lower_tm): ...new pass_data instance and...
7671 (make_pass_lower_tm): ...new function.
7672 (pass_tm_init): Convert from a global struct to a subclass of
7673 gimple_opt_pass along with...
7674 (pass_data_tm_init): ...new pass_data instance and...
7675 (make_pass_tm_init): ...new function.
7676 (pass_tm_mark): Convert from a global struct to a subclass of
7677 gimple_opt_pass along with...
7678 (pass_data_tm_mark): ...new pass_data instance and...
7679 (make_pass_tm_mark): ...new function.
7680 (pass_tm_edges): Convert from a global struct to a subclass of
7681 gimple_opt_pass along with...
7682 (pass_data_tm_edges): ...new pass_data instance and...
7683 (make_pass_tm_edges): ...new function.
7684 (pass_tm_memopt): Convert from a global struct to a subclass of
7685 gimple_opt_pass along with...
7686 (pass_data_tm_memopt): ...new pass_data instance and...
7687 (make_pass_tm_memopt): ...new function.
7688 (pass_ipa_tm): Convert from a global struct to a subclass of
7689 simple_ipa_opt_pass along with...
7690 (pass_data_ipa_tm): ...new pass_data instance and...
7691 (make_pass_ipa_tm): ...new function.
7692 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
7693 subclass of gimple_opt_pass along with...
7694 (pass_data_call_cdce): ...new pass_data instance and...
7695 (make_pass_call_cdce): ...new function.
7696 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
7697 subclass of gimple_opt_pass along with...
7698 (pass_data_build_cfg): ...new pass_data instance and...
7699 (make_pass_build_cfg): ...new function.
7700 (pass_split_crit_edges): Convert from a global struct to a subclass of
7701 gimple_opt_pass along with...
7702 (pass_data_split_crit_edges): ...new pass_data instance and...
7703 (make_pass_split_crit_edges): ...new function.
7704 (pass_warn_function_return): Convert from a global struct to a
7705 subclass of gimple_opt_pass along with...
7706 (pass_data_warn_function_return): ...new pass_data instance and...
7707 (make_pass_warn_function_return): ...new function.
7708 (pass_warn_function_noreturn): Convert from a global struct to a
7709 subclass of gimple_opt_pass along with...
7710 (pass_data_warn_function_noreturn): ...new pass_data instance and...
7711 (make_pass_warn_function_noreturn): ...new function.
7712 (pass_warn_unused_result): Convert from a global struct to a subclass
7713 of gimple_opt_pass along with...
7714 (pass_data_warn_unused_result): ...new pass_data instance and...
7715 (make_pass_warn_unused_result): ...new function.
7716 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
7717 a subclass of gimple_opt_pass along with...
7718 (pass_data_merge_phi): ...new pass_data instance and...
7719 (make_pass_merge_phi): ...new function.
7720 * tree-complex.c (pass_lower_complex): Convert from a global struct to
7721 a subclass of gimple_opt_pass along with...
7722 (pass_data_lower_complex): ...new pass_data instance and...
7723 (make_pass_lower_complex): ...new function.
7724 (pass_lower_complex_O0): Convert from a global struct to a subclass of
7725 gimple_opt_pass along with...
7726 (pass_data_lower_complex_O0): ...new pass_data instance and...
7727 (make_pass_lower_complex_O0): ...new function.
7728 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
7729 subclass of gimple_opt_pass along with...
7730 (pass_data_lower_eh): ...new pass_data instance and...
7731 (make_pass_lower_eh): ...new function.
7732 (pass_refactor_eh): Convert from a global struct to a subclass of
7733 gimple_opt_pass along with...
7734 (pass_data_refactor_eh): ...new pass_data instance and...
7735 (make_pass_refactor_eh): ...new function.
7736 (pass_lower_resx): Convert from a global struct to a subclass of
7737 gimple_opt_pass along with...
7738 (pass_data_lower_resx): ...new pass_data instance and...
7739 (make_pass_lower_resx): ...new function.
7740 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
7741 of gimple_opt_pass along with...
7742 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
7743 (make_pass_lower_eh_dispatch): ...new function.
7744 (pass_cleanup_eh): Convert from a global struct to a subclass of
7745 gimple_opt_pass along with...
7746 (pass_data_cleanup_eh): ...new pass_data instance and...
7747 (make_pass_cleanup_eh): ...new function.
7748 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
7749 to a subclass of simple_ipa_opt_pass along with...
7750 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
7751 (make_pass_ipa_lower_emutls): ...new function.
7752 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
7753 a subclass of gimple_opt_pass along with...
7754 (pass_data_if_conversion): ...new pass_data instance and...
7755 (make_pass_if_conversion): ...new function.
7756 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
7757 subclass of gimple_opt_pass along with...
7758 (pass_data_build_ssa): ...new pass_data instance and...
7759 (make_pass_build_ssa): ...new function.
7760 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
7761 global struct to a subclass of gimple_opt_pass along with...
7762 (pass_data_loop_distribution): ...new pass_data instance and...
7763 (make_pass_loop_distribution): ...new function.
7764 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
7765 subclass of gimple_opt_pass along with...
7766 (pass_data_mudflap_1): ...new pass_data instance and...
7767 (make_pass_mudflap_1): ...new function.
7768 (pass_mudflap_2): Convert from a global struct to a subclass of
7769 gimple_opt_pass along with...
7770 (pass_data_mudflap_2): ...new pass_data instance and...
7771 (make_pass_mudflap_2): ...new function.
7772 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
7773 subclass of gimple_opt_pass along with...
7774 (pass_data_mudflap_1): ...new pass_data instance and...
7775 (make_pass_mudflap_1): ...new function.
7776 (pass_mudflap_2): Convert from a global struct to a subclass of
7777 gimple_opt_pass along with...
7778 (pass_data_mudflap_2): ...new pass_data instance and...
7779 (make_pass_mudflap_2): ...new function.
7780 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
7781 gimple_opt_pass along with...
7782 (pass_data_nrv): ...new pass_data instance and...
7783 (make_pass_nrv): ...new function.
7784 (pass_return_slot): Convert from a global struct to a subclass of
7785 gimple_opt_pass along with...
7786 (pass_data_return_slot): ...new pass_data instance and...
7787 (make_pass_return_slot): ...new function.
7788 * tree-object-size.c (pass_object_sizes): Convert from a global struct
7789 to a subclass of gimple_opt_pass along with...
7790 (pass_data_object_sizes): ...new pass_data instance and...
7791 (make_pass_object_sizes): ...new function.
7792 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
7793 global struct to a subclass of gimple_opt_pass along with...
7794 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
7796 (make_pass_cleanup_cfg_post_optimizing): ...new function.
7797 (pass_fixup_cfg): Convert from a global struct to a subclass of
7798 gimple_opt_pass along with...
7799 (pass_data_fixup_cfg): ...new pass_data instance and...
7800 (make_pass_fixup_cfg): ...new function.
7801 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
7802 (make_pass_mudflap_1): ...new function.
7803 (pass_mudflap_2): Replace declaration with that of...
7804 (make_pass_mudflap_2): ...new function.
7805 (pass_asan): Replace declaration with that of...
7806 (make_pass_asan): ...new function.
7807 (pass_asan_O0): Replace declaration with that of...
7808 (make_pass_asan_O0): ...new function.
7809 (pass_tsan): Replace declaration with that of...
7810 (make_pass_tsan): ...new function.
7811 (pass_tsan_O0): Replace declaration with that of...
7812 (make_pass_tsan_O0): ...new function.
7813 (pass_lower_cf): Replace declaration with that of...
7814 (make_pass_lower_cf): ...new function.
7815 (pass_refactor_eh): Replace declaration with that of...
7816 (make_pass_refactor_eh): ...new function.
7817 (pass_lower_eh): Replace declaration with that of...
7818 (make_pass_lower_eh): ...new function.
7819 (pass_lower_eh_dispatch): Replace declaration with that of...
7820 (make_pass_lower_eh_dispatch): ...new function.
7821 (pass_lower_resx): Replace declaration with that of...
7822 (make_pass_lower_resx): ...new function.
7823 (pass_build_cfg): Replace declaration with that of...
7824 (make_pass_build_cfg): ...new function.
7825 (pass_early_tree_profile): Replace declaration with that of...
7826 (make_pass_early_tree_profile): ...new function.
7827 (pass_cleanup_eh): Replace declaration with that of...
7828 (make_pass_cleanup_eh): ...new function.
7829 (pass_sra): Replace declaration with that of...
7830 (make_pass_sra): ...new function.
7831 (pass_sra_early): Replace declaration with that of...
7832 (make_pass_sra_early): ...new function.
7833 (pass_early_ipa_sra): Replace declaration with that of...
7834 (make_pass_early_ipa_sra): ...new function.
7835 (pass_tail_recursion): Replace declaration with that of...
7836 (make_pass_tail_recursion): ...new function.
7837 (pass_tail_calls): Replace declaration with that of...
7838 (make_pass_tail_calls): ...new function.
7839 (pass_tree_loop): Replace declaration with that of...
7840 (make_pass_tree_loop): ...new function.
7841 (pass_tree_loop_init): Replace declaration with that of...
7842 (make_pass_tree_loop_init): ...new function.
7843 (pass_lim): Replace declaration with that of...
7844 (make_pass_lim): ...new function.
7845 (pass_tree_unswitch): Replace declaration with that of...
7846 (make_pass_tree_unswitch): ...new function.
7847 (pass_predcom): Replace declaration with that of...
7848 (make_pass_predcom): ...new function.
7849 (pass_iv_canon): Replace declaration with that of...
7850 (make_pass_iv_canon): ...new function.
7851 (pass_scev_cprop): Replace declaration with that of...
7852 (make_pass_scev_cprop): ...new function.
7853 (pass_empty_loop): Replace declaration with that of...
7854 (make_pass_empty_loop): ...new function.
7855 (pass_record_bounds): Replace declaration with that of...
7856 (make_pass_record_bounds): ...new function.
7857 (pass_graphite): Replace declaration with that of...
7858 (make_pass_graphite): ...new function.
7859 (pass_graphite_transforms): Replace declaration with that of...
7860 (make_pass_graphite_transforms): ...new function.
7861 (pass_if_conversion): Replace declaration with that of...
7862 (make_pass_if_conversion): ...new function.
7863 (pass_loop_distribution): Replace declaration with that of...
7864 (make_pass_loop_distribution): ...new function.
7865 (pass_vectorize): Replace declaration with that of...
7866 (make_pass_vectorize): ...new function.
7867 (pass_slp_vectorize): Replace declaration with that of...
7868 (make_pass_slp_vectorize): ...new function.
7869 (pass_complete_unroll): Replace declaration with that of...
7870 (make_pass_complete_unroll): ...new function.
7871 (pass_complete_unrolli): Replace declaration with that of...
7872 (make_pass_complete_unrolli): ...new function.
7873 (pass_parallelize_loops): Replace declaration with that of...
7874 (make_pass_parallelize_loops): ...new function.
7875 (pass_loop_prefetch): Replace declaration with that of...
7876 (make_pass_loop_prefetch): ...new function.
7877 (pass_iv_optimize): Replace declaration with that of...
7878 (make_pass_iv_optimize): ...new function.
7879 (pass_tree_loop_done): Replace declaration with that of...
7880 (make_pass_tree_loop_done): ...new function.
7881 (pass_ch): Replace declaration with that of...
7882 (make_pass_ch): ...new function.
7883 (pass_ccp): Replace declaration with that of...
7884 (make_pass_ccp): ...new function.
7885 (pass_phi_only_cprop): Replace declaration with that of...
7886 (make_pass_phi_only_cprop): ...new function.
7887 (pass_build_ssa): Replace declaration with that of...
7888 (make_pass_build_ssa): ...new function.
7889 (pass_build_alias): Replace declaration with that of...
7890 (make_pass_build_alias): ...new function.
7891 (pass_build_ealias): Replace declaration with that of...
7892 (make_pass_build_ealias): ...new function.
7893 (pass_dominator): Replace declaration with that of...
7894 (make_pass_dominator): ...new function.
7895 (pass_dce): Replace declaration with that of...
7896 (make_pass_dce): ...new function.
7897 (pass_dce_loop): Replace declaration with that of...
7898 (make_pass_dce_loop): ...new function.
7899 (pass_cd_dce): Replace declaration with that of...
7900 (make_pass_cd_dce): ...new function.
7901 (pass_call_cdce): Replace declaration with that of...
7902 (make_pass_call_cdce): ...new function.
7903 (pass_merge_phi): Replace declaration with that of...
7904 (make_pass_merge_phi): ...new function.
7905 (pass_split_crit_edges): Replace declaration with that of...
7906 (make_pass_split_crit_edges): ...new function.
7907 (pass_pre): Replace declaration with that of...
7908 (make_pass_pre): ...new function.
7909 (pass_profile): Replace declaration with that of...
7910 (make_pass_profile): ...new function.
7911 (pass_strip_predict_hints): Replace declaration with that of...
7912 (make_pass_strip_predict_hints): ...new function.
7913 (pass_lower_complex_O0): Replace declaration with that of...
7914 (make_pass_lower_complex_O0): ...new function.
7915 (pass_lower_complex): Replace declaration with that of...
7916 (make_pass_lower_complex): ...new function.
7917 (pass_lower_vector): Replace declaration with that of...
7918 (make_pass_lower_vector): ...new function.
7919 (pass_lower_vector_ssa): Replace declaration with that of...
7920 (make_pass_lower_vector_ssa): ...new function.
7921 (pass_lower_omp): Replace declaration with that of...
7922 (make_pass_lower_omp): ...new function.
7923 (pass_diagnose_omp_blocks): Replace declaration with that of...
7924 (make_pass_diagnose_omp_blocks): ...new function.
7925 (pass_expand_omp): Replace declaration with that of...
7926 (make_pass_expand_omp): ...new function.
7927 (pass_expand_omp_ssa): Replace declaration with that of...
7928 (make_pass_expand_omp_ssa): ...new function.
7929 (pass_object_sizes): Replace declaration with that of...
7930 (make_pass_object_sizes): ...new function.
7931 (pass_strlen): Replace declaration with that of...
7932 (make_pass_strlen): ...new function.
7933 (pass_fold_builtins): Replace declaration with that of...
7934 (make_pass_fold_builtins): ...new function.
7935 (pass_stdarg): Replace declaration with that of...
7936 (make_pass_stdarg): ...new function.
7937 (pass_early_warn_uninitialized): Replace declaration with that of...
7938 (make_pass_early_warn_uninitialized): ...new function.
7939 (pass_late_warn_uninitialized): Replace declaration with that of...
7940 (make_pass_late_warn_uninitialized): ...new function.
7941 (pass_cse_reciprocals): Replace declaration with that of...
7942 (make_pass_cse_reciprocals): ...new function.
7943 (pass_cse_sincos): Replace declaration with that of...
7944 (make_pass_cse_sincos): ...new function.
7945 (pass_optimize_bswap): Replace declaration with that of...
7946 (make_pass_optimize_bswap): ...new function.
7947 (pass_optimize_widening_mul): Replace declaration with that of...
7948 (make_pass_optimize_widening_mul): ...new function.
7949 (pass_warn_function_return): Replace declaration with that of...
7950 (make_pass_warn_function_return): ...new function.
7951 (pass_warn_function_noreturn): Replace declaration with that of...
7952 (make_pass_warn_function_noreturn): ...new function.
7953 (pass_cselim): Replace declaration with that of...
7954 (make_pass_cselim): ...new function.
7955 (pass_phiopt): Replace declaration with that of...
7956 (make_pass_phiopt): ...new function.
7957 (pass_forwprop): Replace declaration with that of...
7958 (make_pass_forwprop): ...new function.
7959 (pass_phiprop): Replace declaration with that of...
7960 (make_pass_phiprop): ...new function.
7961 (pass_tree_ifcombine): Replace declaration with that of...
7962 (make_pass_tree_ifcombine): ...new function.
7963 (pass_dse): Replace declaration with that of...
7964 (make_pass_dse): ...new function.
7965 (pass_nrv): Replace declaration with that of...
7966 (make_pass_nrv): ...new function.
7967 (pass_rename_ssa_copies): Replace declaration with that of...
7968 (make_pass_rename_ssa_copies): ...new function.
7969 (pass_sink_code): Replace declaration with that of...
7970 (make_pass_sink_code): ...new function.
7971 (pass_fre): Replace declaration with that of...
7972 (make_pass_fre): ...new function.
7973 (pass_check_data_deps): Replace declaration with that of...
7974 (make_pass_check_data_deps): ...new function.
7975 (pass_copy_prop): Replace declaration with that of...
7976 (make_pass_copy_prop): ...new function.
7977 (pass_vrp): Replace declaration with that of...
7978 (make_pass_vrp): ...new function.
7979 (pass_uncprop): Replace declaration with that of...
7980 (make_pass_uncprop): ...new function.
7981 (pass_return_slot): Replace declaration with that of...
7982 (make_pass_return_slot): ...new function.
7983 (pass_reassoc): Replace declaration with that of...
7984 (make_pass_reassoc): ...new function.
7985 (pass_rebuild_cgraph_edges): Replace declaration with that of...
7986 (make_pass_rebuild_cgraph_edges): ...new function.
7987 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
7988 (make_pass_remove_cgraph_callee_edges): ...new function.
7989 (pass_build_cgraph_edges): Replace declaration with that of...
7990 (make_pass_build_cgraph_edges): ...new function.
7991 (pass_local_pure_const): Replace declaration with that of...
7992 (make_pass_local_pure_const): ...new function.
7993 (pass_tracer): Replace declaration with that of...
7994 (make_pass_tracer): ...new function.
7995 (pass_warn_unused_result): Replace declaration with that of...
7996 (make_pass_warn_unused_result): ...new function.
7997 (pass_diagnose_tm_blocks): Replace declaration with that of...
7998 (make_pass_diagnose_tm_blocks): ...new function.
7999 (pass_lower_tm): Replace declaration with that of...
8000 (make_pass_lower_tm): ...new function.
8001 (pass_tm_init): Replace declaration with that of...
8002 (make_pass_tm_init): ...new function.
8003 (pass_tm_mark): Replace declaration with that of...
8004 (make_pass_tm_mark): ...new function.
8005 (pass_tm_memopt): Replace declaration with that of...
8006 (make_pass_tm_memopt): ...new function.
8007 (pass_tm_edges): Replace declaration with that of...
8008 (make_pass_tm_edges): ...new function.
8009 (pass_split_functions): Replace declaration with that of...
8010 (make_pass_split_functions): ...new function.
8011 (pass_feedback_split_functions): Replace declaration with that of...
8012 (make_pass_feedback_split_functions): ...new function.
8013 (pass_strength_reduction): Replace declaration with that of...
8014 (make_pass_strength_reduction): ...new function.
8015 (pass_ipa_lower_emutls): Replace declaration with that of...
8016 (make_pass_ipa_lower_emutls): ...new function.
8017 (pass_ipa_function_and_variable_visibility): Replace declaration with
8019 (make_pass_ipa_function_and_variable_visibility): ...new function.
8020 (pass_ipa_tree_profile): Replace declaration with that of...
8021 (make_pass_ipa_tree_profile): ...new function.
8022 (pass_early_local_passes): Replace declaration with that of...
8023 (make_pass_early_local_passes): ...new function.
8024 (pass_ipa_whole_program_visibility): Replace declaration with that
8026 (make_pass_ipa_whole_program_visibility): ...new function.
8027 (pass_ipa_lto_gimple_out): Replace declaration with that of...
8028 (make_pass_ipa_lto_gimple_out): ...new function.
8029 (pass_ipa_increase_alignment): Replace declaration with that of...
8030 (make_pass_ipa_increase_alignment): ...new function.
8031 (pass_ipa_inline): Replace declaration with that of...
8032 (make_pass_ipa_inline): ...new function.
8033 (pass_ipa_free_lang_data): Replace declaration with that of...
8034 (make_pass_ipa_free_lang_data): ...new function.
8035 (pass_ipa_free_inline_summary): Replace declaration with that of...
8036 (make_pass_ipa_free_inline_summary): ...new function.
8037 (pass_ipa_cp): Replace declaration with that of...
8038 (make_pass_ipa_cp): ...new function.
8039 (pass_ipa_reference): Replace declaration with that of...
8040 (make_pass_ipa_reference): ...new function.
8041 (pass_ipa_pure_const): Replace declaration with that of...
8042 (make_pass_ipa_pure_const): ...new function.
8043 (pass_ipa_pta): Replace declaration with that of...
8044 (make_pass_ipa_pta): ...new function.
8045 (pass_ipa_lto_finish_out): Replace declaration with that of...
8046 (make_pass_ipa_lto_finish_out): ...new function.
8047 (pass_ipa_tm): Replace declaration with that of...
8048 (make_pass_ipa_tm): ...new function.
8049 (pass_ipa_profile): Replace declaration with that of...
8050 (make_pass_ipa_profile): ...new function.
8051 (pass_ipa_cdtor_merge): Replace declaration with that of...
8052 (make_pass_ipa_cdtor_merge): ...new function.
8053 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
8055 (make_pass_cleanup_cfg_post_optimizing): ...new function.
8056 (pass_init_datastructures): Replace declaration with that of...
8057 (make_pass_init_datastructures): ...new function.
8058 (pass_fixup_cfg): Replace declaration with that of...
8059 (make_pass_fixup_cfg): ...new function.
8060 (pass_expand): Replace declaration with that of...
8061 (make_pass_expand): ...new function.
8062 (pass_instantiate_virtual_regs): Replace declaration with that of...
8063 (make_pass_instantiate_virtual_regs): ...new function.
8064 (pass_rtl_fwprop): Replace declaration with that of...
8065 (make_pass_rtl_fwprop): ...new function.
8066 (pass_rtl_fwprop_addr): Replace declaration with that of...
8067 (make_pass_rtl_fwprop_addr): ...new function.
8068 (pass_jump): Replace declaration with that of...
8069 (make_pass_jump): ...new function.
8070 (pass_jump2): Replace declaration with that of...
8071 (make_pass_jump2): ...new function.
8072 (pass_lower_subreg): Replace declaration with that of...
8073 (make_pass_lower_subreg): ...new function.
8074 (pass_cse): Replace declaration with that of...
8075 (make_pass_cse): ...new function.
8076 (pass_fast_rtl_dce): Replace declaration with that of...
8077 (make_pass_fast_rtl_dce): ...new function.
8078 (pass_ud_rtl_dce): Replace declaration with that of...
8079 (make_pass_ud_rtl_dce): ...new function.
8080 (pass_rtl_dce): Replace declaration with that of...
8081 (make_pass_rtl_dce): ...new function.
8082 (pass_rtl_dse1): Replace declaration with that of...
8083 (make_pass_rtl_dse1): ...new function.
8084 (pass_rtl_dse2): Replace declaration with that of...
8085 (make_pass_rtl_dse2): ...new function.
8086 (pass_rtl_dse3): Replace declaration with that of...
8087 (make_pass_rtl_dse3): ...new function.
8088 (pass_rtl_cprop): Replace declaration with that of...
8089 (make_pass_rtl_cprop): ...new function.
8090 (pass_rtl_pre): Replace declaration with that of...
8091 (make_pass_rtl_pre): ...new function.
8092 (pass_rtl_hoist): Replace declaration with that of...
8093 (make_pass_rtl_hoist): ...new function.
8094 (pass_rtl_store_motion): Replace declaration with that of...
8095 (make_pass_rtl_store_motion): ...new function.
8096 (pass_cse_after_global_opts): Replace declaration with that of...
8097 (make_pass_cse_after_global_opts): ...new function.
8098 (pass_rtl_ifcvt): Replace declaration with that of...
8099 (make_pass_rtl_ifcvt): ...new function.
8100 (pass_into_cfg_layout_mode): Replace declaration with that of...
8101 (make_pass_into_cfg_layout_mode): ...new function.
8102 (pass_outof_cfg_layout_mode): Replace declaration with that of...
8103 (make_pass_outof_cfg_layout_mode): ...new function.
8104 (pass_loop2): Replace declaration with that of...
8105 (make_pass_loop2): ...new function.
8106 (pass_rtl_loop_init): Replace declaration with that of...
8107 (make_pass_rtl_loop_init): ...new function.
8108 (pass_rtl_move_loop_invariants): Replace declaration with that of...
8109 (make_pass_rtl_move_loop_invariants): ...new function.
8110 (pass_rtl_unswitch): Replace declaration with that of...
8111 (make_pass_rtl_unswitch): ...new function.
8112 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
8113 (make_pass_rtl_unroll_and_peel_loops): ...new function.
8114 (pass_rtl_doloop): Replace declaration with that of...
8115 (make_pass_rtl_doloop): ...new function.
8116 (pass_rtl_loop_done): Replace declaration with that of...
8117 (make_pass_rtl_loop_done): ...new function.
8118 (pass_web): Replace declaration with that of...
8119 (make_pass_web): ...new function.
8120 (pass_cse2): Replace declaration with that of...
8121 (make_pass_cse2): ...new function.
8122 (pass_df_initialize_opt): Replace declaration with that of...
8123 (make_pass_df_initialize_opt): ...new function.
8124 (pass_df_initialize_no_opt): Replace declaration with that of...
8125 (make_pass_df_initialize_no_opt): ...new function.
8126 (pass_reginfo_init): Replace declaration with that of...
8127 (make_pass_reginfo_init): ...new function.
8128 (pass_inc_dec): Replace declaration with that of...
8129 (make_pass_inc_dec): ...new function.
8130 (pass_stack_ptr_mod): Replace declaration with that of...
8131 (make_pass_stack_ptr_mod): ...new function.
8132 (pass_initialize_regs): Replace declaration with that of...
8133 (make_pass_initialize_regs): ...new function.
8134 (pass_combine): Replace declaration with that of...
8135 (make_pass_combine): ...new function.
8136 (pass_if_after_combine): Replace declaration with that of...
8137 (make_pass_if_after_combine): ...new function.
8138 (pass_ree): Replace declaration with that of...
8139 (make_pass_ree): ...new function.
8140 (pass_partition_blocks): Replace declaration with that of...
8141 (make_pass_partition_blocks): ...new function.
8142 (pass_match_asm_constraints): Replace declaration with that of...
8143 (make_pass_match_asm_constraints): ...new function.
8144 (pass_regmove): Replace declaration with that of...
8145 (make_pass_regmove): ...new function.
8146 (pass_split_all_insns): Replace declaration with that of...
8147 (make_pass_split_all_insns): ...new function.
8148 (pass_fast_rtl_byte_dce): Replace declaration with that of...
8149 (make_pass_fast_rtl_byte_dce): ...new function.
8150 (pass_lower_subreg2): Replace declaration with that of...
8151 (make_pass_lower_subreg2): ...new function.
8152 (pass_mode_switching): Replace declaration with that of...
8153 (make_pass_mode_switching): ...new function.
8154 (pass_sms): Replace declaration with that of...
8155 (make_pass_sms): ...new function.
8156 (pass_sched): Replace declaration with that of...
8157 (make_pass_sched): ...new function.
8158 (pass_ira): Replace declaration with that of...
8159 (make_pass_ira): ...new function.
8160 (pass_reload): Replace declaration with that of...
8161 (make_pass_reload): ...new function.
8162 (pass_clean_state): Replace declaration with that of...
8163 (make_pass_clean_state): ...new function.
8164 (pass_branch_prob): Replace declaration with that of...
8165 (make_pass_branch_prob): ...new function.
8166 (pass_value_profile_transformations): Replace declaration with that
8168 (make_pass_value_profile_transformations): ...new function.
8169 (pass_postreload_cse): Replace declaration with that of...
8170 (make_pass_postreload_cse): ...new function.
8171 (pass_gcse2): Replace declaration with that of...
8172 (make_pass_gcse2): ...new function.
8173 (pass_split_after_reload): Replace declaration with that of...
8174 (make_pass_split_after_reload): ...new function.
8175 (pass_branch_target_load_optimize1): Replace declaration with that
8177 (make_pass_branch_target_load_optimize1): ...new function.
8178 (pass_thread_prologue_and_epilogue): Replace declaration with that
8180 (make_pass_thread_prologue_and_epilogue): ...new function.
8181 (pass_stack_adjustments): Replace declaration with that of...
8182 (make_pass_stack_adjustments): ...new function.
8183 (pass_peephole2): Replace declaration with that of...
8184 (make_pass_peephole2): ...new function.
8185 (pass_if_after_reload): Replace declaration with that of...
8186 (make_pass_if_after_reload): ...new function.
8187 (pass_regrename): Replace declaration with that of...
8188 (make_pass_regrename): ...new function.
8189 (pass_cprop_hardreg): Replace declaration with that of...
8190 (make_pass_cprop_hardreg): ...new function.
8191 (pass_reorder_blocks): Replace declaration with that of...
8192 (make_pass_reorder_blocks): ...new function.
8193 (pass_branch_target_load_optimize2): Replace declaration with that
8195 (make_pass_branch_target_load_optimize2): ...new function.
8196 (pass_leaf_regs): Replace declaration with that of...
8197 (make_pass_leaf_regs): ...new function.
8198 (pass_split_before_sched2): Replace declaration with that of...
8199 (make_pass_split_before_sched2): ...new function.
8200 (pass_compare_elim_after_reload): Replace declaration with that of...
8201 (make_pass_compare_elim_after_reload): ...new function.
8202 (pass_sched2): Replace declaration with that of...
8203 (make_pass_sched2): ...new function.
8204 (pass_stack_regs): Replace declaration with that of...
8205 (make_pass_stack_regs): ...new function.
8206 (pass_stack_regs_run): Replace declaration with that of...
8207 (make_pass_stack_regs_run): ...new function.
8208 (pass_df_finish): Replace declaration with that of...
8209 (make_pass_df_finish): ...new function.
8210 (pass_compute_alignments): Replace declaration with that of...
8211 (make_pass_compute_alignments): ...new function.
8212 (pass_duplicate_computed_gotos): Replace declaration with that of...
8213 (make_pass_duplicate_computed_gotos): ...new function.
8214 (pass_variable_tracking): Replace declaration with that of...
8215 (make_pass_variable_tracking): ...new function.
8216 (pass_free_cfg): Replace declaration with that of...
8217 (make_pass_free_cfg): ...new function.
8218 (pass_machine_reorg): Replace declaration with that of...
8219 (make_pass_machine_reorg): ...new function.
8220 (pass_cleanup_barriers): Replace declaration with that of...
8221 (make_pass_cleanup_barriers): ...new function.
8222 (pass_delay_slots): Replace declaration with that of...
8223 (make_pass_delay_slots): ...new function.
8224 (pass_split_for_shorten_branches): Replace declaration with that of...
8225 (make_pass_split_for_shorten_branches): ...new function.
8226 (pass_split_before_regstack): Replace declaration with that of...
8227 (make_pass_split_before_regstack): ...new function.
8228 (pass_convert_to_eh_region_ranges): Replace declaration with that
8230 (make_pass_convert_to_eh_region_ranges): ...new function.
8231 (pass_shorten_branches): Replace declaration with that of...
8232 (make_pass_shorten_branches): ...new function.
8233 (pass_set_nothrow_function_flags): Replace declaration with that of...
8234 (make_pass_set_nothrow_function_flags): ...new function.
8235 (pass_dwarf2_frame): Replace declaration with that of...
8236 (make_pass_dwarf2_frame): ...new function.
8237 (pass_final): Replace declaration with that of...
8238 (make_pass_final): ...new function.
8239 (pass_rtl_seqabstr): Replace declaration with that of...
8240 (make_pass_rtl_seqabstr): ...new function.
8241 (pass_release_ssa_names): Replace declaration with that of...
8242 (make_pass_release_ssa_names): ...new function.
8243 (pass_early_inline): Replace declaration with that of...
8244 (make_pass_early_inline): ...new function.
8245 (pass_inline_parameters): Replace declaration with that of...
8246 (make_pass_inline_parameters): ...new function.
8247 (pass_update_address_taken): Replace declaration with that of...
8248 (make_pass_update_address_taken): ...new function.
8249 (pass_convert_switch): Replace declaration with that of...
8250 (make_pass_convert_switch): ...new function.
8251 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
8252 to a subclass of simple_ipa_opt_pass along with...
8253 (pass_data_ipa_tree_profile): ...new pass_data instance and...
8254 (make_pass_ipa_tree_profile): ...new function.
8255 * tree-sra.c (pass_sra_early): Convert from a global struct to a
8256 subclass of gimple_opt_pass along with...
8257 (pass_data_sra_early): ...new pass_data instance and...
8258 (make_pass_sra_early): ...new function.
8259 (pass_sra): Convert from a global struct to a subclass of
8260 gimple_opt_pass along with...
8261 (pass_data_sra): ...new pass_data instance and...
8262 (make_pass_sra): ...new function.
8263 (pass_early_ipa_sra): Convert from a global struct to a subclass of
8264 gimple_opt_pass along with...
8265 (pass_data_early_ipa_sra): ...new pass_data instance and...
8266 (make_pass_early_ipa_sra): ...new function.
8267 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
8268 subclass of gimple_opt_pass along with...
8269 (pass_data_ccp): ...new pass_data instance and...
8270 (make_pass_ccp): ...new function.
8271 (pass_fold_builtins): Convert from a global struct to a subclass of
8272 gimple_opt_pass along with...
8273 (pass_data_fold_builtins): ...new pass_data instance and...
8274 (make_pass_fold_builtins): ...new function.
8275 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
8276 subclass of gimple_opt_pass along with...
8277 (pass_data_copy_prop): ...new pass_data instance and...
8278 (make_pass_copy_prop): ...new function.
8279 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
8280 global struct to a subclass of gimple_opt_pass along with...
8281 (pass_data_rename_ssa_copies): ...new pass_data instance and...
8282 (make_pass_rename_ssa_copies): ...new function.
8283 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
8284 subclass of gimple_opt_pass along with...
8285 (pass_data_dce): ...new pass_data instance and...
8286 (make_pass_dce): ...new function.
8287 (pass_dce_loop): Convert from a global struct to a subclass of
8288 gimple_opt_pass along with...
8289 (pass_data_dce_loop): ...new pass_data instance and...
8290 (make_pass_dce_loop): ...new function.
8291 (pass_cd_dce): Convert from a global struct to a subclass of
8292 gimple_opt_pass along with...
8293 (pass_data_cd_dce): ...new pass_data instance and...
8294 (make_pass_cd_dce): ...new function.
8295 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
8296 subclass of gimple_opt_pass along with...
8297 (pass_data_dominator): ...new pass_data instance and...
8298 (make_pass_dominator): ...new function.
8299 (pass_phi_only_cprop): Convert from a global struct to a subclass of
8300 gimple_opt_pass along with...
8301 (pass_data_phi_only_cprop): ...new pass_data instance and...
8302 (make_pass_phi_only_cprop): ...new function.
8303 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
8304 subclass of gimple_opt_pass along with...
8305 (pass_data_dse): ...new pass_data instance and...
8306 (make_pass_dse): ...new function.
8307 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
8308 a subclass of gimple_opt_pass along with...
8309 (pass_data_forwprop): ...new pass_data instance and...
8310 (make_pass_forwprop): ...new function.
8311 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
8312 struct to a subclass of gimple_opt_pass along with...
8313 (pass_data_tree_ifcombine): ...new pass_data instance and...
8314 (make_pass_tree_ifcombine): ...new function.
8315 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
8316 subclass of gimple_opt_pass along with...
8317 (pass_data_ch): ...new pass_data instance and...
8318 (make_pass_ch): ...new function.
8319 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
8320 subclass of gimple_opt_pass along with...
8321 (pass_data_tree_loop): ...new pass_data instance and...
8322 (make_pass_tree_loop): ...new function.
8323 (pass_tree_loop_init): Convert from a global struct to a subclass of
8324 gimple_opt_pass along with...
8325 (pass_data_tree_loop_init): ...new pass_data instance and...
8326 (make_pass_tree_loop_init): ...new function.
8327 (pass_lim): Convert from a global struct to a subclass of
8328 gimple_opt_pass along with...
8329 (pass_data_lim): ...new pass_data instance and...
8330 (make_pass_lim): ...new function.
8331 (pass_tree_unswitch): Convert from a global struct to a subclass of
8332 gimple_opt_pass along with...
8333 (pass_data_tree_unswitch): ...new pass_data instance and...
8334 (make_pass_tree_unswitch): ...new function.
8335 (pass_predcom): Convert from a global struct to a subclass of
8336 gimple_opt_pass along with...
8337 (pass_data_predcom): ...new pass_data instance and...
8338 (make_pass_predcom): ...new function.
8339 (pass_vectorize): Convert from a global struct to a subclass of
8340 gimple_opt_pass along with...
8341 (pass_data_vectorize): ...new pass_data instance and...
8342 (make_pass_vectorize): ...new function.
8343 (pass_graphite): Convert from a global struct to a subclass of
8344 gimple_opt_pass along with...
8345 (pass_data_graphite): ...new pass_data instance and...
8346 (make_pass_graphite): ...new function.
8347 (pass_graphite_transforms): Convert from a global struct to a subclass
8348 of gimple_opt_pass along with...
8349 (pass_data_graphite_transforms): ...new pass_data instance and...
8350 (make_pass_graphite_transforms): ...new function.
8351 (pass_check_data_deps): Convert from a global struct to a subclass of
8352 gimple_opt_pass along with...
8353 (pass_data_check_data_deps): ...new pass_data instance and...
8354 (make_pass_check_data_deps): ...new function.
8355 (pass_iv_canon): Convert from a global struct to a subclass of
8356 gimple_opt_pass along with...
8357 (pass_data_iv_canon): ...new pass_data instance and...
8358 (make_pass_iv_canon): ...new function.
8359 (pass_scev_cprop): Convert from a global struct to a subclass of
8360 gimple_opt_pass along with...
8361 (pass_data_scev_cprop): ...new pass_data instance and...
8362 (make_pass_scev_cprop): ...new function.
8363 (pass_record_bounds): Convert from a global struct to a subclass of
8364 gimple_opt_pass along with...
8365 (pass_data_record_bounds): ...new pass_data instance and...
8366 (make_pass_record_bounds): ...new function.
8367 (pass_complete_unroll): Convert from a global struct to a subclass of
8368 gimple_opt_pass along with...
8369 (pass_data_complete_unroll): ...new pass_data instance and...
8370 (make_pass_complete_unroll): ...new function.
8371 (pass_complete_unrolli): Convert from a global struct to a subclass of
8372 gimple_opt_pass along with...
8373 (pass_data_complete_unrolli): ...new pass_data instance and...
8374 (make_pass_complete_unrolli): ...new function.
8375 (pass_parallelize_loops): Convert from a global struct to a subclass
8376 of gimple_opt_pass along with...
8377 (pass_data_parallelize_loops): ...new pass_data instance and...
8378 (make_pass_parallelize_loops): ...new function.
8379 (pass_loop_prefetch): Convert from a global struct to a subclass of
8380 gimple_opt_pass along with...
8381 (pass_data_loop_prefetch): ...new pass_data instance and...
8382 (make_pass_loop_prefetch): ...new function.
8383 (pass_iv_optimize): Convert from a global struct to a subclass of
8384 gimple_opt_pass along with...
8385 (pass_data_iv_optimize): ...new pass_data instance and...
8386 (make_pass_iv_optimize): ...new function.
8387 (pass_tree_loop_done): Convert from a global struct to a subclass of
8388 gimple_opt_pass along with...
8389 (pass_data_tree_loop_done): ...new pass_data instance and...
8390 (make_pass_tree_loop_done): ...new function.
8391 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
8392 struct to a subclass of gimple_opt_pass along with...
8393 (pass_data_cse_reciprocals): ...new pass_data instance and...
8394 (make_pass_cse_reciprocals): ...new function.
8395 (pass_cse_sincos): Convert from a global struct to a subclass of
8396 gimple_opt_pass along with...
8397 (pass_data_cse_sincos): ...new pass_data instance and...
8398 (make_pass_cse_sincos): ...new function.
8399 (pass_optimize_bswap): Convert from a global struct to a subclass of
8400 gimple_opt_pass along with...
8401 (pass_data_optimize_bswap): ...new pass_data instance and...
8402 (make_pass_optimize_bswap): ...new function.
8403 (pass_optimize_widening_mul): Convert from a global struct to a
8404 subclass of gimple_opt_pass along with...
8405 (pass_data_optimize_widening_mul): ...new pass_data instance and...
8406 (make_pass_optimize_widening_mul): ...new function.
8407 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
8408 subclass of gimple_opt_pass along with...
8409 (pass_data_phiopt): ...new pass_data instance and...
8410 (make_pass_phiopt): ...new function.
8411 (pass_cselim): Convert from a global struct to a subclass of
8412 gimple_opt_pass along with...
8413 (pass_data_cselim): ...new pass_data instance and...
8414 (make_pass_cselim): ...new function.
8415 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
8416 subclass of gimple_opt_pass along with...
8417 (pass_data_phiprop): ...new pass_data instance and...
8418 (make_pass_phiprop): ...new function.
8419 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
8420 subclass of gimple_opt_pass along with...
8421 (pass_data_pre): ...new pass_data instance and...
8422 (make_pass_pre): ...new function.
8423 (pass_fre): Convert from a global struct to a subclass of
8424 gimple_opt_pass along with...
8425 (pass_data_fre): ...new pass_data instance and...
8426 (make_pass_fre): ...new function.
8427 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
8428 subclass of gimple_opt_pass along with...
8429 (pass_data_reassoc): ...new pass_data instance and...
8430 (make_pass_reassoc): ...new function.
8431 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
8432 subclass of gimple_opt_pass along with...
8433 (pass_data_sink_code): ...new pass_data instance and...
8434 (make_pass_sink_code): ...new function.
8435 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
8436 subclass of gimple_opt_pass along with...
8437 (pass_data_strlen): ...new pass_data instance and...
8438 (make_pass_strlen): ...new function.
8439 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
8440 struct to a subclass of gimple_opt_pass along with...
8441 (pass_data_build_alias): ...new pass_data instance and...
8442 (make_pass_build_alias): ...new function.
8443 (pass_build_ealias): Convert from a global struct to a subclass of
8444 gimple_opt_pass along with...
8445 (pass_data_build_ealias): ...new pass_data instance and...
8446 (make_pass_build_ealias): ...new function.
8447 (pass_ipa_pta): Convert from a global struct to a subclass of
8448 simple_ipa_opt_pass along with...
8449 (pass_data_ipa_pta): ...new pass_data instance and...
8450 (make_pass_ipa_pta): ...new function.
8451 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
8452 subclass of gimple_opt_pass along with...
8453 (pass_data_uncprop): ...new pass_data instance and...
8454 (make_pass_uncprop): ...new function.
8455 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
8456 global struct to a subclass of gimple_opt_pass along with...
8457 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
8458 (make_pass_late_warn_uninitialized): ...new function.
8459 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
8460 to a subclass of gimple_opt_pass along with...
8461 (pass_data_init_datastructures): ...new pass_data instance and...
8462 (make_pass_init_datastructures): ...new function.
8463 (pass_early_warn_uninitialized): Convert from a global struct to a
8464 subclass of gimple_opt_pass along with...
8465 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
8466 (make_pass_early_warn_uninitialized): ...new function.
8467 (pass_update_address_taken): Convert from a global struct to a
8468 subclass of gimple_opt_pass along with...
8469 (pass_data_update_address_taken): ...new pass_data instance and...
8470 (make_pass_update_address_taken): ...new function.
8471 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
8472 struct to a subclass of gimple_opt_pass along with...
8473 (pass_data_release_ssa_names): ...new pass_data instance and...
8474 (make_pass_release_ssa_names): ...new function.
8475 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
8476 subclass of gimple_opt_pass along with...
8477 (pass_data_stdarg): ...new pass_data instance and...
8478 (make_pass_stdarg): ...new function.
8479 * tree-switch-conversion.c (pass_convert_switch): Convert from a
8480 global struct to a subclass of gimple_opt_pass along with...
8481 (pass_data_convert_switch): ...new pass_data instance and...
8482 (make_pass_convert_switch): ...new function.
8483 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
8484 to a subclass of gimple_opt_pass along with...
8485 (pass_data_tail_recursion): ...new pass_data instance and...
8486 (make_pass_tail_recursion): ...new function.
8487 (pass_tail_calls): Convert from a global struct to a subclass of
8488 gimple_opt_pass along with...
8489 (pass_data_tail_calls): ...new pass_data instance and...
8490 (make_pass_tail_calls): ...new function.
8491 * tree-vect-generic.c (pass_lower_vector): Convert from a global
8492 struct to a subclass of gimple_opt_pass along with...
8493 (pass_data_lower_vector): ...new pass_data instance and...
8494 (make_pass_lower_vector): ...new function.
8495 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
8496 gimple_opt_pass along with...
8497 (pass_data_lower_vector_ssa): ...new pass_data instance and...
8498 (make_pass_lower_vector_ssa): ...new function.
8499 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
8500 to a subclass of gimple_opt_pass along with...
8501 (pass_data_slp_vectorize): ...new pass_data instance and...
8502 (make_pass_slp_vectorize): ...new function.
8503 (pass_ipa_increase_alignment): Convert from a global struct to a
8504 subclass of simple_ipa_opt_pass along with...
8505 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
8506 (make_pass_ipa_increase_alignment): ...new function.
8507 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
8508 gimple_opt_pass along with...
8509 (pass_data_vrp): ...new pass_data instance and...
8510 (make_pass_vrp): ...new function.
8511 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
8512 subclass of simple_ipa_opt_pass along with...
8513 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
8514 (make_pass_ipa_free_lang_data): ...new function.
8515 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
8516 gimple_opt_pass along with...
8517 (pass_data_tsan): ...new pass_data instance and...
8518 (make_pass_tsan): ...new function.
8519 (pass_tsan_O0): Convert from a global struct to a subclass of
8520 gimple_opt_pass along with...
8521 (pass_data_tsan_O0): ...new pass_data instance and...
8522 (make_pass_tsan_O0): ...new function.
8523 * var-tracking.c (pass_variable_tracking): Convert from a global
8524 struct to a subclass of rtl_opt_pass along with...
8525 (pass_data_variable_tracking): ...new pass_data instance and...
8526 (make_pass_variable_tracking): ...new function.
8527 * web.c (pass_web): Convert from a global struct to a subclass of
8528 rtl_opt_pass along with...
8529 (pass_data_web): ...new pass_data instance and...
8530 (make_pass_web): ...new function.
8531 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
8532 declaration with that of...
8533 (make_pass_mode_switch_use): ...new function.
8534 (pass_resolve_sw_modes): Replace declaration with that of...
8535 (make_pass_resolve_sw_modes): ...new function.
8536 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
8537 from a global struct to a subclass of rtl_opt_pass along with...
8538 (pass_data_mode_switch_use): ...new pass_data instance and...
8539 (make_pass_mode_switch_use): ...new function.
8540 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
8541 from a global struct to a subclass of rtl_opt_pass along with...
8542 (pass_data_resolve_sw_modes): ...new pass_data instance and...
8543 (make_pass_resolve_sw_modes): ...new function.
8544 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
8545 struct to a subclass of rtl_opt_pass along with...
8546 (pass_data_insert_vzeroupper): ...new pass_data instance and...
8547 (make_pass_insert_vzeroupper): ...new function.
8548 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
8549 global struct to a subclass of rtl_opt_pass along with...
8550 (pass_data_work_around_errata): ...new pass_data instance and...
8551 (make_pass_work_around_errata): ...new function.
8552 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
8553 struct to a subclass of rtl_opt_pass along with...
8554 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
8555 (make_pass_mips_machine_reorg2): ...new function.
8557 2013-08-05 David Malcolm <dmalcolm@redhat.com>
8559 * passes.c (pass_manager::operator new): New.
8561 2013-08-05 David Malcolm <dmalcolm@redhat.com>
8563 Handwritten part of conversion of passes to C++ classes.
8565 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
8566 (toplev.o): Add dep on PASS_MANAGER_H.
8567 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
8568 of early local pases to reflect this moving from a global to a
8569 member of gcc::pass_manager.
8570 (cgraph_add_new_function): Likewise.
8571 * lto-cgraph.c (lto_output_node): Update for conversion of
8572 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
8573 * passes.c (opt_pass::clone): New.
8574 (opt_pass::gate): New.
8575 (opt_pass::execute): New.
8576 (opt_pass::opt_pass): New.
8577 (pass_manager::execute_early_local_passes): New.
8578 (pass_manager::execute_pass_mode_switching): new.
8579 (finish_optimization_passes): Convert to...
8580 (pass_manager::finish_optimization_passes): ...this.
8581 (finish_optimization_passes): Update for conversion of passes to
8583 (register_dump_files_1): Use has_gate since we cannot portably
8584 check a vtable entry against NULL.
8585 (dump_one_pass): Likewise.
8586 (ipa_write_summaries_2): Likewise.
8587 (ipa_write_optimization_summaries_1): Likewise.
8588 (ipa_read_summaries_1): Likewise.
8589 (ipa_read_optimization_summaries_1): Likewise.
8590 (execute_ipa_stmt_fixups): Likewise.
8591 (pass_manager::pass_manager): Rewrite pass-creation, invoking
8592 pass-creation functions rather than wiring up globals, and
8593 storing the results in fields of pass_manager generated using
8595 (pass_manager::dump_profile_report): Update for conversion of
8596 passes to C++ classes.
8597 (pass_manager::execute_ipa_summary_passes): Likewise.
8598 (execute_one_ipa_transform_pass): Likewise.
8599 (execute_one_pass): Use has_gate and has_execute since we cannot
8600 portably check a vtable entry against NULL.
8601 * pass_manager.h (pass_manager::finish_optimization_passes): New.
8602 (pass_manager): Use pass-instances.def to add fields for the
8603 various pass instances.
8604 * toplev.c (finalize): Update for move of
8605 finish_optimization_passes to a method of gcc::pass_manager.
8606 * toplev.h (finish_optimization_passes): Move to method of class
8608 * tree-pass.h (struct pass_data): New.
8609 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
8610 (opt_pass::gate): Convert to virtual function.
8611 (opt_pass::~opt_pass): New.
8612 (opt_pass::clone): New.
8613 (opt_pass::execute): Convert to virtual function.
8614 (opt_pass::opt_pass): New.
8615 (opt_pass::ctxt_): new.
8616 (gimple_opt_pass): Convert to subclass of opt_pass.
8617 (gimple_opt_pass::gimple_opt_pass): New.
8618 (rtl_opt_pass): Convert to subclass of opt_pass.
8619 (rtl_opt_pass::rtl_opt_pass): New.
8620 (ipa_opt_pass_d): Convert to subclass of opt_pass.
8621 (ipa_opt_pass_d::ipa_opt_pass_d): New.
8622 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
8623 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
8624 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
8625 invocation of pass_mode_switching to reflect this moving from a
8626 global to a member of gcc::pass_manager.
8627 (ix86_option_override): Rework how pass_insert_vzeroupper is
8628 added to the pass_manager to reflect autogenerated changes.
8629 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
8631 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
8633 PR rtl-optimization/57708
8634 * recog.c (peep2_find_free_register): Validate all regs in a
8637 2013-08-05 Jan Hubicka <jh@suse.cz>
8640 * cgraph.c (verify_cgraph_node): Accept local flags from other
8642 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
8643 (function_and_variable_visibility): Likewise.
8644 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
8646 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
8648 * graph.c (init_graph_slim_pretty_print): Remove.
8649 (print_graph_cfg): Do not call it. Use local pretty printer.
8650 (start_graph_dump): Likewise.
8652 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
8654 * gimple-pretty-print.c (buffer): Remove.
8655 (initialized): Likewise.
8656 (maybe_init_pretty_print): Likewise.
8657 (print_gimple_stmt): Do not call it. Use non-static local
8658 pretty_printer variable.
8659 (print_gimple_expr): Likewise.
8660 (print_gimple_seq): Likewise.
8661 (gimple_dump_bb): Likewise.
8663 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
8665 * asan.c (asan_pp): Remove.
8666 (asan_pp_initialized): Likewise.
8667 (asan_pp_initialize): Likewise.
8668 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
8669 (asan_emit_stack_protection): Tidy. Use local pretty printer.
8670 (asan_add_global): Likewise.
8672 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
8674 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
8675 * diagnostic.h (diagnostic_flush_buffer): Adjust.
8676 * pretty-print.c (pp_formatted_text_data): Likewise.
8677 (pp_indent): Rename from pp_base_indent.
8678 (pp_format): Rename from pp_base_format.
8679 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
8680 (pp_format_verbatim): Rename from pp_base_format_verbatim.
8681 (pp_flush): Rename from pp_base_flush.
8682 (pp_set_line_maximum_length): Rename from
8683 pp_base_set_line_maximum_length.
8684 (pp_clear_output_area): Rename from pp_base_clear_output_area.
8685 (pp_set_prefix): Rename from pp_base_set_prefix.
8686 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
8687 (pp_emit_prefix): Rename from pp_base_emit_prefix.
8688 (pp_append_text): Rename from pp_base_append_text.
8689 (pp_formatted_text): Rename from pp_base_formatted_text.
8690 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
8691 (pp_remaining_character_count_for_line): Rename from
8692 pp_base_remaining_character_count_for_line.
8693 (pp_newline): Rename from pp_base_newline.
8694 (pp_character): Rename from pp_base_character.
8695 (pp_string): Rename from pp_base_string.
8696 (pp_maybe_space): Rename from pp_base_maybe_space.
8697 * asan.c (asan_pp_string): Adjust.
8698 (asan_emit_stack_protection): Likewise.
8699 (asan_add_global): Likewise.
8700 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
8701 * tree-mudflap.c (mf_varname_tree): Likewise.
8702 * tree-pretty-print.c (pp_tree_identifier): Rename from
8703 pp_base_tree_identifier.
8704 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
8705 Declare as function.
8707 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
8709 * pretty-print.h (pp_bar_bar): New.
8710 (pp_ampersand_ampersand): Likewise.
8711 (pp_less_equal): Likewise.
8712 (pp_greater_equal): Likewise.
8713 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
8714 printer functions instead of pp_string or operators and punctuators.
8715 (dump_gimple_call): Likewise.
8716 (dump_gimple_omp_for): Likewise.
8717 (dump_gimple_transaction): Likewise.
8718 (dump_gimple_phi): Likewise.
8719 (pp_gimple_stmt_1): Likewise.
8720 * sched-vis.c (print_insn): Likewise.
8721 * tree-mudflap.c (mf_varname_tree): Likewise.
8722 * tree-pretty-print.c (dump_block_node): Likewise.
8723 (dump_generic_node): Likewise.
8725 2013-08-02 Jan Hubicka <jh@suse.cz>
8727 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
8729 * lto-streamer-out.c (tree_is_indexable): Results decls and
8730 parm decls are not indexable.
8731 (DFS_write_tree_body): Do not follow args and results.
8732 (hash_tree): Likewise.
8733 (output_functions): Rearrange so struct function is needed
8734 only when real body is output; be able to also ouptut abstract
8735 functions; output DECL_ARGUMENTS and DECL_RESULT.
8736 (lto_output): When not in WPA, ale store abstract functions.
8737 (write_symbol): Do not care about RESULT_DECL.
8738 (output_symbol_p): Handle correctly sbtract decls.
8739 * lto-streamer-in.c (input_function): Rearrange so struct
8740 function can be NULL at entry; allow streaming of
8741 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
8742 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
8743 sanity check during LTO.
8744 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
8745 RESULT_DECl and DECL_ARGUMENTS.
8746 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
8749 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
8751 * pretty-print.h (pp_underscore): New.
8753 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
8754 printer functions instead of pp_character.
8755 (dump_binary_rhs): Likewise.
8756 (dump_ternary_rhs): Likewise.
8757 (dump_gimple_call_args): Likewise.
8758 (pp_points_to_solution): Likewise.
8759 (dump_gimple_call): Likewise.
8760 (dump_gimple_switch): Likewise.
8761 (dump_gimple_cond): Likewise.
8762 (dump_gimple_bind): Likewise.
8763 (dump_gimple_try): Likewise.
8764 (dump_gimple_omp_for): Likewise.
8765 (dump_gimple_omp_continue): Likewise.
8766 (dump_gimple_omp_single): Likewise.
8767 (dump_gimple_omp_sections): Likewise.
8768 (dump_gimple_omp_block): Likewise.
8769 (dump_gimple_omp_critical): Likewise.
8770 (dump_gimple_transaction): Likewise.
8771 (dump_gimple_asm): Likewise.
8772 (dump_gimple_phi): Likewise.
8773 (dump_gimple_omp_parallel): Likewise.
8774 (dump_gimple_omp_task): Likewise.
8775 (dump_gimple_omp_atomic_load): Likewise.
8776 (dump_gimple_omp_atomic_store): Likewise.
8777 (dump_gimple_mem_ops): Likewise.
8778 (pp_gimple_stmt_1): Likewise.
8779 (pp_cfg_jump): Likewise.
8780 (dump_implicit_edges): Likewise.
8781 (gimple_dump_bb_for_graph): Likewise.
8782 * graph.c (draw_cfg_node): Likewise.
8783 * langhooks.c (lhd_print_error_function): Likewise.
8784 * sched-vis.c (print_exp): Likewise.
8785 (print_value): Likewise.
8786 (print_pattern): Likewise.
8787 (print_insn): Likewise.
8788 (rtl_dump_bb_for_graph): Likewise.
8789 * tree-pretty-print.c (dump_function_declaration): Likewise.
8790 (dump_array_domain): Likewise.
8791 (dump_omp_clause): Likewise.
8792 (dump_location): Likewise.
8793 (dump_generic_node): Likewise.
8794 (print_struct_decl): Likewise.
8795 * diagnostic.c (diagnostic_show_locus): Use pp_space.
8797 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8799 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
8800 candidate table when replacing a candidate statement.
8801 (replace_rhs_if_not_dup): Likewise.
8802 (replace_one_candidate): Likewise.
8804 2013-08-02 Jan Hubicka <jh@suse.cz>
8805 Martin Liska <marxin.liska@gmail.com>
8807 * cgraphunit.c (add_new_function): Fix logic when adding from
8809 (assemble_thunk): Rename to ...
8810 (expand_thunk); .. this one; export; get it working with
8811 general functions; make produced gimple valid.
8812 * cgraph.h (expand_thunk): Declare.
8814 2013-08-02 Jan Hubicka <jh@suse.cz>
8816 * ipa-cp.c (gather_context_independent_values): Use
8817 ipa_get_param_move_cost.
8818 (get_replacement_map): Remove PARAM; move parameter folding
8820 (create_specialized_node): Update.
8821 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
8822 assert that we have gimple body; update move_cost.
8823 (count_formal_params): Assert that we have gimple body.
8824 (ipa_dump_param): New function.
8825 (ipa_alloc_node_params): Break out from ...
8826 (ipa_initialize_node_params): ... here.
8827 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
8828 (ipa_write_node_info): Stream move costs.
8829 (ipa_read_node_info): Read move costs.
8830 (ipa_update_after_lto_read): Do not recompute node params.
8831 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
8832 (ipa_get_param): Check we are not in WPA.
8833 (ipa_get_param_move_cost): New.
8834 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
8835 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
8836 parm numbers to be present.
8838 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
8840 PR rtl-optimization/58048
8841 * lra-constraints.c (process_alt_operands): Don't check asm
8842 operand on register.
8844 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
8846 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
8847 the implied StoreLoad barrier for atomic operations if before.
8849 2013-08-02 Jan Hubicka <jh@suse.cz>
8850 Martin Liska <marxin.liska@gmail.com>
8852 * cgraph.c (cgraph_function_body_availability): Do not check
8854 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
8855 symtab_node_availability): Declare.
8856 * ipa.c (can_replace_by_local_alias): New.
8857 (function_and_variable_visibility): Use it.
8858 * symtab.c (symtab_for_node_and_aliases,
8859 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
8861 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
8863 PR rtl-optimization/57963
8864 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
8865 (lra_constraints): Use them.
8867 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
8869 * config/arm/types.md (define_attr "type"): Add "load_acq"
8871 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
8873 (cortex_a53_store1): Likewise.
8875 2013-08-01 Jan Hubicka <jh@suse.cz>
8877 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
8878 partitions are not needed.
8880 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
8882 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
8883 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
8884 MAYBE_NON_Q_CLASS_P where appropriate.
8886 2013-08-01 Jan Hubicka <jh@suse.cz>
8888 * cgraph.h (release_function_body): Declare.
8889 * tree.c (free_lang_data_in_decl): Free, parameters and return values
8890 of unused delcarations.
8892 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8894 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
8895 RTL form when subtracting a constant.
8897 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8899 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
8900 Generate canonical plus rtx with negated immediate instead of minus
8902 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
8904 2013-08-01 Jan Hubicka <jh@suse.cz>
8906 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
8907 (cgraph_release_function_body): Likewise.
8908 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
8909 * cgraph.h (cgrpah_node): Rename abstract_and_needed
8910 to used_as_abstract_origin.
8911 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
8912 symbols used as abstract origins.
8913 * cgraphunit.c (analyze_functions): Update.
8914 * ipa.c (symtab_remove_unreachable_nodes): Recompute
8915 used_as_abstract_origin.
8916 * tree-inline.c (tree_function_versioning): Update
8917 used_as_abstract_origin; be ready for DECL_RESULT and
8918 DECL_ARGUMENTS to be NULL.
8920 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
8921 for abstract functions.
8922 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
8925 2013-08-01 Jan Hubicka <jh@suse.cz>
8927 * profile.c (compute_value_histograms): Fix thinko.
8929 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
8931 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
8932 aarch-common-protos.h to extra_headers.
8933 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
8934 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
8935 * config/aarch64/t-aarch64 (aarch-common.o): Define.
8937 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
8939 * config/aarch64/aarch64.md (define_attr "type"): Delete.
8940 Include "../arm/types.md". Define "type" attribute for all patterns.
8941 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
8944 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
8946 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
8947 to support power8 load fusion.
8948 (fusion_gpr_mem_load): Likewise.
8950 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
8952 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
8953 declarations for power8 load fusion.
8954 (emit_fusion_gpr_load): Likewise.
8956 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
8957 tuning for power8, turn on fusion mode by default. Turn on sign
8958 extending fusion mode if normal fusion mode is on, and we are at
8960 (fusion_gpr_load_p): New function, return true if we can fuse an
8961 addis instruction with a dependent load to a GPR.
8962 (emit_fusion_gpr_load): Emit the instructions for power8 load
8965 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
8966 (VSX load fusion peepholes): New peepholes to fuse together an
8967 addi instruction with a VSX load instruction.
8969 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
8970 peepholes to fuse an addis instruction with a load to a GPR base
8971 register. If we are supporting sign extending fusions, convert
8972 sign extending loads to zero extending loads and add an explicit
8975 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
8977 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
8978 aarch-common-protos.h to extra_headers.
8979 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
8980 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
8981 (arm_early_store_addr_dep): Likewise.
8982 (arm_no_early_alu_shift_dep): Likewise.
8983 (arm_no_early_alu_shift_value_dep): Likewise.
8984 (arm_no_early_mul_dep): Likewise.
8985 (arm_no_early_store_addr_dep): Likewise.
8986 (arm_mac_accumulator_is_mul_result): Likewise.
8987 (arm_mac_accumulator_is_result): Likewise.
8988 * config/arm/aarch-common.c: ... here. New file.
8989 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
8991 (arm_early_store_addr_dep): Likewise.
8992 (arm_no_early_alu_shift_dep): Likewise.
8993 (arm_no_early_alu_shift_value_dep): Likewise.
8994 (arm_no_early_mul_dep): Likewise.
8995 (arm_no_early_store_addr_dep): Likewise.
8996 (arm_mac_accumulator_is_mul_result): Likewise.
8997 (arm_mac_accumulator_is_result): Likewise.
8998 * config/arm/aarch-common-protos.h: ... here. New file.
8999 * config/arm/t-arm (aarch-common.o): Define.
9001 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
9003 * config/arm/arm.md: Include new file "types.md".
9004 (define_attr "type"): Move from here to ...
9005 (define_attr "mul32"): Likewise.
9006 (define_attr "mul64"): Likewise.
9007 * config/arm/types.md: ... here. New file.
9009 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
9011 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
9012 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
9014 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9016 * gen-pass-instances.awk: Fix offset of substr().
9018 2013-07-31 David Malcolm <dmalcolm@redhat.com>
9020 * Makefile.in (pass-instances.def): New.
9021 (passes.o): Replace dependency on passes.def with one on
9024 * gen-pass-instances.awk: New.
9026 * passes.c (pass_manager::pass_manager): Use pass-instances.def
9027 rather than passes.def, updating local definition of NEXT_PASS
9028 macro to add an extra NUM parameter (currently unused).
9030 2013-07-30 David Malcolm <dmalcolm@redhat.com>
9032 * Makefile.in (PASS_MANAGER_H): New.
9033 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
9034 (passes.o): Likewise.
9035 (statistics.o): Likewise.
9036 (cgraphunit.o): Likewise.
9037 (context.o): Depend on PASS_MANAGER_H.
9039 * pass_manager.h: New.
9041 * cgraphunit.c (cgraph_add_new_function): Update for moves
9042 of globals to fields of pass_manager.
9043 (analyze_function): Likewise.
9044 (expand_function): Likewise.
9045 (ipa_passes): Likewise.
9046 (compile): Likewise.
9048 * context.c (context::context): New.
9049 * context.h (context::context): New.
9050 (context::get_passes): New.
9051 (context::passes_): New.
9053 * lto-cgraph.c (input_node): Update for moves of globals to
9054 fields of pass_manager.
9056 * passes.c (all_passes): Remove, in favor of a field of the
9057 same name within the new class pass_manager.
9058 (all_small_ipa_passes): Likewise.
9059 (all_lowering_passes): Likewise.
9060 (all_regular_ipa_passes): Likewise.
9061 (all_late_ipa_passes): Likewise.
9062 (all_lto_gen_passes): Likewise.
9063 (passes_by_id): Likewise.
9064 (passes_by_id_size): Likewise.
9065 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
9066 the new class pass_manager.
9067 (set_pass_for_id): Convert to...
9068 (pass_manager::set_pass_for_id): ...method.
9069 (get_pass_for_id): Convert to...
9070 (pass_manager::get_pass_for_id): ...method.
9071 (register_one_dump_file): Move body of implementation into...
9072 (pass_manager::register_one_dump_file): ...here.
9073 (register_dump_files_1): Convert to...
9074 (pass_manager::register_dump_files_1): ...method.
9075 (register_dump_files): Convert to...
9076 (pass_manager::register_dump_files): ...method.
9077 (create_pass_tab): Update for moves of globals to fields of
9079 (dump_passes): Move body of implementation into...
9080 (pass_manager::dump_passes): ...here.
9081 (register_pass): Move body of implementation into...
9082 (pass_manager::register_pass): ...here.
9083 (init_optimization_passes): Convert into...
9084 (pass_manager::pass_manager): ...constructor for new
9085 pass_manager class, and initialize the pass_lists array.
9086 (check_profile_consistency): Update for moves of globals to
9087 fields of pass_manager.
9088 (dump_profile_report): Move body of implementation into...
9089 (pass_manager::dump_profile_report): ...here.
9090 (ipa_write_summaries_1): Update for moves of pass lists from
9091 being globals to fields of pass_manager.
9092 (ipa_write_optimization_summaries): Likewise.
9093 (ipa_read_summaries): Likewise.
9094 (ipa_read_optimization_summaries): Likewise.
9095 (execute_all_ipa_stmt_fixups): Likewise.
9097 * statistics.c (statistics_fini): Update for moves of globals to
9098 fields of pass_manager.
9100 * toplev.c (general_init): Replace call to
9101 init_optimization_passes with construction of the pass_manager
9104 * tree-pass.h (all_passes): Remove, in favor of a field of the
9105 same name within the new class pass_manager.
9106 (all_small_ipa_passes): Likewise.
9107 (all_lowering_passes): Likewise.
9108 (all_regular_ipa_passes): Likewise.
9109 (all_lto_gen_passes): Likewise.
9110 (all_late_ipa_passes): Likewise.
9111 (passes_by_id): Likewise.
9112 (passes_by_id_size): Likewise.
9113 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
9114 the new class pass_manager.
9115 (get_pass_for_id): Remove.
9117 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
9119 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
9122 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
9124 * arm.md (mulhi3): New expand pattern.
9126 2013-07-30 Jan Hubicka <jh@suse.cz>
9127 Martin Liska <marxin.liska@gmail.com>
9129 * profile.c (compute_value_histograms): Do not ICE when
9130 there is mismatch only on some counters.
9132 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9134 PR rtl-optimization/57637
9135 * function.c (move_insn_for_shrink_wrap): Also check the
9136 GEN set of the LIVE problem for the liveness analysis
9137 if it exists, otherwise give up.
9139 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9141 PR tree-optimization/57993
9142 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
9143 replaced statement in the candidate table.
9144 (phi_add_costs): Return infinite cost when the hidden basis does
9145 not dominate all phis on which the candidate is dependent.
9146 (replace_one_candidate): Record replaced statement in the
9149 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
9151 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
9152 (ashlv2si3): New expander.
9153 (*ashlv2si3_i): New define_insn_and_split.
9154 * predicates.md (float_operation): Allow patterns with three
9157 PR rtl-optimization/58021
9158 * mode-switching.c (create_pre_exit): Always split off preceding
9159 insns if we are not at the basic block head.
9161 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
9163 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
9164 (UCLIBC_DYNAMIC_LINKER): New macro.
9165 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
9167 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
9168 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
9170 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
9171 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
9172 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
9173 for SF and DF modes. Use ieee_quad_format for TF mode.
9174 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
9175 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
9176 (mips_option_override): Handle `-mnan=legacy'.
9177 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
9178 `-mabs=2008' and `-mnan=2008'.
9179 (OPTION_DEFAULT_SPECS): Add "nan" default.
9180 (ASM_SPEC): Handle `-mnan='.
9181 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
9182 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
9183 comment accordingly.
9184 (neg<mode>2): Likewise.
9185 * config/mips/mips.opt (mabs, mnan): New options.
9186 * doc/install.texi (Configuration): Document `--with-nan=' option.
9187 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
9189 (MIPS Options): Document them.
9190 * config.gcc <mips*-*-*>: Handle `--with-nan='.
9191 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
9192 * configure: Regenerate.
9193 * config.in: Regenerate.
9195 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
9197 * config/i386/i386.md (float post-reload splitters): Do not check
9198 for subregs of SSE registers.
9200 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
9201 H.J. Lu <hongjiu.lu@intel.com>
9205 * config/i386/i386.md (post-reload splitter
9206 to avoid partial SSE reg dependency stalls): New pattern.
9208 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
9210 * config/s390/s390.md ("movcc"): Swap load and store instructions.
9212 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
9214 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
9215 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
9217 2013-07-26 Cary Coutant <ccoutant@google.com>
9219 * dwarf2out.c (die_checksum_ordered): Don't include template
9220 instantiations in signature.
9221 (is_template_parameter): New function.
9222 (is_template_instantiation): New function.
9223 (generate_skeleton_bottom_up): Don't include template instantiations
9225 (generate_skeleton): Likewise.
9226 (break_out_comdat_types): Move recursive call to break out nested
9228 (prune_unused_types_mark_generic_parms_dies): Call
9229 is_template_parameter.
9231 2013-07-26 Ian Bolton <ian.bolton@arm.com>
9233 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
9234 uses vector registers.
9235 * config/aarch64/iterators.md: Add attributes rtn and vas.
9237 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9238 Richard Earnshaw <richard.earnshaw@arm.com>
9240 * combine.c (simplify_comparison): Re-canonicalize operands
9242 * config/arm/arm.md (movcond_addsi): New splitter.
9244 2013-07-25 Sterling Augustine <saugustine@google.com>
9246 * dwarf2out.c (size_of_pubnames): Move code to...
9247 (include_pubname_in_output): ...here. New.
9248 (want_pubnames): Rearrange.
9249 (output_pubnames): Call include_pubname_in_output. Move assertion.
9251 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
9253 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
9255 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
9258 * doc/extend.texi: Remove obsolete builtins. Fix
9259 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
9261 2013-07-25 Jan Hubicka <jh@suse.cz>
9263 * cgraph.c (release_function_body): Break out from ...
9264 (cgraph_release_function_body): ... this one; also release DECL_RESULT
9266 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
9267 old_tree in the map.
9268 (create_specialized_node): Update.
9269 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
9271 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
9272 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
9273 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
9274 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
9277 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9279 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
9280 addsi3_carryin_alt2_<optab>): Correct output template.
9282 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9284 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
9285 Adjust for arm_restrict_it.
9286 Remove trailing whitespace.
9288 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
9290 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
9291 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
9293 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
9295 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
9297 PR rtl-optimization/57960
9298 * lra-constraints.c (process_alt_operands): Use the right mode
9299 when checking strict_low.
9301 2013-07-25 Jan Hubicka <jh@suse.cz>
9303 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
9304 * cgraph.c (cgraph_remove_node): Do not release function body
9305 when in cgraph streaming.
9306 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
9307 in other partitions are not considered reachable; fix handling of
9310 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
9312 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
9314 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
9319 * config/arm/arm.md ("*sibcall_insn): Replace use of
9320 Ss with US. Adjust output for v5 and v4t.
9321 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
9323 * config/arm/constraints.md ("Ss"): Rename to US.
9325 2013-07-25 Terry Guo <terry.guo@arm.com>
9327 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
9328 shift_add/shift_sub0/shift_sub1 RTXs.
9330 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
9331 Anton Blanchard <anton@au1.ibm.com>
9333 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
9334 (altivec_vpks<VI_char>ss): Likewise.
9335 (altivec_vpks<VI_char>us): Likewise.
9336 (altivec_vpku<VI_char>us): Likewise.
9337 (altivec_vpku<VI_char>um): Likewise.
9339 2013-07-24 David Malcolm <dmalcolm@redhat.com>
9341 Introduce context class.
9343 * Makefile.in (CONTEXT_H): New.
9344 (OBJS): Add context.o.
9345 (toplev.o): Add CONTEXT_H to dependencies.
9348 * toplev.c (general_init): Create the singleton gcc::context instance.
9354 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
9356 PR rtl-optimization/57968
9357 * mode-switching.c (create_pre_exit): Allow instructions that
9358 don't set a return register to need a non-exit mode.
9360 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9361 Anton Blanchard <anton@au1.ibm.com>
9363 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
9364 operands to vperm for little endian.
9365 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
9366 of lvsl to create the control mask for a vperm for little endian.
9368 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9369 Anton Blanchard <anton@au1.ibm.com>
9371 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
9372 two operands for little-endian.
9374 2013-07-23 Steve Ellcey <sellcey@mips.com>
9376 * config/mips/mips.c (mips_case_values_threshold): New.
9377 (TARGET_CASE_VALUES_THRESHOLD): Define.
9379 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9380 Anton Blanchard <anton@au1.ibm.com>
9382 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
9383 selection of field for vector splat in little endian mode.
9385 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
9387 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
9388 expanders to rs6000.md.
9389 (ior<mode>3): Likewise.
9390 (and<mode>3): Likewise.
9391 (one_cmpl<mode>2): Likewise.
9392 (nor<mode>3): Likewise.
9393 (andc<mode>3): Likewise.
9394 (eqv<mode>3): Likewise.
9395 (nand<mode>3): Likewise.
9396 (orc<mode>3): Likewise.
9398 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
9401 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
9402 to split multi-word logical operations.
9403 (rs6000_split_logical_di): Likewise.
9404 (rs6000_split_logical): Likewise.
9406 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
9407 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
9408 and allow TImode operations in 32-bit.
9409 (vsx_and<mode>3_64bit): Likewise.
9410 (vsx_ior<mode>3_32bit): Likewise.
9411 (vsx_ior<mode>3_64bit): Likewise.
9412 (vsx_xor<mode>3_32bit): Likewise.
9413 (vsx_xor<mode>3_64bit): Likewise.
9414 (vsx_one_cmpl<mode>2_32bit): Likewise.
9415 (vsx_one_cmpl<mode>2_64bit): Likewise.
9416 (vsx_nor<mode>3_32bit): Likewise.
9417 (vsx_nor<mode>3_64bit): Likewise.
9418 (vsx_andc<mode>3_32bit): Likewise.
9419 (vsx_andc<mode>3_64bit): Likewise.
9420 (vsx_eqv<mode>3_32bit): Likewise.
9421 (vsx_eqv<mode>3_64bit): Likewise.
9422 (vsx_nand<mode>3_32bit): Likewise.
9423 (vsx_nand<mode>3_64bit): Likewise.
9424 (vsx_orc<mode>3_32bit): Likewise.
9425 (vsx_orc<mode>3_64bit): Likewise.
9427 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
9428 logical types in GPRs.
9430 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
9431 logical insns to rs6000.md, and allow TImode operations in
9433 (altivec_ior<mode>3): Likewise.
9434 (altivec_xor<mode>3): Likewise.
9435 (altivec_one_cmpl<mode>2): Likewise.
9436 (altivec_nor<mode>3): Likewise.
9437 (altivec_andc<mode>3): Likewise.
9439 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
9440 attributes for moving the 128-bit logical operations into
9442 (BOOL_REGS_OUTPUT): Likewise.
9443 (BOOL_REGS_OP1): Likewise.
9444 (BOOL_REGS_OP2): Likewise.
9445 (BOOL_REGS_UNARY): Likewise.
9446 (BOOL_REGS_AND_CR0): Likewise.
9447 (one_cmpl<mode>2): Add support for DI logical operations on
9448 32-bit, splitting the operations to 32-bit.
9452 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
9453 changes to combine the 32/64-bit code, allow logical operations on
9454 TI mode in 32-bit, and to use similar match_operator patterns like
9455 scalar mode uses. Combine the Altivec and VSX code for logical
9456 operations, and move it here.
9457 (ior<mode>3, 128-bit types): Likewise.
9458 (xor<mode>3, 128-bit types): Likewise.
9459 (one_cmpl<mode>3, 128-bit types): Likewise.
9460 (nor<mode>3, 128-bit types): Likewise.
9461 (andc<mode>3, 128-bit types): Likewise.
9462 (eqv<mode>3, 128-bit types): Likewise.
9463 (nand<mode>3, 128-bit types): Likewise.
9464 (orc<mode>3, 128-bit types): Likewise.
9465 (and<mode>3_internal): Likewise.
9466 (bool<mode>3_internal): Likewise.
9467 (boolc<mode>3_internal1): Likewise.
9468 (boolc<mode>3_internal2): Likewise.
9469 (boolcc<mode>3_internal1): Likewise.
9470 (boolcc<mode>3_internal2): Likewise.
9471 (eqv<mode>3_internal1): Likewise.
9472 (eqv<mode>3_internal2): Likewise.
9473 (one_cmpl1<mode>3_internal): Likewise.
9475 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
9477 * config/microblaze/microblaze.c (microblaze_expand_prologue):
9478 Rename flag_stack_usage to flag_stack_usage_info.
9480 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
9482 * config/microblaze/sync.md: New file.
9483 * config/microblaze/microblaze.md: Include sync.md
9484 * config/microblaze/microblaze.c: Add print_operand 'y'.
9485 * config/microblaze/constraints.md: Add memory_contraint
9486 'Q' which is a single register.
9488 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
9490 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
9492 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
9494 * reload.c (find_reloads): Exit loop once we find this operand
9495 cannot be reloaded somehow for this alternative.
9497 * reload.c (find_reloads): Exit loop once we find a hard register.
9499 * rtlanal.c (computed_jump_p): Exit loop once we find label
9502 * i386.c (ix86_pad_returns): Exit loop after setting replace.
9504 * cfgloopmanip.c (remove_path): Exit loop after setting
9507 * gensupport.c (subst_dup): Avoid loop if code is not
9508 MATCH_DUP nor MATCH_OP_DUP.
9510 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
9512 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
9514 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
9516 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
9517 true for SP_REGNUM if mode == ptr_mode.
9518 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
9519 with value R0_REGNUM + 31.
9521 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
9523 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
9524 pad pointer-typed argument downward.
9526 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
9528 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
9529 and __ILP32__ when the ILP32 model is in use.
9531 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
9533 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
9534 (aarch64_load_symref_appropriately): In the case of
9535 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
9536 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
9537 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
9538 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
9539 if 'mode' doesn't equal to 'ptr_mode'.
9540 (aarch64_output_mi_thunk): Add an assertion on the alignment of
9541 'vcall_offset'; change to call aarch64_emit_move differently depending
9542 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
9543 to calculate the upper bound of 'vcall_offset'.
9544 (aarch64_cannot_force_const_mem): Change to also return true if
9546 (aarch64_legitimize_reload_address): In the case of large
9547 displacements, add new local variable 'xmode' and an assertion
9548 based on it; change to use 'xmode' to generate the new rtx and
9550 (aarch64_asm_trampoline_template): Change to generate the template
9551 differently depending on TARGET_ILP32 or not; change to use
9552 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
9553 (aarch64_trampoline_size): Removed.
9554 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
9555 and replace immediate literals with it. Change to use 'ptr_mode'
9556 instead of 'DImode' and call convert_memory_address if the mode
9557 of 'fnaddr' doesn't equal to 'ptr_mode'.
9558 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
9560 (aarch64_elf_asm_destructor): Likewise.
9561 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
9562 on TARGET_ILP32 instead of aarch64_trampoline_size.
9563 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
9564 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
9565 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
9566 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
9567 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
9568 (storewb_pair<GPI:mode>_<P:mode>): ... this.
9569 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
9570 depending on the value of 'mode'.
9571 (add_losym_<mode>): New.
9572 (ldr_got_small_<mode>): New, based on ldr_got_small.
9573 (ldr_got_small): Remove.
9574 (ldr_got_small_sidi): New.
9575 * config/aarch64/iterators.md (P): New.
9576 (PTR): Change to 'ptr_mode' in the condition.
9578 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
9580 * config.gcc (aarch64*-*-*): Support --with-abi.
9581 (aarch64*-*-elf): Support --with-multilib-list.
9582 (aarch64*-*-linux*): Likewise.
9583 (supported_defaults): Add abi to aarch64*-*-*.
9584 * configure.ac: Mention AArch64 for --with-multilib-list.
9585 * configure: Re-generated.
9586 * config/aarch64/biarchilp32.h: New file.
9587 * config/aarch64/biarchlp64.h: New file.
9588 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
9590 (MULTILIB_DEFAULTS): Ditto.
9591 (DRIVER_SELF_SPECS): Ditto.
9592 (ASM_SPEC): Update to also substitute -mabi.
9593 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
9594 file whose name depends on -mabi= and -mbig-endian.
9595 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
9597 (POINTER_SIZE): New define.
9598 (POINTERS_EXTEND_UNSIGNED): Ditto.
9599 (enum aarch64_abi_type): New enumeration tag.
9600 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
9601 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
9602 (TARGET_ILP32): New define.
9603 * config/aarch64/aarch64.opt (mabi): New.
9605 (ilp32, lp64): New values for -mabi.
9606 * config/aarch64/t-aarch64 (comma): New define.
9607 (MULTILIB_OPTIONS): Ditto.
9608 (MULTILIB_DIRNAMES): Ditto.
9609 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
9610 * doc/invoke.texi: Document -mabi for AArch64.
9612 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
9614 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
9616 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9617 Anton Blanchard <anton@au1.ibm.com>
9619 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
9620 endianness when selecting field to splat.
9622 2013-07-22 Eric Christopher <echristo@gmail.com>
9624 * dwarf2out.c (die_odr_checksum): New function to use
9625 CHECKSUM_ macros and ULEB128 for DIE tag.
9626 (generate_type_signature): Use.
9628 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
9630 * config.gcc (sparc*-*-*): Accept leon3 processor.
9631 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
9632 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
9633 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
9634 * config/sparc/sparc.opt (enum processor_type): Add leon3.
9635 (mfix-ut699): Adjust comment.
9636 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
9637 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
9638 (CPP_CPU_SPEC): Likewise.
9639 (ASM_CPU_SPEC): Likewise.
9640 * config/sparc/sparc.c (leon3_cost): New constant.
9641 (sparc_option_override): Add leon3 support.
9642 (mem_ref): New function.
9643 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
9644 (sparc_do_work_around_errata): Look into the instruction in the delay
9645 slot and adjust accordingly. Add fix for the data cache nullify issues
9646 of the UT699. Change insertion position for the NOP.
9647 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
9648 (leon3_load): New reservation.
9649 (leon_store): Bump latency to 2.
9650 (grfpu): New automaton.
9651 (grfpu_alu): New unit.
9652 (grfpu_ds): Likewise.
9653 (leon_fp_alu): Adjust.
9654 (leon_fp_mult): Delete.
9655 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
9656 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
9657 * config/sparc/sparc.md (cpu): Add leon3.
9658 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
9660 (atomic_test_and_set): Likewise.
9663 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
9665 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
9666 default for R5900 targets.
9667 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
9668 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
9669 * config/mips/mips.c (mips_option_override): Report an error for
9670 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
9671 for -march=r5900 -mhard-float.
9673 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
9675 * df-problems.c (can_move_insns_across): Exit loop once we
9676 find a non-fixed, non-global register.
9678 * ipa-pure-const.c (propagate_nothrow): Exit loop after
9681 * omega.c (omega_eliminate_red): Break after setting red_found.
9682 (omega_problem_has_red_equations): Similarly after setting found.
9683 (omega_query_variable): Similarly after setting coupled.
9685 2013-07-22 Marek Polacek <polacek@redhat.com>
9687 * gimplify.c: Don't include gimple.h twice.
9689 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9691 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
9692 instead of TARGET_THUMB1.
9693 (Pz): New constraint.
9694 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
9696 (compare_negsi_si): Likewise.
9697 (compare_addsi2_op0): Likewise.
9698 (compare_addsi2_op1): Likewise.
9699 (addsi3_carryin_<optab>): Likewise.
9700 (addsi3_carryin_alt2_<optab>): Likewise.
9701 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
9702 for arm_restrict_it.
9703 (subsi3_carryin): Likewise.
9704 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
9705 (minmax_arithsi): Disable for arm_restrict_it.
9706 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
9707 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
9708 (satsi_<SAT:code>_shift): Likewise.
9709 (arm_shiftsi3): Add alternative for 16-bit encoding.
9710 (arm32_movhf): Disable for arm_restrict_it.
9711 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
9712 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
9714 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
9716 * config/arm/arm.md (attribute "insn"): Delete.
9717 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
9718 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
9719 (not_shiftsi): Update for attribute change.
9720 (not_shiftsi_compare0): Likewise.
9721 (not_shiftsi_compare0_scratch): Likewise.
9722 (arm_one_cmplsi2): Likewise.
9723 (thumb1_one_cmplsi2): Likewise.
9724 (notsi_compare0): Likewise.
9725 (notsi_compare0_scratch): Likewise.
9726 (thumb1_movdi_insn): Likewise.
9727 (arm_movsi_insn): Likewise.
9728 (movhi_insn_arch4): Likewise.
9729 (movhi_bytes): Likewise.
9730 (arm_movqi_insn): Likewise.
9731 (thumb1_movqi_insn): Likewise.
9732 (arm32_movhf): Likewise.
9733 (thumb1_movhf): Likewise.
9734 (arm_movsf_soft_insn): Likewise.
9735 (thumb1_movsf_insn): Likewise.
9736 (thumb_movdf_insn): Likewise.
9737 (movsicc_insn): Likewise.
9738 (movsfcc_soft_insn): Likewise.
9739 (and_scc): Likewise.
9740 (cond_move): Likewise.
9741 (if_move_not): Likewise.
9742 (if_not_move): Likewise.
9743 (if_shift_move): Likewise.
9744 (if_move_shift): Likewise.
9745 (if_shift_shift): Likewise.
9746 (if_not_arith): Likewise.
9747 (if_arith_not): Likewise.
9748 (cond_move_not): Likewise.
9749 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
9750 (neon_mov<mode>): Likewise.
9751 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
9752 (thumb2_movsi_vfp): Likewise.
9753 (movsf_vfp): Likewise.
9754 (thumb2_movsf_vfp): Likewise.
9755 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
9757 (cortexa7_older_only): Likewise.
9758 (cortexa7_younger): Likewise.
9759 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
9760 (1020alu_shift_op): Likewise.
9761 (1020alu_shift_reg_op): Likewise.
9762 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
9763 (alu_shift_op): Likewise.
9764 (alu_shift_reg_op): Likewise.
9765 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
9766 (11_alu_shift_op): Likewise.
9767 (11_alu_shift_reg_op): Likewise.
9768 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
9769 (9_alu_shift_reg_op): Likewise.
9770 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
9772 (cortex_a15_alu_shift): Likewise.
9773 (cortex_a15_alu_shift_reg): Likewise.
9774 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
9775 (cortex_a5_alu_shift): Likewise.
9776 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
9778 (cortex_a53_alu_shift): Likewise.
9779 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
9781 (cortex_a7_alu_reg): Likewise.
9782 (cortex_a7_alu_shift): Likewise.
9783 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
9784 (cortex_a8_alu_shift): Likewise.
9785 (cortex_a8_alu_shift_reg): Likewise.
9786 (cortex_a8_mov): Likewise.
9787 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
9788 (cortex_a9_dp_shift): Likewise.
9789 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
9790 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
9791 (cortex_r4_mov): Likewise.
9792 (cortex_r4_alu_shift): Likewise.
9793 (cortex_r4_alu_shift_reg): Likewise.
9794 * config/arm/fa526.md (526_alu_op): Update for attribute change.
9795 (526_alu_shift_op): Likewise.
9796 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
9797 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
9798 (626te_alu_shift_op): Likewise.
9799 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
9800 (726te_alu_op): Likewise.
9801 (726te_alu_shift_op): Likewise.
9802 (726te_alu_shift_reg_op): Likewise.
9803 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
9804 (mp626_alu_shift_op): Likewise.
9805 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
9806 (pj4_alu_e1_conds): Likewise.
9807 (pj4_alu): Likewise.
9808 (pj4_alu_conds): Likewise.
9809 (pj4_shift): Likewise.
9810 (pj4_shift_conds): Likewise.
9811 (pj4_alu_shift): Likewise.
9812 (pj4_alu_shift_conds): Likewise.
9814 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9816 * config/arm/predicates.md (shiftable_operator_strict_it):
9818 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
9819 Disable cond_exec version for arm_restrict_it.
9820 (thumb2_smaxsi3): Convert to generate cond_exec.
9821 (thumb2_sminsi3): Likewise.
9822 (thumb32_umaxsi3): Likewise.
9823 (thumb2_uminsi3): Likewise.
9824 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
9825 (thumb2_neg_abssi2): Likewise.
9826 (thumb2_mov_scc): Add alternative for 16-bit encoding.
9827 (thumb2_movsicc_insn): Adjust alternatives.
9828 (thumb2_mov_negscc): Disable for arm_restrict_it.
9829 (thumb2_mov_negscc_strict_it): New pattern.
9830 (thumb2_mov_notscc_strict_it): New pattern.
9831 (thumb2_mov_notscc): Disable for arm_restrict_it.
9832 (thumb2_ior_scc): Likewise.
9833 (thumb2_ior_scc_strict_it): New pattern.
9834 (thumb2_cond_move): Adjust for arm_restrict_it.
9835 (thumb2_cond_arith): Disable for arm_restrict_it.
9836 (thumb2_cond_arith_strict_it): New pattern.
9837 (thumb2_cond_sub): Adjust for arm_restrict_it.
9838 (thumb2_movcond): Likewise.
9839 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
9840 (thumb2_zero_extendhisi2_v6): Likewise.
9841 (thumb2_zero_extendqisi2_v6): Likewise.
9842 (orsi_notsi_si): Likewise.
9843 (orsi_not_shiftsi_si): Likewise.
9845 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
9847 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
9848 instruction sequence is 1 byte shorter.
9850 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
9852 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
9853 it is not needed after split.
9855 2013-07-20 Iain Sandoe <iain@codesourcery.com>
9858 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
9859 second label for nonlocal goto receivers. Don't output pic base labels
9860 unless we're producing PIC; mark that action unreachable().
9861 (ix86_save_reg): If the function contains a nonlocal label, save the
9863 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
9864 * config/darwin.c (emitted_pic_label_num): New GTY.
9865 (update_pic_label_number_if_needed): New.
9866 (machopic_output_function_base_name): Adjust for nonlocal receiver
9868 (machopic_should_output_picbase_label): New.
9869 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
9870 (nonlocal_goto_receiver): New insn and split.
9872 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
9874 * config/aarch64/aarch64-builtins.c
9875 (aarch64_fold_builtin): Fold abs in all modes.
9876 * config/aarch64/aarch64-simd-builtins.def
9877 (abs): Enable for all modes.
9878 * config/aarch64/arm_neon.h
9879 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
9880 (vabs_f64): Add missing intrinsic.
9882 2013-07-19 Ian Bolton <ian.bolton@arm.com>
9884 * config/aarch64/arm_neon.h (vabs_s64): New function
9886 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
9889 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
9890 * config/avr/avr.md (adjust_len): Add `round'.
9891 * config/avr/avr-protos.h (avr_out_round): New prototype.
9892 (avr_out_plus): Add `out_label' argument.
9893 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
9894 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
9895 Handle the case where `insn' is just a pattern.
9896 (avr_out_bitop): Handle the case where `insn' is just a pattern.
9897 (avr_out_round): New function.
9898 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
9900 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
9902 * config/microblaze/microblaze.c (microblaze_expand_prologue):
9903 Add check for flag_stack_usage to handle -fstack-usage support
9905 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
9907 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
9908 interaction for new Power8 flags and VSX.
9910 2013-07-18 Sriraman Tallam <tmsriram@google.com>
9913 * tree-inline.c (expand_call_inline): Emit errors during
9914 early_inlining only if optimization is not turned on.
9916 2013-07-18 David Malcolm <dmalcolm@redhat.com>
9920 * passes.c (init_optimization_passes): Move the construction of
9921 the pass hierarchy into a new passes.def file.
9923 * Makefile.in (passes.o): Add dependency on passes.def.
9925 2013-07-18 David Malcolm <dmalcolm@redhat.com>
9927 * passes.c (init_optimization_passes): Introduce macros for
9928 constructing the tree of passes (INSERT_PASSES_AFTER,
9929 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
9930 TERMINATE_PASS_LIST).
9932 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
9933 Wei Mi <wmi@google.com>
9935 PR rtl-optimization/57878
9936 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
9938 (reload_pseudo_compare_func): Check nregs first for reload
9941 2013-07-18 David Malcolm <dmalcolm@redhat.com>
9943 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
9945 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
9947 * read-rtl.c (validate_const_int): Once an invalid character is
9948 seen, quit the loop.
9950 * gengtype.c (write_roots): Similarly once we find the "deletable"
9951 or "if_marked" option.
9953 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
9955 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
9956 "xtab" and "sat". Move value "clz" from here to ...
9957 (attriubte "type"): ... here.
9958 (satsi_<SAT:code>): Delete "insn" attribute.
9959 (satsi_<SAT:code>_shift): Likewise.
9960 (arm_zero_extendqisi2addsi): Likewise.
9961 (arm_extendqisi2addsi): Likewise.
9962 (clzsi2): Update for attribute changes.
9963 (rbitsi2): Likewise.
9964 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
9966 (arm_usatsihi): Likewise.
9967 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
9969 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
9971 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
9972 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
9973 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
9974 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
9975 in alphabetical order.
9976 (attribute "core_cycles"): Update for attribute changes.
9977 (arm_addsi3): Likewise.
9978 (addsi3_compare0): Likewise.
9979 (addsi3_compare0_scratch): Likewise.
9980 (addsi3_compare_op1): Likewise.
9981 (addsi3_compare_op2): Likewise.
9982 (compare_addsi2_op0): Likewise.
9983 (compare_addsi2_op1): Likewise.
9984 (addsi3_carryin_shift_<optab>): Likewise.
9985 (subsi3_carryin_shift): Likewise.
9986 (rsbsi3_carryin_shift): Likewise.
9987 (arm_subsi3_insn): Likewise.
9988 (subsi3_compare0): Likewise.
9989 (subsi3_compare): Likewise.
9990 (arm_andsi3_insn): Likewise.
9991 (thumb1_andsi3_insn): Likewise.
9992 (andsi3_compare0): Likewise.
9993 (andsi3_compare0_scratch): Likewise.
9994 (zeroextractsi_compare0_scratch
9995 (andsi_not_shiftsi_si): Likewise.
9996 (iorsi3_insn): Likewise.
9997 (iorsi3_compare0): Likewise.
9998 (iorsi3_compare0_scratch): Likewise.
9999 (arm_xorsi3): Likewise.
10000 (thumb1_xorsi3_insn): Likewise.
10001 (xorsi3_compare0): Likewise.
10002 (xorsi3_compare0_scratch): Likewise.
10003 (satsi_<SAT:code>_shift): Likewise.
10005 (arm_shiftsi3): Likewise.
10006 (shiftsi3_compare0): Likewise.
10007 (not_shiftsi): Likewise.
10008 (not_shiftsi_compare0): Likewise.
10009 (not_shiftsi_compare0_scratch): Likewise.
10010 (arm_one_cmplsi2): Likewise.
10011 (thumb_one_complsi2): Likewise.
10012 (notsi_compare0): Likewise.
10013 (notsi_compare0_scratch): Likewise.
10014 (thumb1_zero_extendhisi2): Likewise.
10015 (arm_zero_extendhisi2): Likewise.
10016 (arm_zero_extendhisi2_v6): Likewise.
10017 (arm_zero_extendhisi2addsi): Likewise.
10018 (thumb1_zero_extendqisi2): Likewise.
10019 (thumb1_zero_extendqisi2_v6): Likewise.
10020 (arm_zero_extendqisi2): Likewise.
10021 (arm_zero_extendqisi2_v6): Likewise.
10022 (arm_zero_extendqisi2addsi): Likewise.
10023 (thumb1_extendhisi2): Likewise.
10024 (arm_extendhisi2): Likewise.
10025 (arm_extendhisi2_v6): Likewise.
10026 (arm_extendqisi): Likewise.
10027 (arm_extendqisi_v6): Likewise.
10028 (arm_extendqisi2addsi): Likewise.
10029 (thumb1_extendqisi2): Likewise.
10030 (thumb1_movdi_insn): Likewise.
10031 (arm_movsi_insn): Likewise.
10032 (movsi_compare0): Likewise.
10033 (movhi_insn_arch4): Likewise.
10034 (movhi_bytes): Likewise.
10035 (arm_movqi_insn): Likewise.
10036 (thumb1_movqi_insn): Likewise.
10037 (arm32_movhf): Likewise.
10038 (thumb1_movhf): Likewise.
10039 (arm_movsf_soft_insn): Likewise.
10040 (thumb1_movsf_insn): Likewise.
10041 (movdf_soft_insn): Likewise.
10042 (thumb_movdf_insn): Likewise.
10043 (arm_cmpsi_insn): Likewise.
10044 (cmpsi_shiftsi): Likewise.
10045 (cmpsi_shiftsi_swp): Likewise.
10046 (arm_cmpsi_negshiftsi_si): Likewise.
10047 (movsicc_insn): Likewise.
10048 (movsfcc_soft_insn): Likewise.
10049 (arith_shiftsi): Likewise.
10050 (arith_shiftsi_compare0
10051 (arith_shiftsi_compare0_scratch
10052 (sub_shiftsi): Likewise.
10053 (sub_shiftsi_compare0
10054 (sub_shiftsi_compare0_scratch
10055 (and_scc): Likewise.
10056 (cond_move): Likewise.
10057 (if_plus_move): Likewise.
10058 (if_move_plus): Likewise.
10059 (if_move_not): Likewise.
10060 (if_not_move): Likewise.
10061 (if_shift_move): Likewise.
10062 (if_move_shift): Likewise.
10063 (if_shift_shift): Likewise.
10064 (if_not_arith): Likewise.
10065 (if_arith_not): Likewise.
10066 (cond_move_not): Likewise.
10067 (thumb1_ashlsi3): Set type attribute.
10068 (thumb1_ashrsi3): Likewise.
10069 (thumb1_lshrsi3): Likewise.
10070 (thumb1_rotrsi3): Likewise.
10071 (shiftsi3_compare0_scratch): Likewise.
10072 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
10073 (neon_mov<mode>): Likewise.
10074 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
10076 (thumb2_movsi_insn): Likewise.
10077 (thumb2_cmpsi_neg_shiftsi): Likewise.
10078 (thumb2_extendqisi_v6): Likewise.
10079 (thumb2_zero_extendhisi2_v6): Likewise.
10080 (thumb2_zero_extendqisi2_v6): Likewise.
10081 (thumb2_shiftsi3_short): Likewise.
10082 (thumb2_addsi3_compare0_scratch): Likewise.
10083 (orsi_not_shiftsi_si): Likewise.
10084 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
10085 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
10087 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
10088 (1020alu_shift_op): Likewise.
10089 (1020alu_shift_reg_op): Likewise.
10090 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
10091 (alu_shift_op): Likewise.
10092 (alu_shift_reg_op): Likewise.
10093 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
10094 (11_alu_shift_op): Likewise.
10095 (11_alu_shift_reg_op): Likewise.
10096 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
10097 (9_alu_shift_reg_op): Likewise.
10098 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
10100 (cortex_a15_alu_shift): Likewise.
10101 (cortex_a15_alu_shift_reg): Likewise.
10102 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
10104 (cortex_a5_alu_shift): Likewise.
10105 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
10107 (cortex_a53_alu_shift): Likewise.
10108 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
10110 (cortex_a7_alu_reg): Likewise.
10111 (cortex_a7_alu_shift): Likewise.
10112 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
10114 (cortex_a8_alu_shift): Likewise.
10115 (cortex_a8_alu_shift_reg): Likewise.
10116 (cortex_a8_mov): Likewise.
10117 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
10118 (cortex_a9_dp_shift): Likewise.
10119 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
10121 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
10123 (cortex_r4_mov): Likewise.
10124 (cortex_r4_alu_shift): Likewise.
10125 (cortex_r4_alu_shift_reg): Likewise.
10126 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
10127 (526_alu_shift_op): Likewise.
10128 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
10129 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
10130 (626te_alu_shift_op): Likewise.
10131 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
10132 (726te_alu_op): Likewise.
10133 (726te_alu_shift_op): Likewise.
10134 (726te_alu_shift_reg_op): Likewise.
10135 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
10136 (mp626_alu_shift_op): Likewise.
10137 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
10138 (pj4_alu_e1_conds): Likewise.
10139 (pj4_alu): Likewise.
10140 (pj4_alu_conds): Likewise.
10141 (pj4_shift): Likewise.
10142 (pj4_shift_conds): Likewise.
10143 (pj4_alu_shift): Likewise.
10144 (pj4_alu_shift_conds): Likewise.
10145 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
10147 (cortexa7_older_only): Likewise.
10148 (cortexa7_younger): Likewise.
10150 2013-07-18 David Malcolm <dmalcolm@redhat.com>
10152 * ipa-pure-const.c (generate_summary): Rename to...
10153 (pure_const_generate_summary): ... this.
10155 2013-07-17 Iain Sandoe <iain@codesourcery.com>
10157 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
10159 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
10162 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
10165 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10167 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
10168 enabled without -march=zEC12.
10169 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
10172 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
10174 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
10175 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
10176 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
10177 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
10178 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
10179 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
10180 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
10182 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
10183 and ISA_HAS_NMADD3_NMSUB3.
10184 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
10185 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
10186 (nmsub4<mode>, nmsub3<mode>): Likewise.
10187 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
10189 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
10191 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
10192 TARGET_MIPS5400 checking.
10194 2013-07-16 Jakub Jelinek <jakub@redhat.com>
10195 Peter Bergner <bergner@vnet.ibm.com>
10197 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
10198 registers in the comment.
10199 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
10200 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
10201 rather than FIRST_PSEUDO_REGISTERS.
10203 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
10205 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
10206 enable extra ISA flags with TARGET_HTM.
10208 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
10210 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
10213 2013-07-15 Cong Hou <congh@google.com>
10215 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
10216 in compare function for sorting.
10218 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
10220 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
10221 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
10222 * config/rs6000/rs6000.opt: Add -mhtm option.
10223 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
10224 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
10225 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
10226 __HTM__ if the HTM instructions are available.
10227 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
10228 htm_spr_reg_operand): New define_predicates.
10229 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
10230 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
10232 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
10233 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
10234 HTM builtin functions.
10235 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
10236 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
10237 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
10238 (rs6000_builtin_mask_calculate): Likewise.
10239 (rs6000_option_override_internal): Likewise.
10240 (bdesc_htm): Add new HTM builtin support.
10241 (htm_spr_num): New function.
10242 (htm_spr_regno): Likewise.
10243 (rs6000_htm_spr_icode): Likewise.
10244 (htm_expand_builtin): Likewise.
10245 (htm_init_builtins): Likewise.
10246 (rs6000_expand_builtin): Add support for HTM builtin functions.
10247 (rs6000_init_builtins): Likewise.
10248 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
10250 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
10251 (TARGET_HTM, MASK_HTM): Define macros.
10252 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
10253 (FIXED_REGISTERS): Likewise.
10254 (CALL_USED_REGISTERS): Likewise.
10255 (CALL_REALLY_USED_REGISTERS): Likewise.
10256 (REG_ALLOC_ORDER): Likewise.
10257 (enum reg_class): Likewise.
10258 (REG_CLASS_NAMES): Likewise.
10259 (REG_CLASS_CONTENTS): Likewise.
10260 (REGISTER_NAMES): Likewise.
10261 (ADDITIONAL_REGISTER_NAMES): Likewise.
10262 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
10263 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
10264 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
10265 * config/rs6000/htm.md: New file.
10266 * config/rs6000/htmintrin.h: New file.
10267 * config/rs6000/htmxlintrin.h: New file.
10269 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
10271 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
10272 Define SYMBOL_TINY_GOT, update comment.
10273 * config/aarch64/aarch64.c
10274 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
10275 (aarch64_expand_mov_immediate): Likewise.
10276 (aarch64_print_operand): Likewise.
10277 (aarch64_classify_symbol): Likewise.
10278 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
10279 (ldr_got_tiny): Define.
10281 2013-07-13 Tobias Grosser <tobias@grosser.es>
10283 PR tree-optimization/54094
10284 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
10285 scheduling dimension for the parallelism check from the polyhedral
10286 information in the AST.
10287 * graphite-dependences.c (carries_deps): Do not assume the schedule is
10290 2013-07-13 Jason Merrill <jason@redhat.com>
10292 * print-tree.c (debug_vec_tree): Use debug_raw.
10293 (debug_raw (vec<tree, va_gc> &)): New.
10294 (debug_raw (vec<tree, va_gc> *)): New.
10295 * tree.h: Declare them.
10297 2013-07-13 Bin Cheng <bin.cheng@arm.com>
10299 * ifcvt.c (ifcvt_after_combine): New static variable.
10300 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
10302 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
10303 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
10304 rest_of_handle_if_after_reload): Pass new argument for if_convert.
10306 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
10308 * config/mips/mips.c (mips_expand_call): Remove empty statement.
10310 2013-07-12 Michael Matz <matz@suse.de>
10312 PR middle-end/55771
10313 * convert.c (convert_to_real): Reject non-float inner types.
10315 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
10317 * config/aarch64/aarch64-protos.h
10318 (aarch64_simd_immediate_valid_for_move): Remove.
10319 * config/aarch64/aarch64.c (simd_immediate_info): New member.
10320 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
10322 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
10324 2013-07-11 Steve Ellcey <sellcey@mips.com>
10326 * config/mips/mips.c (mips_conditional_register_usage): Do not
10327 use t[0-7] registers in MIPS16 mode when optimizing for size.
10329 2013-07-11 Sriraman Tallam <tmsriram@google.com>
10331 * config/i386/i386.c (dispatch_function_versions): Fix array
10332 indexing of function_version_info to match actual_versions.
10334 2013-07-11 Teresa Johnson <tejohnson@google.com>
10336 * vec.h (struct va_gc): Move release out-of-line.
10337 (va_gc::release): Call ggc_free on released vec.
10339 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
10341 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
10342 Require GOT register as additional operand in UNSPEC.
10343 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
10344 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
10345 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
10346 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
10347 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
10348 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
10349 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
10351 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
10354 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
10355 name seen by assembler/linker rather if available.
10357 2013-07-11 Andreas Schwab <schwab@suse.de>
10359 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
10361 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
10363 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
10365 2013-07-10 Joseph Myers <joseph@codesourcery.com>
10367 * doc/tm.texi.in: Move hook documentation to ....
10368 * target.def: ... here.
10370 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
10371 text on @hook line.
10372 * doc/tm.texi: Regenerate.
10374 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
10377 * doc/invoke.texi: Document -Wconditionally-supported.
10379 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
10382 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
10385 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
10388 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
10389 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
10390 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
10391 Remove duplicate devices.
10392 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
10393 * config/avr/t-multilib: Regenerate.
10394 * config/avr/avr-tables.opt: Regenerate.
10395 * doc/avr-mmcu.texi: Regenerate.
10397 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
10400 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
10402 2013-07-10 Graham Stott <graham.stott@btinternet.com>
10404 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
10405 the cost of MULT when optimizing for size.
10407 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10409 * config/cr16/cr16-protos.h: Don't include target.h.
10411 2013-07-09 Joseph Myers <joseph@codesourcery.com>
10413 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
10414 adjust register size for TDmode and TFmode for VSX registers.
10416 2013-07-08 Kai Tietz <ktietz@redhat.com>
10419 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
10420 hook_bool_const_tree_true.
10422 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10424 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
10425 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
10426 * config/s390/s390.md: Define FPR*_REGNUM constants.
10427 Fix FPR2_REGNUM constant (18 -> 17).
10428 ("*trunc<BFP:mode><DFP_ALL:mode>2")
10429 ("*trunc<DFP_ALL:mode><BFP:mode>2")
10430 ("trunc<BFP:mode><DFP_ALL:mode>2")
10431 ("trunc<DFP_ALL:mode><BFP:mode>2")
10432 ("*extend<BFP:mode><DFP_ALL:mode>2")
10433 ("*extend<DFP_ALL:mode><BFP:mode>2")
10434 ("extend<BFP:mode><DFP_ALL:mode>2")
10435 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
10438 2013-07-08 Graham Stott <graham.stott@btinternet.com>
10440 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
10442 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10444 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
10445 and cfun_fpr_bit_p to cfun_fpr_save_p.
10446 (s390_frame_area, s390_register_info, s390_frame_info)
10447 (s390_emit_prologue, s390_emit_epilogue)
10448 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
10450 * config/s390/s390.h: Define *_REGNUM macros for floating point
10453 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
10455 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
10457 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
10459 PR rtl-optimization/57786
10460 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
10461 and break out of the loop when it is set to false.
10463 2013-07-08 Jakub Jelinek <jakub@redhat.com>
10466 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
10467 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
10468 (const_int 63)) 0)).
10469 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
10470 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
10471 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
10473 PR rtl-optimization/57829
10474 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
10475 mask bits outside of mode are just sign-extension from mode to HWI.
10477 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
10479 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
10480 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
10481 adjust_address instead of change_address to keep info about alignment.
10482 (emit_strmov): Remove.
10483 (emit_memmov): New function.
10484 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
10485 (expand_movmem_epilogue): Likewise and return updated rtx for
10487 (expand_constant_movmem_prologue): Likewise and return updated rtx for
10488 destination and source.
10489 (decide_alignment): Refactor, handle vector_loop.
10490 (ix86_expand_movmem): Likewise.
10491 (ix86_expand_setmem): Likewise.
10492 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
10494 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
10496 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
10497 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
10499 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
10501 * config/i386/sse.md (sse_movlhps): Change alternative 3
10502 of operand 2 to "m".
10504 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
10507 * config/i386/sse.md (iptr): New mode attribute.
10508 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
10509 (<sse>_vm<plusminus_insn><mode>3): Ditto.
10510 (<sse>_vmmul<mode>3): Ditto.
10511 (<sse>_vmdiv<mode>3): Ditto.
10512 (sse_vmrcpv4sf2): Ditto.
10513 (<sse>_vmsqrt<mode>2): Ditto.
10514 (sse_vmrsqrtv4sf2): Ditto.
10515 (<sse>_vm<code><mode>3): Ditto.
10516 (avx_vmcmp<mode>3): Ditto.
10517 (<sse>_vmmaskcmp<mode>3): Ditto.
10518 (<sse>_comi): Ditto.
10519 (<sse>_ucomi): Ditto.
10520 (*xop_vmfrcz_<mode>): Ditto.
10521 (*fmai_fmadd_<mode>): Ditto.
10522 (*fmai_fmsub_<mode>): Ditto.
10523 (*fmai_fnmadd_<mode>): Ditto.
10524 (*fmai_fnmsub_<mode>): Ditto.
10525 (*fma4i_vmfmadd_<mode>): Ditto.
10526 (*fma4i_vmfmsub_<mode>): Ditto.
10527 (*fma4i_vmfnmadd_<mode>): Ditto.
10528 (*fma4i_vmfnmsub_<mode>): Ditto.
10529 (*xop_vmfrcz_<mode>): Ditto.
10530 (sse_cvtps2pi): Ditto.
10531 (sse_cvttps2pi): Ditto.
10532 (sse_cvtss2si): Ditto.
10533 (sse_cvtss2si_2): Ditto.
10534 (sse_cvtss2siq_2): Ditto.
10535 (sse_cvttss2si): Ditto.
10536 (sse_cvttss2siq): Ditto.
10537 (sse_cvtsd2si): Ditto.
10538 (sse_cvtsd2si_2): Ditto.
10539 (sse_cvtsd2siq_2): Ditto.
10540 (sse_cvttsd2si): Ditto.
10541 (sse_cvttsd2siq): Ditto.
10542 (sse_cvtsd2ss): Ditto.
10543 (sse_cvtss2sd): Ditto.
10544 (avx2_pbroadcast<mode>): Ditto.
10545 (avx2_pbroadcast<mode>_1): Ditto.
10546 (*avx_vperm_broadcast_v4sf): Ditto.
10548 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
10549 (sse_movlhps): Ditto.
10550 (sse_storehps): Ditto.
10551 (sse_loadhps): Ditto.
10552 (sse_storelps): Ditto.
10553 (sse_loadlps): Ditto.
10554 (*vec_concatv4sf): Ditto.
10555 (*vec_interleave_highv2df): Ditto.
10556 (*vec_interleave_lowv2df): Ditto.
10557 (*vec_extractv2df_1_sse): Ditto.
10558 (*vec_extractv2df_0_sse): Ditto.
10559 (sse2_storelpd): Ditto.
10560 (sse2_loadlpd): Ditto.
10561 (sse2_movsd): Ditto.
10562 (*vec_concatv4si): Ditto.
10563 (vec_concatv2di): Ditto.
10565 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
10566 for Intel asm dialect.
10567 (mmx_punpcklwd): Ditto.
10568 (mmx_punpckldq): Ditto.
10570 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
10571 for intel assembler dialect.
10573 2013-07-06 Jakub Jelinek <jakub@redhat.com>
10576 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
10577 for BUILT_IN_C{LZ,LRSB}*.
10578 * tree.h (CASE_INT_FN): Add FN##IMAX case.
10579 * tree-vrp.c (extract_range_basic): Handle
10580 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
10581 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
10582 fall thru to code calling set_value*.
10583 * builtins.c (expand_builtin): Remove *IMAX cases.
10584 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
10585 if width is bigger than 2*HWI.
10587 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
10589 PR rtl-optimization/55342
10590 * lra-int.h (lra_subreg_reload_pseudos): New.
10591 * lra.c: Add undoing optional reloads to the block diagram.
10592 (lra_subreg_reload_pseudos): New.
10593 (lra_optional_reload_pseudos): Change comments.
10594 (lra): Init and clear lra_subreg_reload_pseudos. Clear
10595 lra_optional_reload_pseudos after undo transformations.
10596 * lra-assigns.c (pseudo_prefix_title): New.
10597 (lra_setup_reg_renumber): Use it.
10598 (spill_for): Ditto. Check subreg reload pseudos too.
10599 (assign_by_spills): Consider subreg reload pseudos too.
10600 * lra-constraints.c (simplify_operand_subreg): Use
10601 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
10602 (curr_insn_transform): Recognize and do optional reloads.
10603 (undo_optional_reloads): New.
10604 (lra_undo_inheritance): Call undo_optional_reloads.
10606 2013-07-05 Thomas Quinot <quinot@adacore.com>
10608 * tree-complex.c (expand_complex_operations_1): Fix typo.
10610 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
10612 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
10613 (tune_params): New member 'const vec_costs'.
10614 * config/aarch64/aarch64.c (generic_vector_cost): New.
10615 (generic_tunings): New member 'generic_vector_cost'.
10616 (aarch64_builtin_vectorization_cost): New.
10617 (aarch64_add_stmt_cost): New.
10618 (TARGET_VECTORIZE_ADD_STMT_COST): New.
10619 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
10621 2013-07-03 Jakub Jelinek <jakub@redhat.com>
10624 * config/i386/predicates.md (vsib_address_operand): Disallow
10625 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
10627 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
10629 PR middle-end/55030
10630 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
10631 expand_builtin_setjmp_receiver.
10632 (expand_label): Adjust, call expand_builtin_setjmp_receiver
10633 with NULL for the label parameter.
10634 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
10635 the frame-pointer. Adjust comments.
10636 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
10637 only if LABEL is non-NULL.
10639 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
10641 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
10642 (ARM_ABI_AAPCS64): Ditto.
10644 (ARM_DEFAULT_ABI): Ditto.
10646 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
10648 * config/aarch64/aarch64-builtins.c
10649 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
10650 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
10652 * config/aarch64/aarch64-simd.md
10653 (aarch64_ld1<VALL:mode>): New.
10654 (aarch64_st1<VALL:mode>): Likewise.
10655 * config/aarch64/arm_neon.h
10656 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
10658 2013-07-02 Sriraman Tallam <tmsriram@google.com>
10660 * config/i386/i386.c (gate_insert_vzeroupper): Check if
10662 (ix86_option_override_internal):Turn on all -mavx target flags by
10663 default as they are dependent on AVX anyway.
10665 2013-07-02 Cary Coutant <ccoutant@google.com>
10667 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
10668 deterministic hash.
10669 (loc_checksum_ordered): Likewise.
10670 (hash_loc_operands): Remove inline keyword.
10672 2013-07-02 Jakub Jelinek <jakub@redhat.com>
10674 PR tree-optimization/57741
10675 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
10676 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
10677 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
10678 Allow REAL_CST step_exprs if flag_associative_math.
10679 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
10681 2013-07-02 Ian Bolton <ian.bolton@arm.com>
10683 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
10685 2013-07-02 Ian Bolton <ian.bolton@arm.com>
10687 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
10689 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10691 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
10693 (iorsi3_insn): Likewise.
10694 (arm_xorsi3): Likewise.
10696 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
10698 * arm.md (attribute "wtype"): Delete. Move attribute values from here
10700 (attribute "type"): ... here, and prefix with "wmmx_".
10701 (attribute "core_cycles"): Update for attribute changes.
10702 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
10703 (tbcstv4hi): Likewise.
10704 (tbcstv2si): Likewise.
10705 (iwmmxt_iordi3): Likewise.
10706 (iwmmxt_xordi3): Likewise.
10707 (iwmmxt_anddi3): Likewise.
10708 (iwmmxt_nanddi3): Likewise.
10709 (iwmmxt_arm_movdi): Likewise.
10710 (iwmmxt_movsi_insn): Likewise.
10711 (mov<mode>_internal): Likewise.
10712 (and<mode>3_iwmmxt): Likewise.
10713 (ior<mode>3_iwmmxt): Likewise.
10714 (xor<mode>3_iwmmxt): Likewise.
10715 (add<mode>3_iwmmxt): Likewise.
10716 (ssaddv8qi3): Likewise.
10717 (ssaddv4hi3): Likewise.
10718 (ssaddv2si3): Likewise.
10719 (usaddv8qi3): Likewise.
10720 (usaddv4hi3): Likewise.
10721 (usaddv2si3): Likewise.
10722 (sub<mode>3_iwmmxt): Likewise.
10723 (sssubv8qi3): Likewise.
10724 (sssubv4hi3): Likewise.
10725 (sssubv2si3): Likewise.
10726 (ussubv8qi3): Likewise.
10727 (ussubv4hi3): Likewise.
10728 (ussubv2si3): Likewise.
10729 (mulv4hi3_iwmmxt): Likewise.
10730 (smulv4hi3_highpart): Likewise.
10731 (umulv4hi3_highpart): Likewise.
10732 (iwmmxt_wmacs): Likewise.
10733 (iwmmxt_wmacsz): Likewise.
10734 (iwmmxt_wmacu): Likewise.
10735 (iwmmxt_wmacuz): Likewise.
10736 (iwmmxt_clrdi): Likewise.
10737 (iwmmxt_clrv8qi): Likewise.
10738 (iwmmxt_clr4hi): Likewise.
10739 (iwmmxt_clr2si): Likewise.
10740 (iwmmxt_uavgrndv8qi3): Likewise.
10741 (iwmmxt_uavgrndv4hi3): Likewise.
10742 (iwmmxt_uavgv8qi3): Likewise.
10743 (iwmmxt_uavgv4hi3): Likewise.
10744 (iwmmxt_tinsrb): Likewise.
10745 (iwmmxt_tinsrh): Likewise.
10746 (iwmmxt_tinsrw): Likewise.
10747 (iwmmxt_textrmub): Likewise.
10748 (iwmmxt_textrmsb): Likewise.
10749 (iwmmxt_textrmuh): Likewise.
10750 (iwmmxt_textrmsh): Likewise.
10751 (iwmmxt_textrmw): Likewise.
10752 (iwmxxt_wshufh): Likewise.
10753 (eqv8qi3): Likewise.
10754 (eqv4hi3): Likewise.
10755 (eqv2si3): Likewise.
10756 (gtuv8qi3): Likewise.
10757 (gtuv4hi3): Likewise.
10758 (gtuv2si3): Likewise.
10759 (gtv8qi3): Likewise.
10760 (gtv4hi3): Likewise.
10761 (gtv2si3): Likewise.
10762 (smax<mode>3_iwmmxt): Likewise.
10763 (umax<mode>3_iwmmxt): Likewise.
10764 (smin<mode>3_iwmmxt): Likewise.
10765 (umin<mode>3_iwmmxt): Likewise.
10766 (iwmmxt_wpackhss): Likewise.
10767 (iwmmxt_wpackwss): Likewise.
10768 (iwmmxt_wpackdss): Likewise.
10769 (iwmmxt_wpackhus): Likewise.
10770 (iwmmxt_wpackwus): Likewise.
10771 (iwmmxt_wpackdus): Likewise.
10772 (iwmmxt_wunpckihb): Likewise.
10773 (iwmmxt_wunpckihh): Likewise.
10774 (iwmmxt_wunpckihw): Likewise.
10775 (iwmmxt_wunpckilb): Likewise.
10776 (iwmmxt_wunpckilh): Likewise.
10777 (iwmmxt_wunpckilw): Likewise.
10778 (iwmmxt_wunpckehub): Likewise.
10779 (iwmmxt_wunpckehuh): Likewise.
10780 (iwmmxt_wunpckehuw): Likewise.
10781 (iwmmxt_wunpckehsb): Likewise.
10782 (iwmmxt_wunpckehsh): Likewise.
10783 (iwmmxt_wunpckehsw): Likewise.
10784 (iwmmxt_wunpckelub): Likewise.
10785 (iwmmxt_wunpckeluh): Likewise.
10786 (iwmmxt_wunpckeluw): Likewise.
10787 (iwmmxt_wunpckelsb): Likewise.
10788 (iwmmxt_wunpckelsh): Likewise.
10789 (iwmmxt_wunpckelsw): Likewise.
10790 (ror<mode>3): Likewise.
10791 (ashr<mode>3_iwmmxt): Likewise.
10792 (lshr<mode>3_iwmmxt): Likewise.
10793 (ashl<mode>3_iwmmxt): Likewise.
10794 (ror<mode>3_di): Likewise.
10795 (ashr<mode>3_di): Likewise.
10796 (lshr<mode>3_di): Likewise.
10797 (ashl<mode>3_di): Likewise.
10798 (iwmmxt_wmadds): Likewise.
10799 (iwmmxt_wmaddu): Likewise.
10800 (iwmmxt_tmia): Likewise.
10801 (iwmmxt_tmiaph): Likewise.
10802 (iwmmxt_tmiabb): Likewise.
10803 (iwmmxt_tmiatb): Likewise.
10804 (iwmmxt_tmiabt): Likewise.
10805 (iwmmxt_tmiatt): Likewise.
10806 (iwmmxt_tmovmskb): Likewise.
10807 (iwmmxt_tmovmskh): Likewise.
10808 (iwmmxt_tmovmskw): Likewise.
10809 (iwmmxt_waccb): Likewise.
10810 (iwmmxt_wacch): Likewise.
10811 (iwmmxt_waccw): Likewise.
10812 (iwmmxt_waligni): Likewise.
10813 (iwmmxt_walignr): Likewise.
10814 (iwmmxt_walignr0): Likewise.
10815 (iwmmxt_walignr1): Likewise.
10816 (iwmmxt_walignr2): Likewise.
10817 (iwmmxt_walignr3): Likewise.
10818 (iwmmxt_wsadb): Likewise.
10819 (iwmmxt_wsadh): Likewise.
10820 (iwmmxt_wsadbz): Likewise.
10821 (iwmmxt_wsadhz): Likewise.
10822 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
10823 (iwmmxt_wabsdiffb): Likewise.
10824 (iwmmxt_wabsdiffh): Likewise.
10825 (iwmmxt_wabsdiffw): Likewise.
10826 (iwmmxt_waddsubhx): Likewise
10827 (iwmmxt_wsubaddhx): Likewise.
10828 (addc<mode>3): Likewise.
10829 (iwmmxt_avg4): Likewise.
10830 (iwmmxt_avg4r): Likewise.
10831 (iwmmxt_wmaddsx): Likewise.
10832 (iwmmxt_wmaddux): Likewise.
10833 (iwmmxt_wmaddsn): Likewise.
10834 (iwmmxt_wmaddun): Likewise.
10835 (iwmmxt_wmulwsm): Likewise.
10836 (iwmmxt_wmulwum): Likewise.
10837 (iwmmxt_wmulsmr): Likewise.
10838 (iwmmxt_wmulumr): Likewise.
10839 (iwmmxt_wmulwsmr): Likewise.
10840 (iwmmxt_wmulwumr): Likewise.
10841 (iwmmxt_wmulwl): Likewise.
10842 (iwmmxt_wqmulm): Likewise.
10843 (iwmmxt_wqmulwm): Likewise.
10844 (iwmmxt_wqmulmr): Likewise.
10845 (iwmmxt_wqmulwmr): Likewise.
10846 (iwmmxt_waddbhusm): Likewise.
10847 (iwmmxt_waddbhusl): Likewise.
10848 (iwmmxt_wqmiabb): Likewise.
10849 (iwmmxt_wqmiabt): Likewise.
10850 (iwmmxt_wqmiatb): Likewise.
10851 (iwmmxt_wqmiatt): Likewise.
10852 (iwmmxt_wqmiabbn): Likewise.
10853 (iwmmxt_wqmiabtn): Likewise.
10854 (iwmmxt_wqmiatbn): Likewise.
10855 (iwmmxt_wqmiattn): Likewise.
10856 (iwmmxt_wmiabb): Likewise.
10857 (iwmmxt_wmiabt): Likewise.
10858 (iwmmxt_wmiatb): Likewise.
10859 (iwmmxt_wmiatt): Likewise.
10860 (iwmmxt_wmiabbn): Likewise.
10861 (iwmmxt_wmiabtn): Likewise.
10862 (iwmmxt_wmiatbn): Likewise.
10863 (iwmmxt_wmiattn): Likewise.
10864 (iwmmxt_wmiawbb): Likewise.
10865 (iwmmxt_wmiawbt): Likewise.
10866 (iwmmxt_wmiawtb): Likewise.
10867 (iwmmxt_wmiawtt): Likewise.
10868 (iwmmxt_wmiawbbn): Likewise.
10869 (iwmmxt_wmiawbtn): Likewise.
10870 (iwmmxt_wmiawtbn): Likewise.
10871 (iwmmxt_wmiawttn): Likewise.
10872 (iwmmxt_wmerge): Likewise.
10873 (iwmmxt_tandc<mode>3): Likewise.
10874 (iwmmxt_torc<mode>3): Likewise.
10875 (iwmmxt_torvsc<mode>3): Likewise.
10876 (iwmmxt_textrc<mode>3): Likewise.
10877 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
10878 (wmmxt_pack): Likewise.
10879 (wmmxt_mult_c1): Likewise.
10880 (wmmxt_mult_c2): Likewise.
10881 (wmmxt_alu_c1): Likewise.
10882 (wmmxt_alu_c2): Likewise.
10883 (wmmxt_alu_c3): Likewise.
10884 (wmmxt_transfer_c1): Likewise.
10885 (wmmxt_transfer_c2): Likewise.
10886 (wmmxt_transfer_c3): Likewise.
10887 (marvell_f_iwmmxt_wstr): Likewise.
10888 (marvell_f_iwmmxt_wldr): Likewise.
10890 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
10892 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
10894 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
10897 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
10898 * lra-constraints.c (need_for_split_p): Check call used hard regs
10899 living through calls.
10901 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
10902 call used regs for call insn.
10904 2013-06-28 Jakub Jelinek <jakub@redhat.com>
10907 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
10908 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
10910 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
10912 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
10914 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
10916 * lra-constraints.c (need_for_split_p): Check call used hard regs
10917 living through calls.
10919 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
10922 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
10923 to tie with any other modes. Eliminate Altivec vector mode tests,
10924 since these are a subset of ALTIVEC or VSX vector modes. Simplify
10925 code, to return 0 if testing MODE2 for a condition, if we've
10926 already tested MODE1 for the same condition.
10928 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
10930 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
10933 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
10935 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
10936 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
10938 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
10940 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
10942 (aarch64_symbolic_constant_p): Remove.
10943 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
10944 static. Fix line length and white space.
10945 (aarch64_symbolic_constant_p): Remove.
10946 * config/aarch64/predicates.md (aarch64_valid_symref):
10947 Use aarch64_classify_symbol_expression.
10949 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10951 * config/arm/constraints.md (Ts): New constraint.
10952 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
10954 (compare_scc): Use "Ts" constraint for operand 0.
10955 (ior_scc_scc): Likewise.
10956 (and_scc_scc): Likewise.
10957 (and_scc_scc_nodom): Likewise.
10958 (ior_scc_scc_cmp): Likewise for operand 7.
10959 (and_scc_scc_cmp): Likewise.
10960 * config/arm/thumb2.md (thumb2_movsi_insn):
10961 Add alternatives for 16-bit encodings.
10962 (thumb2_movhi_insn): Likewise.
10963 (thumb2_movsicc_insn): Likewise.
10964 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
10965 (thumb2_negscc): Use "Ts" constraint.
10966 Move mvn instruction outside cond_exec block.
10967 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
10968 for 16-bit encodings.
10970 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10972 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
10974 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
10975 (mulsi3subsi): Likewise.
10976 (mulsidi3adddi): Likewise.
10977 (mulsidi3_v6): Likewise.
10978 (umulsidi3_v6): Likewise.
10979 (umulsidi3adddi_v6): Likewise.
10980 (smulsi3_highpart_v6): Likewise.
10981 (umulsi3_highpart_v6): Likewise.
10982 (mulhisi3tb): Likewise.
10983 (mulhisi3bt): Likewise.
10984 (mulhisi3tt): Likewise.
10985 (maddhisi4): Likewise.
10986 (maddhisi4tb): Likewise.
10987 (maddhisi4tt): Likewise.
10988 (maddhidi4): Likewise.
10989 (maddhidi4tb): Likewise.
10990 (maddhidi4tt): Likewise.
10991 (zeroextractsi_compare0_scratch): Likewise.
10992 (insv_zero): Likewise.
10993 (insv_t2): Likewise.
10994 (anddi_notzesidi_di): Likewise.
10995 (anddi_notsesidi_di): Likewise.
10996 (andsi_notsi_si): Likewise.
10997 (iordi_zesidi_di): Likewise.
10998 (xordi_zesidi_di): Likewise.
10999 (andsi_iorsi3_notsi): Likewise.
11000 (smax_0): Likewise.
11001 (smax_m1): Likewise.
11002 (smin_0): Likewise.
11003 (not_shiftsi): Likewise.
11004 (unaligned_loadsi): Likewise.
11005 (unaligned_loadhis): Likewise.
11006 (unaligned_loadhiu): Likewise.
11007 (unaligned_storesi): Likewise.
11008 (unaligned_storehi): Likewise.
11009 (extv_reg): Likewise.
11010 (extzv_t2): Likewise.
11011 (divsi3): Likewise.
11012 (udivsi3): Likewise.
11013 (arm_zero_extendhisi2addsi): Likewise.
11014 (arm_zero_extendqisi2addsi): Likewise.
11015 (compareqi_eq0): Likewise.
11016 (arm_extendhisi2_v6): Likewise.
11017 (arm_extendqisi2addsi): Likewise.
11018 (arm_movt): Likewise.
11019 (thumb2_ldrd): Likewise.
11020 (thumb2_ldrd_base): Likewise.
11021 (thumb2_ldrd_base_neg): Likewise.
11022 (thumb2_strd): Likewise.
11023 (thumb2_strd_base): Likewise.
11024 (thumb2_strd_base_neg): Likewise.
11025 (arm_negsi2): Add alternative for 16-bit encoding.
11026 (arm_one_cmplsi2): Likewise.
11028 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11030 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
11031 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
11032 (movdfcc): Likewise.
11033 * config/arm/vfp.md (*thumb2_movsf_vfp):
11034 Disable predication for arm_restrict_it.
11035 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
11036 (*thumb2_movdfcc_vfp): Likewise.
11037 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
11038 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
11039 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
11040 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
11041 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
11042 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
11043 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
11044 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
11045 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
11046 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
11047 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
11048 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
11049 Disable predication for arm_restrict_it.
11051 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
11053 * config/i386/bmiintrin.h (_bextr_u32): New.
11054 (_bextr_u64): Ditto.
11056 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
11058 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
11059 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
11060 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
11061 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
11062 * config/mips/n32-elf.h: ...this new file.
11064 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
11067 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
11068 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
11070 2013-06-27 Catherine Moore <clm@codesourcery.com>
11072 * config/mips/mips-tables.opt: Regenerate.
11073 * config/mips/mips-cpus.def: Add m14ke and m14kec.
11074 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
11075 * doc/invoke.texi: Add -m14kc.
11077 2013-06-27 Jakub Jelinek <jakub@redhat.com>
11080 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
11081 constraints of operand 1 and 2.
11084 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
11085 to match RTL canonicalization. Swap predicates and
11086 constraints of operand 1 and 2.
11088 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
11090 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
11091 Process OP_INOUT regs for splitting too.
11093 2013-06-27 Jakub Jelinek <jakub@redhat.com>
11095 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
11096 decl before the loop, initialize to NULL.
11097 (vectorizable_load): Initialize ptr_incr to NULL.
11099 2013-06-27 Martin Jambor <mjambor@suse.cz>
11102 * ipa-ref.h (ipa_maybe_record_reference): Declare.
11103 * ipa-ref.c (ipa_maybe_record_reference): New function.
11104 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
11105 * ipa-cp.c (create_specialized_node): Record potential references from
11107 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
11109 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
11111 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
11112 parameter 'mode' of type 'enum machine_mode mode'; change to pass
11113 'mode' to force_reg.
11114 (aarch64_add_offset): Update calls to aarch64_force_temporary.
11115 (aarch64_expand_mov_immediate): Likewise.
11117 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
11119 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
11120 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
11122 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11124 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
11125 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
11126 (struct machine_function): Add tbegin_p.
11127 (s390_canonicalize_comparison): Fold CC mode compares to
11128 conditional jump if possible.
11129 (s390_emit_jump): Return the emitted jump.
11130 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
11131 Handle CCRAWmode compares.
11132 (s390_option_override): Default to -mhtm if available.
11133 (s390_reg_clobbered_rtx): Handle floating point regs as well.
11134 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
11135 FPRs instead of df_regs_ever_live_p.
11136 (s390_optimize_nonescaping_tx): New function.
11137 (s390_init_frame_layout): Extend clobbered_regs array to cover
11139 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
11140 (s390_expand_tbegin): New function.
11141 (enum s390_builtin): New enum definition.
11142 (code_for_builtin): New array definition.
11143 (s390_init_builtins): New function.
11144 (s390_expand_builtin): New function.
11145 (TARGET_INIT_BUILTINS): Define.
11146 (TARGET_EXPAND_BUILTIN): Define.
11147 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
11148 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
11149 (s390_alc_comparison): Likewise.
11150 * config/s390/s390-modes.def: Add CCRAWmode.
11151 * config/s390/s390.h (processor_flags): Add PF_TX.
11152 (TARGET_CPU_HTM): Define macro.
11153 (TARGET_HTM): Define macro.
11154 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
11155 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
11156 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
11157 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
11158 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
11159 (TBEGIN_MASK, TBEGINC_MASK): New constants.
11160 ("*cc_to_int"): Move up.
11161 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
11162 constants other than 0.
11163 ("*ccraw_to_int"): New insn and splitter definition.
11164 ("tbegin", "tbegin_nofloat", "tbegin_retry")
11165 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
11166 ("tx_assist"): New expander.
11167 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
11168 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
11169 * config/s390/s390.opt: Add -mhtm option.
11170 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
11171 * config/s390/htmxlintrin.h: New file.
11172 * config/s390/htmintrin.h: New file.
11173 * config/s390/s390intrin.h: New file.
11174 * doc/extend.texi: Document htm builtins.
11175 * config.gcc: Add the new header files to extra_headers.
11177 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
11179 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
11180 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
11182 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
11183 Pat Haugen <pthaugen@us.ibm.com>
11184 Peter Bergner <bergner@vnet.ibm.com>
11186 * config/rs6000/power8.md: New.
11187 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
11188 setting for power8 entry.
11189 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
11190 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
11191 test for Power4/Power5 only.
11192 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
11194 (force_new_group): Adjust comment.
11195 * config/rs6000/rs6000.md: Include power8.md.
11197 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
11199 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
11200 * config/arm/arm-protos.h (arm_max_conditional_execute): New
11202 (tune_params): Update comment.
11203 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
11204 (arm_max_conditional_execute): New function.
11205 (thumb2_final_prescan_insn): Use max_insn_skipped and
11206 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
11208 2013-06-25 Jakub Jelinek <jakub@redhat.com>
11210 PR tree-optimization/57705
11211 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
11212 SSA_NAME step, provided that it is not defined inside the loop.
11213 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
11214 (get_initial_def_for_induction): Handle SSA_NAME IV step.
11216 2013-06-25 Martin Jambor <mjambor@suse.cz>
11218 PR middle-end/57670
11219 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
11220 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
11222 (ipa_note_param_call): Initialize member_ptr flag.
11223 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
11224 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
11225 (ipa_write_indirect_edge_info): Stream member_ptr flag.
11226 (ipa_read_indirect_edge_info): Likewise.
11228 2013-06-25 Richard Biener <rguenther@suse.de>
11230 PR middle-end/56977
11231 * passes.c (init_optimization_passes): Move pass_fold_builtins
11232 and pass_dce earlier with -Og.
11234 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
11236 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
11237 <BIT_FIELD_REF>: Remove trailing TAB.
11238 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
11241 2013-06-24 Martin Jambor <mjambor@suse.cz>
11243 PR tree-optimization/57358
11244 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
11245 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
11246 (ipa_analyze_params_uses): Generate pessimistic info when true.
11248 2013-06-24 Martin Jambor <mjambor@suse.cz>
11250 PR tree-optimization/57539
11251 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
11252 global.inlined_to of the new node to it. All callers changed.
11253 * ipa-inline-transform.c (clone_inlined_nodes): New variable
11254 inlining_into, pass it to cgraph_clone_node.
11255 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
11256 ipa_free_edge_args_substructures.
11257 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
11258 rdesc linked list. Do not assert rdesc edges have inlined caller.
11259 Assert we have found an rdesc in the rdesc list.
11261 2013-06-24 Richard Biener <rguenther@suse.de>
11263 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
11264 (pointer_set_lookup): Declare.
11265 (class pointer_map): New template class implementing a
11266 generic pointer to T map.
11267 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
11268 pointer_map<T>::contains, pointer_map<T>::insert,
11269 pointer_map<T>::traverse): New functions.
11270 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
11271 (pointer_set_lookup): New function.
11272 (pointer_set_contains): Use pointer_set_lookup.
11273 (pointer_set_insert): Likewise.
11274 (insert_aux): Remove.
11275 (struct pointer_map_t): Embed a pointer_set_t.
11276 (pointer_map_create): Adjust.
11277 (pointer_map_destroy): Likewise.
11278 (pointer_map_contains): Likewise.
11279 (pointer_map_insert): Likewise.
11280 (pointer_map_traverse): Likewise.
11281 * tree-streamer.h (struct streamer_tree_cache_d): Use a
11282 pointer_map<unsigned> instead of a pointer_map_t.
11283 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
11284 (streamer_tree_cache_lookup): Likewise.
11285 (streamer_tree_cache_create): Likewise.
11286 (streamer_tree_cache_delete): Likewise.
11287 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
11288 pointer_map<unsigned> instead of a pointer_map_t.
11289 (lto_init_tree_ref_encoder): Adjust.
11290 (lto_destroy_tree_ref_encoder): Likewise.
11291 * lto-section-out.c (lto_output_decl_index): Likewise.
11292 (lto_record_function_out_decl_state): Likewise.
11293 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
11295 2013-06-24 Richard Biener <rguenther@suse.de>
11297 PR tree-optimization/57488
11298 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
11300 2013-06-24 Alan Modra <amodra@gmail.com>
11302 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
11303 (gen_easy_altivec_constant): Likewise.
11304 * config/rs6000/predicates.md (easy_vector_constant_add_self,
11305 easy_vector_constant_msb): Likewise.
11307 2013-06-23 Jakub Jelinek <jakub@redhat.com>
11310 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
11311 add missing return true.
11313 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
11316 * config/sh/predicates.md (general_extend_operand): Invoke
11317 general_movsrc_operand for memory operands.
11318 (general_movsrc_operand): Allow reg+reg addressing, do not use
11319 general_operand for memory operands.
11321 2013-06-23 Sriraman Tallam <tmsriram@google.com>
11323 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
11324 when current target options does not apply.
11325 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
11326 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
11327 * config/i386/bmiintrin.h: Pass appropriate target
11328 attributes to header.
11329 * config/i386/mmintrin.h: Ditto.
11330 * config/i386/nmmintrin.h: Ditto.
11331 * config/i386/avx2intrin.h: Ditto.
11332 * config/i386/fxsrintrin.h: Ditto.
11333 * config/i386/tbmintrin.h: Ditto.
11334 * config/i386/xsaveintrin.h: Ditto.
11335 * config/i386/f16cintrin.h: Ditto.
11336 * config/i386/xtestintrin.h: Ditto.
11337 * config/i386/xsaveoptintrin.h: Ditto.
11338 * config/i386/bmi2intrin.h: Ditto.
11339 * config/i386/lzcntintrin.h: Ditto.
11340 * config/i386/smmintrin.h: Ditto.
11341 * config/i386/wmmintrin.h: Ditto.
11342 * config/i386/x86intrin.h: Remove all header include guards.
11343 * config/i386/prfchwintrin.h: Ditto.
11344 * config/i386/pmmintrin.h: Ditto.
11345 * config/i386/tmmintrin.h: Ditto.
11346 * config/i386/xmmintrin.h: Ditto.
11347 * config/i386/popcntintrin.h: Ditto.
11348 * config/i386/rdseedintrin.h: Ditto.
11349 * config/i386/ammintrin.h: Ditto.
11350 * config/i386/emmintrin.h: Ditto.
11351 * config/i386/immintrin.h: Remove all header include guards.
11352 * config/i386/fma4intrin.h: Ditto.
11353 * config/i386/lwpintrin.h: Ditto.
11354 * config/i386/xopintrin.h: Ditto.
11355 * config/i386/ia32intrin.h: Ditto.
11356 * config/i386/avxintrin.h: Ditto.
11357 * config/i386/rtmintrin.h: Ditto.
11358 * config/i386/fmaintrin.h: Ditto.
11359 * config/i386/mm3dnow.h: Ditto.
11361 2013-06-22 Sriraman Tallam <tmsriram@google.com>
11363 * common/config/i386/i386-common.c: Handle LZCNT.
11365 2013-06-22 Andi Kleen <ak@linux.intel.com>
11367 * doc/extend.texi: Use __atomic_store_n instead of
11368 __atomic_store in HLE example.
11370 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
11372 * config/sh/sh.c: Remove <cstdlib> workaround.
11374 2013-06-21 Andi Kleen <ak@linux.intel.com>
11376 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
11378 2013-06-21 Andi Kleen <ak@linux.intel.com>
11380 * doc/extend.texi: Document that __atomic_clear and
11381 __atomic_test_and_set should only be used with bool.
11383 2013-06-20 Jan Hubicka <jh@suse.cz>
11385 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
11386 types_same_for_odr.
11387 * tree.c (decls_same_for_odr): New function.
11388 (same_for_edr): New function.
11389 (types_same_for_odr): New function.
11390 (get_binfo_at_offset): Use it.
11391 * tree.h (types_same_for_odr): Declare.
11393 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
11394 Jason Merrill <jason@redhat.com>
11396 * system.h: Include <cstdlib> as well as <stdlib.h>.
11398 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
11401 * config/i386/i386.c (construct_container): Report error if
11402 long double is used with disabled x87 float returns.
11404 2013-06-20 Jan Hubicka <jh@suse.cz>
11406 * lto-cgraph.c (input_symtab): Do not set cgraph state.
11408 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
11410 PR rtl-optimization/57425
11411 PR rtl-optimization/57569
11412 * alias.c (write_dependence_p): Remove parameters mem_mode and
11413 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
11414 Changed all callers.
11415 (canon_anti_dependence): Get comments and semantics in sync.
11416 Add parameter mem_canonicalized. Changed all callers.
11417 * rtl.h (canon_anti_dependence): Update prototype.
11419 2013-06-20 Richard Biener <rguenther@suse.de>
11421 * data-streamer-in.c (streamer_read_uhwi): Optimize single
11422 byte case, inline streamer_read_uchar and defer section
11425 2013-06-20 Richard Biener <rguenther@suse.de>
11427 PR tree-optimization/57584
11428 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
11429 SSA names into the expanded expression that take part in
11430 abnormal coalescing.
11432 2013-06-19 Sharad Singhai <singhai@google.com>
11434 * gcov.c (print_usage): Handle new option.
11435 (process_args): Ditto.
11436 (get_gcov_intermediate_filename): New function.
11437 (output_intermediate_file): New function.
11438 (output_gcov_file): New function
11439 (generate_results): Handle new option.
11440 (release_function): Relase demangled name.
11441 (read_graph_file): Handle demangled name.
11442 (output_lines): Ditto.
11443 * doc/gcov.texi: Document gcov intermediate format.
11445 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
11448 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
11449 (lra_emit_add): Use the functions. Add comment about Y as an
11452 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
11455 * collect2.c (collect_atexit): New.
11456 (collect_exit): Delete.
11457 (main): Register collect_atexit with atexit.
11458 (collect_wait): Change collect_exit to exit.
11460 * collect2.h (collect_exit): Delete.
11461 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
11463 2013-06-19 Wei Mi <wmi@google.com>
11465 PR rtl-optimization/57518
11466 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
11467 if regno is used in paradoxical subreg.
11468 (update_equiv_regs): Check pdx_subregs[regno] before
11469 set a reg to be equivalent with a mem.
11471 2013-06-19 Matthias Klose <doko@ubuntu.com>
11474 * file-find.h (find_a_file): Add a mode parameter.
11475 * file-find.c (find_a_file): Likewise.
11476 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
11477 with X_OK for the executables.
11478 * collect2.c (main): Call find_a_file with X_OK.
11480 2013-06-19 Steve Ellcey <sellcey@mips.com>
11483 * config/mips/mips.md (casesi_internal_mips16_<mode>):
11484 Use NEXT_INSN instead of next_real_insn.
11486 2013-06-19 Jan Hubicka <jh@suse.cz>
11488 * cgraph.h (const_value_known_p): Replace by ...
11489 (ctor_for_folding): .. this one.
11490 * cgraphunit.c (process_function_and_variable_attributes): Use it.
11491 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
11492 * expr.c (expand_expr_real_1): Likewise.
11493 (string_constant): Likewise.
11494 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
11495 * ipa.c (process_references): Likewise.
11496 (symtab_remove_unreachable_nodes): Likewise.
11497 * ipa-inline-analysis.c (param_change_prob): Likewise.
11498 * gimple-fold.c (canonicalize_constructor_val): Likewise.
11499 (get_base_constructor): Likwise.
11500 * varpool.c (varpool_remove_node): Likewise.
11501 (varpool_remove_initializer): LIkewise.
11502 (dump_varpool_node): LIkwise.
11503 (const_value_known_p): Rewrite to ...
11504 (ctor_for_folding): ... this one.
11506 2013-06-19 Jakub Jelinek <jakub@redhat.com>
11509 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
11510 PERSONALITY in $PATH derived prefixes.
11512 2013-06-19 Jeff Law <law@redhat.com>
11514 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
11517 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
11518 (simplify_bitwise_binary): Use it to simpify certain binary ops on
11521 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
11523 * config/arm/vfp.md: Move VFP instruction classification documentation
11525 * config/arm/arm.md: ... here. Update instruction classification
11528 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
11530 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
11531 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
11532 pattern. Use more efficient sequences on ARMv5 and Thumb2.
11534 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
11537 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
11538 with NEXT_INSN. Use tablejump_p to check for jump table data
11541 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
11544 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
11545 that now in C++ the value is correct per the C++ standards.
11547 2013-06-19 Richard Biener <rguenther@suse.de>
11549 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
11550 for global context.
11552 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11555 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11558 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
11559 with next_active_insn.
11561 2013-06-18 Sriraman Tallam <tmsriram@google.com>
11563 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
11564 functions are inlined during failures to flag an error.
11565 * tree-inline.c (expand_call_inline): Allow the error to be flagged
11566 in early inline pass.
11568 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
11570 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
11573 2013-06-18 Julian Brown <julian@codesourcery.com>
11575 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
11576 Permit virtual register pre-reload if !strict.
11577 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
11579 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
11581 * config/arm/neon.md (movmisalign<mode>): Use
11582 neon_perm_struct_or_reg_operand instead of
11583 neon_struct_or_register_operand.
11584 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
11585 neon_permissive_struct_operand instead of neon_struct_operand.
11586 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
11587 neon_vector_mem_operand.
11588 * config/arm/predicates.md (neon_struct_operand): Adjust call to
11589 neon_vector_mem_operand.
11590 (neon_permissive_struct_operand): New.
11591 (neon_struct_or_register_operand): Rename to...
11592 (neon_perm_struct_or_reg_operand): This. Adjust call to
11593 neon_vector_mem_operand.
11595 2013-06-18 Richard Biener <rguenther@suse.de>
11597 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
11598 * lto-streamer.h: Include pointer-set.h.
11599 (struct lto_decl_slot): Remove.
11600 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
11601 Remove next_index entry.
11602 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
11603 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
11604 (lto_init_tree_ref_encoder): Adjust.
11605 (lto_destroy_tree_ref_encoder): Likewise.
11606 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
11607 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
11608 (lto_output_decl_index): Adjust.
11609 (lto_new_out_decl_state): Likewise.
11610 (lto_record_function_out_decl_state): Likewise.
11611 * lto-streamer-out.c (copy_function): Likewise.
11613 2013-06-18 Richard Biener <rguenther@suse.de>
11615 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
11616 * cgraphunit.c: Include cfgloop.h.
11617 (init_lowered_empty_function): Initialize the loop tree.
11618 (assemble_thunk): Insert new BBs into loops.
11620 2013-06-18 Richard Biener <rguenther@suse.de>
11622 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
11623 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
11624 the map from cache entry to cache index optional.
11625 (streamer_tree_cache_replace_tree): Adjust accordingly.
11626 (streamer_tree_cache_append): Likewise.
11627 (streamer_tree_cache_delete): Likewise.
11628 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
11629 streamer cache map from cache entry to cache index.
11630 * lto-streamer-out.c (create_output_block): Adjust.
11632 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
11634 * config/arm/arm.md (attribute "insn"): Move multiplication and
11635 division attributes to...
11636 (attribute "type"): ... here. Remove mult.
11637 (attribute "mul32"): New attribute.
11638 (attribute "mul64"): Add umaal.
11639 (*arm_mulsi3): Update attributes.
11640 (*arm_mulsi3_v6): Likewise.
11641 (*thumb_mulsi3): Likewise.
11642 (*thumb_mulsi3_v6): Likewise.
11643 (*mulsi3_compare0): Likewise.
11644 (*mulsi3_compare0_v6): Likewise.
11645 (*mulsi_compare0_scratch): Likewise.
11646 (*mulsi_compare0_scratch_v6): Likewise.
11647 (*mulsi3addsi): Likewise.
11648 (*mulsi3addsi_v6): Likewise.
11649 (*mulsi3addsi_compare0): Likewise.
11650 (*mulsi3addsi_compare0_v6): Likewise.
11651 (*mulsi3addsi_compare0_scratch): Likewise.
11652 (*mulsi3addsi_compare0_scratch_v6): Likewise.
11653 (*mulsi3subsi): Likewise.
11654 (*mulsidi3adddi): Likewise.
11655 (*mulsi3addsi_v6): Likewise.
11656 (*mulsidi3adddi_v6): Likewise.
11657 (*mulsidi3_nov6): Likewise.
11658 (*mulsidi3_v6): Likewise.
11659 (*umulsidi3_nov6): Likewise.
11660 (*umulsidi3_v6): Likewise.
11661 (*umulsidi3adddi): Likewise.
11662 (*umulsidi3adddi_v6): Likewise.
11663 (*smulsi3_highpart_nov6): Likewise.
11664 (*smulsi3_highpart_v6): Likewise.
11665 (*umulsi3_highpart_nov6): Likewise.
11666 (*umulsi3_highpart_v6): Likewise.
11667 (mulhisi3): Likewise.
11668 (*mulhisi3tb): Likewise.
11669 (*mulhisi3bt): Likewise.
11670 (*mulhisi3tt): Likewise.
11671 (maddhisi4): Likewise.
11672 (*maddhisi4tb): Likewise.
11673 (*maddhisi4tt): Likewise.
11674 (maddhidi4): Likewise.
11675 (*maddhidi4tb): Likewise.
11676 (*maddhidi4tt): Likewise.
11677 (divsi3): Likewise.
11678 (udivsi3): Likewise.
11679 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
11680 (thumb2_mulsi_short_compare0): Likewise.
11681 (thumb2_mulsi_short_compare0_scratch): Likewise.
11682 * config/arm/arm1020e.md (1020mult1): Update attribute change.
11683 (1020mult2): Likewise.
11684 (1020mult3): Likewise.
11685 (1020mult4): Likewise.
11686 (1020mult5): Likewise.
11687 (1020mult6): Likewise.
11688 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
11690 (cortex_a15_mult64): Likewise.
11691 (cortex_a15_sdiv): Likewise.
11692 (cortex_a15_udiv): Likewise.
11693 * config/arm/arm1026ejs.md (mult1): Update attribute change.
11699 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
11700 (pj4_ir_div): Likewise.
11701 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
11702 (11_mult2): Likewise.
11703 (11_mult3): Likewise.
11704 (11_mult4): Likewise.
11705 (11_mult5): Likewise.
11706 (11_mult6): Likewise.
11707 (11_mult7): Likewise.
11708 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
11709 (cortex_a8_mla): Likewise.
11710 (cortex_a8_mull): Likewise.
11711 (cortex_a8_smulwy): Likewise.
11712 (cortex_a8_smlald): Likewise.
11713 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
11714 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
11715 (cortex_r4_mul_3): Likewise.
11716 (cortex_r4_mla_4): Likewise.
11717 (cortex_r4_mla_3): Likewise.
11718 (cortex_r4_smlald): Likewise.
11719 (cortex_r4_mull): Likewise.
11720 (cortex_r4_sdiv): Likewise.
11721 (cortex_r4_udiv): Likewise.
11722 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
11723 (cortex_a7_idiv): Likewise.
11724 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
11725 (9_mult2): Likewise.
11726 (9_mult3): Likewise.
11727 (9_mult4): Likewise.
11728 (9_mult5): Likewise.
11729 (9_mult6): Likewise.
11730 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
11731 (cortex_a53_sdiv): Likewise.
11732 (cortex_a53_udiv): Likewise.
11733 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
11734 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
11735 (mp626_mult2): Likewise.
11736 (mp626_mult3): Likewise.
11737 (mp626_mult4): Likewise.
11738 * config/arm/fa526.md (526_mult1): Update attribute change.
11739 (526_mult2): Likewise.
11740 * config/arm/arm-generic.md (mult): Update attribute change.
11741 (mult_ldsched_strongarm): Likewise.
11742 (mult_ldsched): Likewise.
11743 (multi_cycle): Likewise.
11744 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
11745 * config/arm/fa606te.md (606te_mult1): Update attribute change.
11746 (606te_mult2): Likewise.
11747 (606te_mult3): Likewise.
11748 (606te_mult4): Likewise.
11749 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
11750 (cortex_a9_mac16): Likewise.
11751 (cortex_a9_multiply): Likewise.
11752 (cortex_a9_mac): Likewise.
11753 (cortex_a9_multiply_long): Likewise.
11754 * config/arm/fa626te.md (626te_mult1): Update attribute change.
11755 (626te_mult2): Likewise.
11756 (626te_mult3): Likewise.
11757 (626te_mult4): Likewise.
11759 2013-06-18 Richard Biener <rguenther@suse.de>
11762 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
11764 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11767 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
11768 with next_active_insn.
11770 2013-06-18 Alan Modra <amodra@gmail.com>
11772 * config/rs6000/rs6000.h (enum data_align): New.
11773 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
11774 (DATA_ABI_ALIGNMENT): Define.
11775 (CONSTANT_ALIGNMENT): Correct comment.
11776 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
11777 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
11779 2013-06-17 David Malcolm <dmalcolm@redhat.com>
11781 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
11782 ATTRIBUTE_UNUSED marking.
11784 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
11786 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
11787 alternative and update.
11788 (aarch64_dup_lanedi): Delete.
11789 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
11790 * config/aarch64/aarch64-simd-builtins.def: Update.
11792 2013-06-17 Richard Biener <rguenther@suse.de>
11794 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
11795 (lto_input_scc): Declare.
11796 (lto_input_tree_1): Likewise.
11797 (struct lto_stats_d): Add num_tree_bodies_output and
11798 num_pickle_refs_output.
11799 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
11800 (lto_read_tree_1): Split out from ...
11801 (lto_read_tree): ... this.
11802 (lto_input_scc): New function.
11803 (lto_input_tree_1): Split out from ...
11804 (lto_input_tree): ... this. Handle LTO_tree_scc.
11805 (lto_data_in_create): Create the streamer cache without hashes.
11806 * lto-streamer-out.c (create_output_block): Create the streamer
11807 cache with hashes when not doing WPA.
11808 (lto_write_tree_1): Split out from ...
11809 (lto_write_tree): ... this.
11810 (get_symbol_initial_value): New function.
11811 (lto_output_tree_1): Split out from ...
11812 (lto_output_tree): ... this. Write trees as series of SCCs
11813 using a DFS walk via DFS_write_tree.
11814 (struct sccs, struct scc_entry): New types.
11815 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
11816 (DFS_write_tree_body): New function.
11817 (DFS_write_tree): Likewise.
11818 (hash_tree): Likewise.
11819 (scc_entry_compare): Likewise.
11820 (hash_scc): Likewise.
11821 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
11822 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
11823 TREE_CHAIN as regular reference.
11824 (streamer_read_integer_cst): Remove.
11825 (streamer_get_pickled_tree): Adjust.
11826 * tree-streamer-out.c (streamer_write_chain): Disable streaming
11827 of DECL_EXTERNALs in BLOCK_VARS for now.
11828 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
11830 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
11831 Add hash value argument and record that if hashes are recorded
11833 (streamer_tree_cache_insert_1): Adjust.
11834 (streamer_tree_cache_insert): Likewise.
11835 (streamer_tree_cache_insert_at): Rename to ...
11836 (streamer_tree_cache_replace_tree): ... this and adjust.
11837 (streamer_tree_cache_append): Adjust.
11838 (record_common_node): Likewise.
11839 (streamer_tree_cache_create): Add argument whether to
11840 record hash values together with trees.
11841 (streamer_tree_cache_delete): Adjust.
11842 * tree-streamer.h (struct streamer_tree_cache_d): Add
11844 (streamer_read_integer_cst): Remove.
11845 (streamer_tree_cache_insert): Adjust.
11846 (streamer_tree_cache_append): Likewise.
11847 (streamer_tree_cache_insert_at): Rename to ...
11848 (streamer_tree_cache_replace_tree): ... this and adjust.
11849 (streamer_tree_cache_create): Add argument whether to record hashes.
11850 (streamer_tree_cache_get): Rename to ...
11851 (streamer_tree_cache_get_tree): ... this.
11852 (streamer_tree_cache_get_hash): New function.
11853 * tree.c (cache_integer_cst): New function.
11854 * tree.h (cache_integer_cst): Declare.
11855 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
11856 * lto-symtab.c (lto_varpool_replace_node): Only release
11857 DECL_INITIAL of non-prevailing decls.
11858 * varpool.c (varpool_remove_initializer): Do not release
11859 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
11861 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
11863 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
11864 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
11865 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
11866 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
11867 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
11868 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
11869 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
11870 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
11871 instead of TARGET_64BIT.
11872 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
11873 Require ISA_HAS_<D>DIV.
11875 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
11877 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
11878 (mips*-*-linux*): Move default with_llsc setting to where other
11880 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
11882 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
11883 Likewise. Remove default with_tune setting. Move default float
11884 setting to its own block. Handle with_llsc in the same block as above.
11886 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
11888 PR rtl-optimization/57425
11889 PR rtl-optimization/57569
11890 * alias.c (write_dependence_p): Add new parameters mem_mode,
11891 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
11892 Changed all callers.
11893 (canon_anti_dependence): New function.
11894 * cse.c (check_dependence): Use canon_anti_dependence.
11895 * cselib.c (cselib_invalidate_mem): Likewise.
11896 * rtl.h (canon_anti_dependence): Declare.
11898 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
11900 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
11901 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
11902 ".set mips3" for 64-bit targets.
11904 2013-06-15 Dehao Chen <dehao@google.com>
11906 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
11907 * gimple-low.c (gimple_check_call_matching_types): Likewise.
11908 (gimple_check_call_args): Likewise.
11909 * value-prof.c (check_ic_target): Likewise.
11910 * ipa-inline.c (early_inliner): Likewise.
11911 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
11912 * cgraph.c (cgraph_create_edge_1): Likewise.
11913 (cgraph_make_edge_direct): Likewise.
11915 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
11918 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
11919 rs6000_output_move_128bit to handle emitting quad memory
11920 operations. Set attribute length to 8 bytes.
11922 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
11924 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
11926 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
11927 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
11928 (aarch64_<su>mlsl<mode>): Likewise.
11930 2013-06-14 Mike Stump <mikestump@comcast.net>
11932 * Makefile.in (TARGET_H): Add insn-codes.h.
11934 2013-06-14 Alan Modra <amodra@gmail.com>
11936 PR middle-end/57134
11937 PR middle-end/57586
11938 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
11939 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
11940 bitfield expansion when EXPAND_MEMORY.
11941 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
11943 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
11945 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
11946 test for clearing quad memory on 32-bit later.
11948 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
11950 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
11951 (fold_negate_expr): Likewise.
11952 (fold_real_zero_addition_p): Handle vectors.
11953 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
11955 2013-06-14 Alan Modra <amodra@gmail.com>
11957 * varasm.c (force_const_mem): Revert 2013-06-07 change.
11959 2013-06-13 Jan Hubicka <jh@suse.cz>
11961 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
11962 Local comdats are not externally visible.
11963 * symtab.c (dump_symtab_base): Dump externally visible.
11964 (verify_symtab_base): Verify back links in the symtab hash.
11966 2013-06-13 Bin Cheng <bin.cheng@arm.com>
11968 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
11969 CONVERT_EXPR as equal nodes.
11971 2013-06-13 Bin Cheng <bin.cheng@arm.com>
11973 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
11975 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
11977 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
11978 Generalize to complex and vector.
11979 * tree.c (build_all_ones_cst): New function.
11980 * tree.h (build_all_ones_cst): Declare it.
11982 2013-06-13 Alan Modra <amodra@gmail.com>
11984 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
11985 * config/rs6000/rs6000.md (signbittf2): New insn.
11986 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
11987 (abstf2_internal, cmptf_internal2): Likewise.
11988 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
11990 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
11991 Pat Haugen <pthaugen@us.ibm.com>
11992 Peter Bergner <bergner@vnet.ibm.com>
11994 * config/rs6000/rs6000.c (emit_load_locked): Add support for
11995 power8 byte, half-word, and quad-word atomic instructions.
11996 (emit_store_conditional): Likewise.
11997 (rs6000_expand_atomic_compare_and_swap): Likewise.
11998 (rs6000_expand_atomic_op): Likewise.
12000 * config/rs6000/sync.md (larx): Add new modes for power8.
12002 (AINT): New mode iterator to include TImode as well as normal
12003 integer modes on power8.
12004 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
12005 that VSX registers are not considered. Use AINT mode iterator
12006 instead of INT1 to allow inclusion of quad word atomic operations
12008 (load_locked<mode>): Likewise.
12009 (store_conditional<mode>): Likewise.
12010 (atomic_compare_and_swap<mode>): Likewise.
12011 (atomic_exchange<mode>): Likewise.
12012 (atomic_nand<mode>): Likewise.
12013 (atomic_fetch_<fetchop_name><mode>): Likewise.
12014 (atomic_nand_fetch<mode>): Likewise.
12015 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
12017 (ATOMIC): On power8, add QImode, HImode modes.
12018 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
12019 modes that promote to SImode.
12020 (load_lockedti): Convert TImode arguments to PTImode, so that we
12021 get a guaranteed even/odd register pair.
12022 (load_lockedpti): Likewise.
12023 (store_conditionalti): Likewise.
12024 (store_conditionalpti): Likewise.
12026 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
12027 atomic load/store instructions.
12030 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
12032 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
12034 (insn_count): New attribute, with most cases extracted from...
12035 (length): ...here. Redefine most cases in terms of insn_count.
12036 (single_insn): Delete.
12037 (can_delay): Use insn_count to check for single instructions.
12038 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
12039 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
12040 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
12041 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
12042 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
12043 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
12044 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
12045 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
12046 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
12047 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
12048 rather than "length".
12049 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
12050 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
12051 Use "insn_count" rather than "length".
12052 * config/mips/mips-dsp.md
12053 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
12054 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
12057 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
12059 PR tree-optimization/57361
12060 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
12062 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
12064 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
12066 (aarch64_simd_combine<mode>): New instruction expansion.
12067 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
12068 function prototype.
12069 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
12070 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
12072 2013-06-12 Jan Hubicka <jh@suse.cz>
12074 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
12075 decl has when in streaming stage.
12076 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
12077 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
12079 2013-06-12 Roland Stigge <stigge@antcom.de>
12082 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
12084 2013-06-12 Jakub Jelinek <jakub@redhat.com>
12086 PR tree-optimization/57537
12087 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
12088 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
12090 2013-06-12 Richard Biener <rguenther@suse.de>
12092 * data-streamer.h (streamer_write_char_stream): CSE
12093 obs->current_pointer.
12094 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
12095 streamer_write_char_stream manually and optimize the resulting loop.
12096 (streamer_write_hwi_stream): Likewise.
12098 2013-06-12 Jan Hubicka <jh@suse.cz>
12100 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
12101 * cgraph.h (varpool_create_empty_node): Declare.
12102 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
12104 * symtab.c (symtab_unregister_node): Be lax about missin entries
12106 (symtab_get_node): Update comment.
12107 * varpool.c (varpool_create_empty_node): Break out from ...
12108 (varpool_node_for_decl): ... here.
12109 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
12111 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
12113 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
12114 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
12115 part. Use straight-line flow at the end.
12116 <COMPONENT_REF>: Remove superfluous else.
12117 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
12119 2013-06-12 Jakub Jelinek <jakub@redhat.com>
12122 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
12123 target hook even for !TREE_PUBLIC decls. If no resolution info
12124 is available, return false for common and external decls.
12126 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
12128 * config/rl78/constraints.md (U): New constraint.
12129 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
12132 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
12135 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
12136 to allow returning address to AT_PLATFORM name.
12138 2013-06-11 Jan Hubicka <jh@suse.cz>
12140 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
12141 * cgraph.h (symtab_node_base): Add weakref flag.
12142 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
12143 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
12144 (output_weakrefs): Use weakref flag.
12145 * fold-const.c (simple_operand_p): Handle WEAK.
12146 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
12147 * ipa.c (varpool_externally_visible_p): Drop weakref.
12148 (function_and_variable_visibility): Update comment; fix weakref
12149 sanity checks; do not clear DECL_WEAK on them.
12150 * lto-cgraph.c (lto_output_node): update.
12151 (lto_output_varpool_node): Update.
12152 (input_overwrite_node): Update.
12153 (input_node): Update.
12154 (input_varpool_node): Update.
12155 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
12156 (lto_symtab_merge_symbols): Add sanity check.
12157 (lto_symtab_prevailing_decl): Do not special case weakrefs.
12158 * passes.c (rest_of_decl_compilation): Set static flag, too.
12159 * symtab.c (dump_symtab_base): Dump weakref.
12160 (verify_symtab_base): Sanity check weakrefs.
12161 (symtab_make_decl_local): Remove duplicated code.
12162 (symtab_alias_ultimate_target): Simplify.
12163 * varpool.c (varpool_create_variable_alias): Set weakref flag.
12165 2013-06-11 Tom de Vries <tom@codesourcery.com>
12167 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
12168 sequence_vect == NULL.
12170 2013-06-11 DJ Delorie <dj@redhat.com>
12172 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
12173 (rl78_unwind_word_mode): New.
12175 2013-06-11 David Malcolm <dmalcolm@redhat.com>
12177 * final.c (debug_prefix_maps): Make static.
12179 2013-06-11 David Malcolm <dmalcolm@redhat.com>
12181 * function.c (initial_trampoline): Remove stray copy.
12183 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
12185 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
12187 2013-06-11 Martin Jambor <mjambor@suse.cz>
12189 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
12190 within bounds at the beginning of the function.
12192 2013-06-11 Alan Modra <amodra@gmail.com>
12194 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
12196 (get_named_section): Don't NULL !DECL_P decl.
12198 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
12200 * doc/invoke.texi (core-avx2): Document.
12202 (atom): Updated with MOVBE.
12204 2013-06-11 Richard Biener <rguenther@suse.de>
12206 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
12208 2013-06-11 Anton Blanchard <anton@samba.org>
12210 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
12211 correct shift value in little-endian mode.
12213 2013-06-11 Jakub Jelinek <jakub@redhat.com>
12216 * varasm.c (get_variable_align): Move #endif to the right place.
12218 2013-06-10 Cary Coutant <ccoutant@google.com>
12220 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
12221 for hash so that hash table traversal order is deterministic.
12223 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
12224 Pat Haugen <pthaugen@us.ibm.com>
12225 Peter Bergner <bergner@vnet.ibm.com>
12227 * config/rs6000/vector.md (GPR move splitter): Do not split moves
12228 of vectors in GPRS if they are direct moves or quad word load or
12231 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
12233 (direct_move_p): Likewise.
12234 (quad_load_store_p): Likewise.
12236 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
12237 classes into bins based on the physical register type.
12238 (reg_class_to_reg_type): Likewise.
12239 (IS_STD_REG_TYPE): Likewise.
12240 (IS_FP_VECT_REG_TYPE): Likewise.
12241 (reload_fpr_gpr): Arrays to determine what insn to use if we can
12242 use direct move instructions.
12243 (reload_gpr_vsx): Likewise.
12244 (reload_vsx_gpr): Likewise.
12245 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
12246 information that is a simplification of register classes. Also
12247 precalculate direct move reload helpers.
12248 (direct_move_p): New function to return true if the operation can
12249 be done as a direct move instruciton.
12250 (quad_load_store_p): New function to return true if the operation
12251 is a quad memory operation.
12252 (rs6000_legitimize_address): If quad memory, only allow register
12253 indirect for TImode addresses.
12254 (rs6000_legitimate_address_p): Likewise.
12255 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
12256 (rs6000_reload_register_type): Likewise.
12257 (register_to_reg_type): Return register type.
12258 (rs6000_secondary_reload_simple_move): New helper function for
12259 secondary reload and secondary memory needed to identify anything
12260 that is a simple move, and does not need reloading.
12261 (rs6000_secondary_reload_direct_move): New helper function for
12262 secondary reload to identify cases that can be done with several
12263 instructions via the direct move instructions.
12264 (rs6000_secondary_reload_move): New helper function for secondary
12265 reload to identify moves between register types that can be done.
12266 (rs6000_secondary_reload): Add support for quad memory operations
12267 and for direct move.
12268 (rs6000_secondary_memory_needed): Likewise.
12269 (rs6000_debug_secondary_memory_needed): Change argument names.
12270 (rs6000_output_move_128bit): New function to return the move to
12271 use for 128-bit moves, including knowing about the various
12272 limitations of quad memory operations.
12274 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
12275 memory operations. call rs6000_output_move_128bit for the actual
12276 instruciton(s) to generate.
12277 (vsx_movti_64bit): Likewise.
12279 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
12280 (UNSPEC_P8V_MTVSRWZ): Likewise.
12281 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
12282 (UNSPEC_P8V_MTVSRD): Likewise.
12283 (UNSPEC_P8V_XXPERMDI): Likewise.
12284 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
12285 (UNSPEC_FUSION_GPR): Likewise.
12286 (FMOVE128_GPR): New iterator for direct move.
12287 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
12288 (f32_sv): Likewise.
12289 (f32_dm): Likewise.
12290 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
12291 loads and direct move instructions.
12292 (zero_extendsidi2_lfiwzx): Likewise.
12293 (extendsidi2_lfiwax): Likewise.
12294 (extendsidi2_nocell): Likewise.
12295 (floatsi<mode>2_lfiwax): Likewise.
12296 (lfiwax): Likewise.
12297 (floatunssi<mode>2_lfiwzx): Likewise.
12298 (lfiwzx): Likewise.
12299 (fix_trunc<mode>_stfiwx): Likewise.
12300 (fixuns_trunc<mode>_stfiwx): Likewise.
12301 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
12302 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
12303 (parity<mode>2_cmpb): Set length/type attr.
12304 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
12305 for 'mr.' to fast_compare.
12306 (bpermd_<mode>): Change type attr to popcnt.
12307 (p8_fmrgow_<mode>): New insns for power8 direct move support.
12308 (p8_mtvsrwz_1): Likewise.
12309 (p8_mtvsrwz_2): Likewise.
12310 (reload_fpr_from_gpr<mode>): Likewise.
12311 (p8_mtvsrd_1): Likewise.
12312 (p8_mtvsrd_2): Likewise.
12313 (p8_xxpermdi_<mode>): Likewise.
12314 (reload_vsx_from_gpr<mode>): Likewise.
12315 (reload_vsx_from_gprsf): Likewise.
12316 (p8_mfvsrd_3_<mode>): LIkewise.
12317 (reload_gpr_from_vsx<mode>): Likewise.
12318 (reload_gpr_from_vsxsf): Likewise.
12319 (p8_mfvsrd_4_disf): Likewise.
12320 (multi-word GPR splits): Do not split direct moves or quad memory
12323 2013-06-10 David Malcolm <dmalcolm@redhat.com>
12325 * tree-into-ssa.c (interesting_blocks): Make static.
12327 2013-06-10 Jakub Jelinek <jakub@redhat.com>
12330 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
12331 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
12332 Use DATA_ABI_ALIGNMENT for that case instead if defined.
12333 (get_variable_align): New function.
12334 (get_variable_section, emit_bss, emit_common,
12335 assemble_variable_contents, place_block_symbol): Use
12336 get_variable_align instead of DECL_ALIGN.
12337 (assemble_noswitch_variable): Add align argument, use it
12338 instead of DECL_ALIGN.
12339 (assemble_variable): Adjust caller. Use get_variable_align
12340 instead of DECL_ALIGN.
12341 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
12343 (DATA_ABI_ALIGNMENT): Define.
12344 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
12345 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
12346 opt is false, only return the psABI mandated alignment increase.
12347 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
12348 (DATA_ABI_ALIGNMENT): ... this.
12349 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
12350 (DATA_ABI_ALIGNMENT): ... this.
12351 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
12352 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
12353 (DATA_ABI_ALIGNMENT): ... this.
12354 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
12355 * doc/tm.texi: Regenerated.
12357 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
12359 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
12360 cmp_code to construct REG_EQUAL note.
12362 2013-06-09 Jakub Jelinek <jakub@redhat.com>
12365 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
12366 that operands[2] doesn't overlap with operands[0].
12368 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
12369 Jan Hubicka <jh@suse.cz>
12371 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
12372 hack to mark symbols as used.
12374 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
12376 PR rtl-optimization/57559
12377 * lra-constraints.c (process_alt_operands): Don't discourage
12378 memory with known offset for offsetable memory constraint.
12379 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
12381 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
12383 * varasm.c (struct oc_local_state): Reorder fields.
12384 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
12385 and adjust accordingly.
12386 (output_constructor): Reorder initialization code and adjust call to
12387 output_constructor_bitfield.
12389 2013-06-07 Jan Hubicka <jh@suse.cz>
12391 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
12393 2013-06-07 David Malcolm <dmalcolm@redhat.com>
12395 * tree-object-size.c (unknown): Make const.
12397 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12399 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
12400 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
12401 for last alternative in the cpu_facility attribute.
12403 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12406 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
12407 (xordi3): Change operand 2 constraint to arm_xordi_operand.
12408 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
12409 * config/arm/constraints.md (Dg): New constraint.
12410 * config/arm/neon.md (xordi3_neon): Remove.
12411 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
12412 * config/arm/predicates.md (arm_xordi_operand): New predicate.
12414 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12416 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
12417 Clean up alternatives.
12419 2013-06-07 Alan Modra <amodra@gmail.com>
12421 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
12424 2013-06-07 Alan Modra <amodra@gmail.com>
12426 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
12428 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12430 * config/arm/constraints.md (Df): New constraint.
12431 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
12432 Correct length attribute for last two alternatives.
12434 2013-06-07 Alan Modra <amodra@gmail.com>
12436 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
12437 override user -mfp-in-toc.
12438 (offsettable_ok_by_alignment): Consider just the current access
12439 rather than the whole object, unless BLKmode. Handle
12440 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
12441 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
12442 for -mcmodel=medium.
12443 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
12444 override user -mfp-in-toc or -msum-in-toc. Default to
12445 -mno-fp-in-toc for -mcmodel=medium.
12447 2013-06-06 DJ Delorie <dj@redhat.com>
12449 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
12450 TARGET_VALID_POINTER_MODE.
12452 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
12453 Pat Haugen <pthaugen@us.ibm.com>
12454 Peter Bergner <bergner@vnet.ibm.com>
12456 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
12457 Document new power8 builtins.
12459 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
12460 condition code register, to allow 128-bit logical operations to be
12461 done in the VSX or GPR registers.
12462 (nor<mode>3): Use the canonical form for nor.
12463 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
12464 vclz*, and vpopcnt* vector instructions.
12465 (nand<mode>3): Likewise.
12466 (orc<mode>3): Likewise.
12467 (clz<mode>2): LIkewise.
12468 (popcount<mode>2): Likewise.
12470 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
12471 that only the GPRs are recognized.
12473 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
12474 support for new power8 builtins.
12476 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
12478 (xscvdpspn): Likewise.
12484 (vpopcnt): Likewise.
12485 (vpopcntb): Likewise.
12486 (vpopcnth): Likewise.
12487 (vpopcntw): Likewise.
12488 (vpopcntd): Likewise.
12490 (vmrgew): Likewise.
12491 (vmrgow): Likewise.
12493 (eqv_v16qi3): Likewise.
12494 (eqv_v8hi3): Likewise.
12495 (eqv_v4si3): Likewise.
12496 (eqv_v2di3): Likewise.
12497 (eqv_v4sf3): Likewise.
12498 (eqv_v2df3): Likewise.
12500 (nand_v16qi3): Likewise.
12501 (nand_v8hi3): Likewise.
12502 (nand_v4si3): Likewise.
12503 (nand_v2di3): Likewise.
12504 (nand_v4sf3): Likewise.
12505 (nand_v2df3): Likewise.
12507 (orc_v16qi3): Likewise.
12508 (orc_v8hi3): Likewise.
12509 (orc_v4si3): Likewise.
12510 (orc_v2di3): Likewise.
12511 (orc_v4sf3): Likewise.
12512 (orc_v2df3): Likewise.
12514 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
12515 allow power8 quad mode in 64-bit.
12516 (rs6000_builtin_vectorized_function): Add support to vectorize
12517 ISA 2.07 count leading zeros, population count builtins.
12518 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
12519 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
12520 (builtin_function_type): Add vgbbd builtin function which takes an
12522 (altivec_expand_vec_perm_const): Add support for new power8 merge
12525 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
12526 that does not include TImdoe for use with 32-bit.
12527 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
12529 (UNSPEC_VSX_CVDPSPN): Likewise.
12530 (vsx_xscvdpspn): Likewise.
12531 (vsx_xscvspdpn): Likewise.
12532 (vsx_xscvdpspn_scalar): Likewise.
12533 (vsx_xscvspdpn_directmove): Likewise.
12534 (vsx_and<mode>3): Split logical operations into 32-bit and
12535 64-bit. Add support to do logical operations on TImode as well as
12536 VSX vector types. Allow logical operations to be done in either
12537 VSX registers or in general purpose registers in 64-bit mode. Add
12538 splitters if GPRs were used. For AND, add clobber of CCmode to
12539 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
12541 (vsx_and<mode>3_32bit): Likewise.
12542 (vsx_and<mode>3_64bit): Likewise.
12543 (vsx_ior<mode>3): Likewise.
12544 (vsx_ior<mode>3_32bit): Likewise.
12545 (vsx_ior<mode>3_64bit): Likewise.
12546 (vsx_xor<mode>3): Likewise.
12547 (vsx_xor<mode>3_32bit): Likewise.
12548 (vsx_xor<mode>3_64bit): Likewise.
12549 (vsx_one_cmpl<mode>2): Likewise.
12550 (vsx_one_cmpl<mode>2_32bit): Likewise.
12551 (vsx_one_cmpl<mode>2_64bit): Likewise.
12552 (vsx_nor<mode>3): Likewise.
12553 (vsx_nor<mode>3_32bit): Likewise.
12554 (vsx_nor<mode>3_64bit): Likewise.
12555 (vsx_andc<mode>3): Likewise.
12556 (vsx_andc<mode>3_32bit): Likewise.
12557 (vsx_andc<mode>3_64bit): Likewise.
12558 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
12559 and xxlorc instructions.
12560 (vsx_eqv<mode>3_64bit): Likewise.
12561 (vsx_nand<mode>3_32bit): Likewise.
12562 (vsx_nand<mode>3_64bit): Likewise.
12563 (vsx_orc<mode>3_32bit): Likewise.
12564 (vsx_orc<mode>3_64bit): Likewise.
12566 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
12568 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
12570 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
12571 (p8_vmrgow): Likewise.
12572 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
12573 GPRs to be split under VSX.
12574 (p8v_clz<mode>2): Add power8 count leading zero support.
12575 (p8v_popcount<mode>2): Add power8 population count support.
12576 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
12579 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
12582 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
12584 (vec_nand): Likewise.
12585 (vec_vclz): Likewise.
12586 (vec_vclzb): Likewise.
12587 (vec_vclzd): Likewise.
12588 (vec_vclzh): Likewise.
12589 (vec_vclzw): Likewise.
12590 (vec_vgbbd): Likewise.
12591 (vec_vmrgew): Likewise.
12592 (vec_vmrgow): Likewise.
12593 (vec_vpopcnt): Likewise.
12594 (vec_vpopcntb): Likewise.
12595 (vec_vpopcntd): Likewise.
12596 (vec_vpopcnth): Likewise.
12597 (vec_vpopcntw): Likewise.
12599 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
12601 PR rtl-optimization/57468
12602 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
12605 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
12607 PR rtl-optimization/57459
12608 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
12609 type when setting live regs.
12611 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
12613 * config/s390/s390.opt (mlra): New option.
12614 * config/s390/s390.c (s390_decompose_address): Check displacement
12615 for all registers for LRA.
12616 (s390_secondary_reload): Don't used secondary reloads for LRA.
12617 (s390_lra_p): New function.
12618 (TARGET_LRA_P): Define.
12619 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
12620 of attribute cpu_facility to zarch for the last alternative.
12621 (*cmpmem_short): Ditto.
12623 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
12625 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
12626 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
12627 (arm_expand_prologue): Likewise.
12629 2013-06-06 Teresa Johnson <tejohnson@google.com>
12632 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
12633 as this is now done by redirect_edge_and_branch_force.
12634 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
12635 barriers, and fix interaction with splitting.
12636 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
12637 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
12638 reflect changes made in the routine.
12639 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
12640 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
12641 since this is called in cfglayout mode, and replace partition fixup
12642 with assert as that is now done by force_nonfallthru_and_redirect.
12643 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
12644 already be marked with region crossing note.
12645 (insert_section_boundary_note): Make non-static, gate on flag
12646 has_bb_partition, rewrite to also check for multiple partitions.
12647 (rest_of_handle_reorder_blocks): Remove call to
12648 insert_section_boundary_note, now done later during free_cfg.
12649 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
12650 * bb-reorder.h (insert_section_boundary_note): Declare.
12651 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
12652 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
12653 invoke insert_section_boundary_note.
12654 (try_redirect_by_replacing_jump): Remove unnecessary
12655 check for region crossing note.
12656 (fixup_partition_crossing): New function.
12657 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
12658 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
12659 in non-cfglayout mode.
12660 (force_nonfallthru_and_redirect): Fixup partition boundaries,
12661 remove old code that tried to do this. Emit barrier correctly
12662 when we are in cfglayout mode.
12663 (last_bb_in_partition): New function.
12664 (rtl_split_edge): Correctly fixup partition boundaries.
12665 (commit_one_edge_insertion): Remove old code that tried to
12666 fixup region crossing edge since this is now handled in
12667 split_block, and set up insertion point correctly since
12668 block may now end in a jump.
12669 (verify_hot_cold_block_grouping): Guard against checking when not in
12670 linearized RTL mode.
12671 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
12673 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
12674 rtl_verify_flow_info, so not called in cfglayout mode.
12675 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
12676 (fixup_reorder_chain): Remove old code that attempted to fixup region
12677 crossing note as this is now handled in force_nonfallthru_and_redirect.
12678 (duplicate_insn_chain): Don't duplicate switch section notes.
12679 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
12681 * basic-block.h (emit_barrier_after_bb): Declare.
12683 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12685 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
12686 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
12687 arm_usatsihi): Adjust alternatives for arm_restrict_it.
12689 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12691 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
12693 * config/arm/ldmstm.md: Regenerate.
12695 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12697 * config/arm/sync.md (atomic_loaddi_1):
12698 Disable predication for arm_restrict_it.
12699 (arm_load_exclusive<mode>): Likewise.
12700 (arm_load_exclusivesi): Likewise.
12701 (arm_load_exclusivedi): Likewise.
12702 (arm_load_acquire_exclusive<mode>): Likewise.
12703 (arm_load_acquire_exclusivesi): Likewise.
12704 (arm_load_acquire_exclusivedi): Likewise.
12705 (arm_store_exclusive<mode>): Likewise.
12706 (arm_store_exclusive<mode>): Likewise.
12707 (arm_store_release_exclusivedi): Likewise.
12708 (arm_store_release_exclusive<mode>): Likewise.
12710 2013-06-06 Richard Biener <rguenther@suse.de>
12712 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
12714 (lto_tag_is_tree_code_p): Adjust.
12715 (lto_tag_is_gimple_code_p): Likewise.
12716 (lto_gimple_code_to_tag): Likewise.
12717 (lto_tag_to_gimple_code): Likewise.
12718 (lto_tree_code_to_tag): Likewise.
12719 (lto_tag_to_tree_code): Likewise.
12720 * data-streamer.h (streamer_write_hwi_in_range): Use
12721 uhwi streaming to stream the normalized range.
12722 (streamer_read_hwi_in_range): Likewise.
12724 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12726 * config/arm/arm.md (enabled_for_depr_it): New attribute.
12727 (predicable_short_it): Likewise.
12728 (predicated): Likewise.
12729 (enabled): Handle above.
12730 (define_cond_exec): Set predicated attribute to yes.
12732 2013-06-05 Mike Stump <mikestump@comcast.net>
12734 * gdbinit.in (__FUNCTION__): Add.
12736 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
12738 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
12739 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
12741 2013-06-05 Jan Hubicka <jh@suse.cz>
12743 * varasm.c (mark_decl_referenced): Revert the removal until targets
12746 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
12748 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
12749 instead of mark_decl_referenced.
12751 2013-06-05 Jan Hubicka <jh@suse.cz>
12753 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
12754 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
12755 and symtab_used_from_object_file_p.
12756 (cgraph_make_node_local_1): Clear forced_by_abi.
12757 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
12758 * cgraph.h (symtab_node_base): Add forced_by_abi.
12759 (decide_is_variable_needed): Remove.
12760 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
12761 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
12762 (decide_is_symbol_needed): ... this one; handle symbols in general;
12763 always analyze virtuals; honnor forced_by_abi.
12764 (cgraph_finalize_function): Update.
12765 (varpool_finalize_decl): Update.
12766 (symbol_defined_and_needed): Remove.
12767 (analyze_functions): Update.
12768 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
12769 output_refs, input_overwrite_node): Handle forced_by_abi.
12770 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
12771 (address_taken_from_non_vtable_p): ... this one.
12772 (comdat_can_be_unshared_p_1): New function.
12773 (cgraph_comdat_can_be_unshared_p): Rename to ...
12774 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
12775 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
12776 (function_and_variable_visibility): Clear forced_by_abi as needed.
12777 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
12778 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
12779 * symtab.c (dump_symtab_base): Dump forced_by_abi.
12780 * varpool.c (decide_is_variable_needed): Remove.
12782 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12784 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
12785 (arm_option_override): Override arm_restrict_it where appropriate.
12786 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
12787 * config/arm/arm.opt (mrestrict-it): New command-line option.
12788 * doc/invoke.texi: Document -mrestrict-it.
12790 2013-06-05 David Malcolm <dmalcolm@redhat.com>
12792 * tsan.c (tsan_atomic_table): Make const.
12794 2013-06-05 Richard Biener <rguenther@suse.de>
12796 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
12797 index associated with the tree we are supposed to replace.
12798 * tree-streamer-out.c (pack_ts_base_value_fields): Output
12799 TREE_ASM_WRITTEN as zero for everything but SSA names.
12801 2013-06-05 David Malcolm <dmalcolm@redhat.com>
12803 * tree-ssa-structalias.c (call_stmt_vars): Make static.
12805 2013-06-04 Jan Hubicka <jh@suse.cz>
12807 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
12808 (input_node, input_varpool_node): Handle correctly external same
12810 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
12811 nodes at ltrans stage.
12813 2013-06-04 Jan Hubicka <jh@suse.cz>
12815 * ipa-inline.c (update_caller_keys): Fix availability test.
12816 (update_callee_keys): Likewise.
12817 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
12818 to follow ELF standard.
12820 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
12822 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
12823 (mips64r5900el-*-elf*): New configurations.
12824 * config/mips/mips-cpus.def (r5900): New processor.
12825 * config/mips/mips-tables.opt: Regenerate.
12826 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
12827 (mips_issue_rate): Handle PROCESSOR_R5900.
12828 (mips_reorg_process_insns): Force reorder mode for the R5900.
12829 * config/mips/mips.h (TARGET_MIPS5900): Define.
12830 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
12832 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
12834 * config/mips/mips.md (processor): Add r5900.
12835 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
12837 2013-06-04 Ian Bolton <ian.bolton@arm.com>
12839 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
12840 into function to generate MOVI instruction.
12841 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
12842 (aarch64_preferred_simd_mode): Turn into wrapper.
12843 (aarch64_output_scalar_simd_mov_immediate): New function.
12844 * config/aarch64/aarch64-protos.h: Add prototype for above.
12846 2013-06-04 Ian Bolton <ian.bolton@arm.com>
12848 * config/aarch64/aarch64.c (simd_immediate_info): Remove
12849 element_char member.
12850 (sizetochar): Return signed char.
12851 (aarch64_simd_valid_immediate): Remove elchar and other
12852 unnecessary variables.
12853 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
12854 Calculate element_char as required.
12855 * config/aarch64/aarch64-protos.h: Update and move prototype
12856 for aarch64_output_simd_mov_immediate.
12857 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
12860 2013-06-04 Ian Bolton <ian.bolton@arm.com>
12862 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
12863 information completed by aarch64_simd_valid_immediate.
12864 (aarch64_legitimate_constant_p): Update arguments.
12865 (aarch64_simd_valid_immediate): Work with struct rather than many
12867 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
12868 (aarch64_simd_make_constant): Update arguments.
12869 (aarch64_output_simd_mov_immediate): Work with struct rather than
12870 many pointers. Output immediate directly rather than as operand.
12871 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
12873 * config/aarch64/constraints.md (Dn): Update arguments.
12875 2013-06-04 Ian Bolton <ian.bolton@arm.com>
12877 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
12879 (aarch64_simd_immediate_valid_for_move): Remove.
12880 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
12881 (aarch64_simd_make_constant): Update call.
12882 (aarch64_output_simd_mov_immediate): Update call.
12883 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
12885 * config/aarch64/constraints.md (Dn): Update call.
12887 2013-06-04 Ian Bolton <ian.bolton@arm.com>
12889 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
12890 return type to bool for prototype.
12891 (aarch64_legitimate_constant_p): Check for true instead of not -1.
12892 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
12893 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
12895 2013-06-04 Catherine Moore <clm@codesourcery.com>
12897 * config/mips/mips.opt (meva): New.
12898 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
12899 (ASM_SPEC): Handle -meva.
12900 * doc/invoke.texi (meva): Document.
12902 2013-06-04 Alan Modra <amodra@gmail.com>
12904 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
12907 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12909 * rtl.def: Add extra fourth optional field to define_cond_exec.
12910 * gensupport.c (process_one_cond_exec): Process attributes from
12912 * doc/md.texi: Document fourth field in define_cond_exec.
12914 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
12916 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
12917 out the processing order as in store_bit_field_1.
12919 2013-06-04 Jan Hubicka <jh@suse.cz>
12921 PR middle-end/57500
12922 * cgraphunit.c (cgraph_process_same_body_aliases): Create
12923 non-VAR_DECL node if it does not exist yet.
12925 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
12927 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
12928 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
12929 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
12930 target_cpu_default setting.
12932 2013-06-03 Teresa Johnson <tejohnson@google.com>
12934 * dumpfile.c (opt_info_switch_p): Change -fopt-info
12935 default to -fopt-info=optimized instead of all.
12936 * doc/invoke.texi: Ditto.
12937 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
12938 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
12939 (execute_vect_slp): Emit BB vectorization success under
12940 MSG_OPTIMIZED_LOCATIONS.
12941 * tree-vect-slp.c (vect_slp_transform_bb): Change
12942 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
12943 * tree-vect-loop.c (vect_transform_loop): Ditto.
12945 2013-06-03 Jason Merrill <jason@redhat.com>
12948 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
12949 Use TARGET_EXPR for C++.
12951 2013-06-03 Jakub Jelinek <jakub@redhat.com>
12953 PR rtl-optimization/57268
12954 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
12955 if DEBUG_INSN_P (insn).
12958 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
12960 PR rtl-optimization/57268
12961 * sched-deps.c (sched_analyze_2): Flush dependence lists if
12962 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
12964 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
12966 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
12967 (ix86_avoid_lea_for_addr): Likewise.
12968 (exact_dependency_1): Likewise.
12969 (ix86_adjust_cost): Likewise.
12970 (swap_top_of_ready_list): Fix formatting and !reload_completed check
12972 (do_reorder_for_imul): Fix typo, formatting and
12973 !reload_completed check removed.
12974 (ix86_sched_reorder): Fix typo and formatting.
12975 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
12978 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
12980 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
12982 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
12984 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
12985 <STRING_CST>: Likewise.
12986 <VECTOR_CST>: Likewise.
12988 2013-06-01 Janus Weil <janus@gcc.gnu.org>
12989 Mikael Morin <mikael@gcc.gnu.org>
12991 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
12992 * config.in: Regenerated.
12993 * configure: Regenerated.
12995 2013-06-01 Jan Hubicka <jh@suse.cz>
12997 PR middle-end/57366
12998 * cgraphunit.c (compile): When weakref is not supported,
12999 set up transparent aliases before final output pass.
13000 * varasm.c (assemble_alias): Do not try to do it here.
13002 2013-06-01 Jan Hubicka <jh@suse.cz>
13004 PR middle-end/57467
13005 * passes.c (for_per_function): Skip unanalyzed functions.
13007 2013-06-01 Jan Hubicka <jh@suse.cz>
13009 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
13010 (lto_symtab_merge_symbols_1): ... this one.
13011 (lto_symtab_merge_cgraph_nodes): Rename to ...
13012 (lto_symtab_merge_symbols): ... this one; simplify.
13013 * cgraph.c (same_body_aliases_done): Rename to ...
13014 (cpp_implicit_aliases_done): ... this one.
13015 (cgraph_create_function_alias): Update.
13016 (cgraph_same_body_alias): Update.
13017 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
13018 (verify_edge_corresponds_to_fndecl): Simplify.
13019 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
13020 (cgraph_node): Remove same_body_alias.
13021 (varpool_node): Remove alias_of and extra_name_alias.
13022 (same_body_aliases_done): Rename to ..
13023 (cpp_implicit_aliases_done): ... this one.
13024 (symtab_alias_ultimate_target): Add default parameter.
13025 (symtab_resolve_alias): New function.
13026 (fixup_same_cpp_alias_visibility): Declare.
13027 (cgraph_function_node): Add default parameter.
13028 (cgraph_node_asm_name): Likewise.
13029 (cgraph_function_or_thunk_node): Add default parameter; do
13030 not ICE when it is NULL.
13031 (varpool_variable_node): Likewise.
13032 * tree-emutls.c (create_emultls_var): Update.
13033 (ipa_lower_emutls): Update.
13034 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
13035 (cgraph_reset_node): Reset alias info.
13036 (cgraph_finalize_function): Update.
13037 (fixup_same_cpp_alias_visibility): Move to symtab.c.
13038 (analyze_function): Simplify.
13039 (cgraph_process_same_body_aliases): Simplify.
13040 (analyze_functions): Fixup same body aliases.
13041 (handle_alias_pairs): Simplify.
13042 (assemble_thunk): Update.
13043 (assemble_thunks_and_aliases): Update.
13044 (output_weakrefs): Rewrite.
13045 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
13046 (lto_output_varpool_node): Likewise.
13047 (compute_ltrans_boundary): Remve assert.
13048 (get_alias_symbol): New functoin.
13049 (input_node): Rewrite alias handling.
13050 (input_varpool_node): Likewise.
13051 * ipa-pure-const.c (propagate_pure_const): Fix formating.
13052 * ipa.c (process_references): Handle weakrefs correctly.
13053 (symtab_remove_unreachable_nodes): Likewise.
13054 * trans-mem.c (get_cg_data): Update.
13055 (ipa_tm_create_version_alias): Update.
13056 (ipa_tm_execute): Update.
13057 * symtab.c (dump_symtab_base): Dump aliases.
13058 (verify_symtab_base): Verify aliases.
13059 (symtab_node_availability): New function.
13060 (symtab_alias_ultimate_target): Simplify.
13061 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
13062 handle all the fixup cases.
13063 (symtab_resolve_alias): New function.
13064 * passes.c (ipa_write_summaries): Handle weakrefs.
13065 * varpool.c (varpool_analyze_node): Simplify.
13066 (assemble_aliases): Update.
13067 (varpool_create_variable_alias): Simplify.
13068 (varpool_extra_name_alias): Simplify.
13069 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
13070 (lto_symtab_merge_symbols): ... this one.
13072 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
13075 PR rtl-optimization/57268
13076 * sched-deps.c (sched_analyze_2): Flush dependence lists if
13077 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
13079 2013-06-01 Tobias Burnus <burnus@net-b.de>
13081 Partially reverted:
13082 2013-05-31 Tobias Burnus <burnus@net-b.de>
13084 PR middle-end/57073
13085 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
13088 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
13090 PR rtl-optimization/57268
13091 * sched-deps.c (sched_analyze_2): Flush dependence lists if
13092 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
13094 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
13096 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
13097 unordered comparison operators when -fno-trapping-math is in effect
13099 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
13100 and implement unordered comparison operators properly on the e500.
13102 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
13104 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
13105 for constant scalar integers.
13106 (simplify_relational_operation_1): Likewise.
13108 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
13110 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
13111 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
13114 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
13115 Igor Zamyatin <igor.zamyatin@intel.com>
13117 Silvermont (SLM) architecture performance tuning.
13118 * config/i386/i386.h (enum ix86_tune_indices): Add
13119 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
13120 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
13122 * config/i386/i386.c (initial_ix86_tune_features)
13123 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
13124 (ix86_lea_outperforms): Handle Silvermont tuning.
13125 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
13127 (ix86_use_lea_for_mov): Likewise.
13128 (ix86_avoid_lea_for_addr): Likewise.
13129 (ix86_lea_for_add_ok): Likewise.
13130 (exact_dependency_1): New function.
13131 (exact_store_load_dependency): Likewise.
13132 (ix86_adjust_cost): Handle Silvermont tuning.
13133 (do_reoder_for_imul): Likewise.
13134 (swap_top_of_ready_list): New function.
13135 (ix86_sched_reorder): Changed to handle Silvermont tuning.
13137 * config/i386/i386.md (peepholes that split memory operand in fp
13140 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
13142 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
13143 Remove un-necessary braces.
13145 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
13147 * config/aarch64/aarch64.c (aarch64_classify_symbol):
13148 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
13150 2013-05-31 Tobias Burnus <burnus@net-b.de>
13152 PR middle-end/57073
13153 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
13155 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13158 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
13159 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
13160 * config/arm/neon.md (iordi3_neon): Remove.
13161 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
13162 * config/arm/predicates.md (imm_for_neon_logic_operand):
13163 Move to earlier in the file.
13164 (neon_logic_op2): Likewise.
13165 (arm_iordi_operand_neon): New predicate.
13167 2013-05-31 Richard Biener <rguenther@suse.de>
13169 PR tree-optimization/57478
13170 PR tree-optimization/57453
13171 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
13174 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
13176 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
13177 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
13179 2013-05-30 Tobias Burnus <burnus@net-b.de>
13180 Thomas Koenig <tkoenig@gcc.gnu.org>
13182 PR middle-end/57073
13183 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
13184 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
13186 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
13188 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
13190 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
13192 * target.def (register_usage_leveling_p): New hook.
13193 * targhooks.c (default_register_usage_leveling_p): New.
13194 * targhooks.h (default_register_usage_leveling_p): New prototype.
13195 * lra-assigns.c (register_usage_leveling_p): Use the hook.
13196 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
13197 * doc/tm.texi: Update.
13198 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
13200 2013-05-30 Ian Bolton <ian.bolton@arm.com>
13202 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
13203 (*insv_reg<mode>): New define_insn.
13205 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
13207 PR rtl-optimization/57439
13208 * postreload.c (move2add_valid_value_p): Check that we have
13209 a zero subreg_regno_offset when accessing the register in
13210 the requested mode.
13212 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
13213 Igor Zamyatin <igor.zamyatin@intel.com>
13215 Silvermont (SLM) architecture pipeline model, tuning and
13217 * config.gcc: Add slm config options and target.
13219 * config/i386/slm.md: New.
13221 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
13223 * config/i386/i386-c.c (ix86_target_macros_internal): New case
13225 (ix86_target_macros_internal): Likewise.
13227 * config/i386/i386.c (slm_cost): New cost.
13228 (m_SLM): New macro flag.
13229 (initial_ix86_tune_features): Set m_SLM.
13230 (x86_accumulate_outgoing_args): Likewise.
13231 (x86_arch_always_fancy_math_387): Likewise.
13232 (processor_target_table): Add slm cost.
13233 (cpu_names): Add slm cpu name.
13234 (x86_option_override_internal): Set SLM ISA.
13235 (ix86_issue_rate): New case PROCESSOR_SLM.
13236 (ia32_multipass_dfa_lookahead): Likewise.
13237 (fold_builtin_cpu): Add slm.
13239 * config/i386/i386.h (TARGET_SLM): New target macro.
13240 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
13241 (processor_type): Add PROCESSOR_SLM.
13243 * config/i386/i386.md (cpu): Add new value "slm".
13244 (slm.md): Include slm.md.
13246 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
13247 Zhenqiang Chen <zhenqiang.chen@linaro.org>
13249 * config/arm/arm-protos.h: Add and update function protos.
13250 * config/arm/arm.c (use_simple_return_p): New added.
13251 (thumb2_expand_return): Check simple_return flag.
13252 * config/arm/arm.md: Add simple_return and conditional simple_return.
13253 * config/arm/iterators.md: Add iterator for return and simple_return.
13255 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
13257 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
13258 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
13259 (arm_emit_vfp_multi_reg_pop): Likewise.
13260 (thumb2_emit_ldrd_pop): Likewise.
13261 (arm_expand_epilogue): Add misc REG_CFA notes.
13262 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
13264 2013-05-29 Lawrence Crowl <crowl@google.com>
13266 * config/arm/t-arm: Update for below.
13268 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
13269 Change type to hash_table. Update dependent calls and types.
13271 * config/i386/t-cygming: Update for below.
13273 * config/i386/t-interix: Update for below.
13275 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
13276 Change type to hash_table. Update dependent calls and types.
13277 (i386_find_on_wrapper_list::wrappers): Likewise.
13279 * config/ia64/t-ia64: Update for below.
13281 * config/ia64/ia64.c (bundle_state_table):
13282 Change type to hash_table. Update dependent calls and types.
13284 * config/mips/mips.c (mips_reorg_process_insns::htab):
13285 Change type to hash_table. Update dependent calls and types.
13287 * config/sol2.c (solaris_comdat_htab):
13288 Change type to hash_table. Update dependent calls and types.
13290 * config/t-sol2: Update for above.
13292 2013-05-29 Teresa Johnson <tejohnson@google.com>
13294 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
13295 functions are not yet marked as defined.
13297 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
13298 Pat Haugen <pthaugen@us.ibm.com>
13299 Peter Bergner <bergner@vnet.ibm.com>
13301 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
13305 (vrotl<mode>3): Likewise.
13306 (vashl<mode>3): Likewise.
13307 (vlshr<mode>3): Likewise.
13308 (vashr<mode>3): Likewise.
13310 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
13311 support for power8 V2DI builtins.
13313 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
13314 power8 V2DI builtins.
13315 (vupkhsw): Likewise.
13316 (vupklsw): Likewise.
13317 (vaddudm): Likewise.
13318 (vminsd): Likewise.
13319 (vmaxsd): Likewise.
13320 (vminud): Likewise.
13321 (vmaxud): Likewise.
13322 (vpkudum): Likewise.
13323 (vpksdss): Likewise.
13324 (vpkudus): Likewise.
13325 (vpksdus): Likewise.
13330 (vsubudm): Likewise.
13331 (vcmpequd): Likewise.
13332 (vcmpgtsd): Likewise.
13333 (vcmpgtud): Likewise.
13334 (vcmpequd_p): Likewise.
13335 (vcmpgtsd_p): Likewise.
13336 (vcmpgtud_p): Likewise.
13337 (vupkhsw): Likewise.
13338 (vupklsw): Likewise.
13339 (vaddudm): Likewise.
13340 (vmaxsd): Likewise.
13341 (vmaxud): Likewise.
13342 (vminsd): Likewise.
13343 (vminud): Likewise.
13344 (vpksdss): Likewise.
13345 (vpksdus): Likewise.
13346 (vpkudum): Likewise.
13347 (vpkudus): Likewise.
13352 (vsubudm): Likewise.
13354 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
13355 support for power8 V2DI instructions.
13357 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
13358 power8 V2DI instructions. Combine pack and unpack insns to use an
13359 iterator for each mode. Check whether a particular mode supports
13360 Altivec instructions instead of just checking TARGET_ALTIVEC.
13361 (UNSPEC_VPKUWUM): Likewise.
13362 (UNSPEC_VPKSHSS): Likewise.
13363 (UNSPEC_VPKSWSS): Likewise.
13364 (UNSPEC_VPKUHUS): Likewise.
13365 (UNSPEC_VPKSHUS): Likewise.
13366 (UNSPEC_VPKUWUS): Likewise.
13367 (UNSPEC_VPKSWUS): Likewise.
13368 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
13369 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
13370 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
13371 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
13372 (UNSPEC_VUPKHSB): Likewise.
13373 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
13374 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
13375 (UNSPEC_VUPKHSH): Likewise.
13376 (UNSPEC_VUPKLSB): Likewise.
13377 (UNSPEC_VUPKLSH): Likewise.
13379 (VI_char): Likewise.
13380 (VI_scalar): Likewise.
13381 (VI_unit): Likewise.
13383 (VP_small): Likewise.
13384 (VP_small_lc): Likewise.
13385 (VU_char): Likewise.
13386 (add<mode>3): Likewise.
13387 (altivec_vaddcuw): Likewise.
13388 (altivec_vaddu<VI_char>s): Likewise.
13389 (altivec_vadds<VI_char>s): Likewise.
13390 (sub<mode>3): Likewise.
13391 (altivec_vsubcuw): Likewise.
13392 (altivec_vsubu<VI_char>s): Likewise.
13393 (altivec_vsubs<VI_char>s): Likewise.
13394 (altivec_vavgs<VI_char>): Likewise.
13395 (altivec_vcmpbfp): Likewise.
13396 (altivec_eq<mode>): Likewise.
13397 (altivec_gt<mode>): Likewise.
13398 (altivec_gtu<mode>): Likewise.
13399 (umax<mode>3): Likewise.
13400 (smax<mode>3): Likewise.
13401 (umin<mode>3): Likewise.
13402 (smin<mode>3): Likewise.
13403 (altivec_vpkuhum): Likewise.
13404 (altivec_vpkuwum): Likewise.
13405 (altivec_vpkshss): Likewise.
13406 (altivec_vpkswss): Likewise.
13407 (altivec_vpkuhus): Likewise.
13408 (altivec_vpkshus): Likewise.
13409 (altivec_vpkuwus): Likewise.
13410 (altivec_vpkswus): Likewise.
13411 (altivec_vpks<VI_char>ss): Likewise.
13412 (altivec_vpks<VI_char>us): Likewise.
13413 (altivec_vpku<VI_char>us): Likewise.
13414 (altivec_vpku<VI_char>um): Likewise.
13415 (altivec_vrl<VI_char>): Likewise.
13416 (altivec_vsl<VI_char>): Likewise.
13417 (altivec_vsr<VI_char>): Likewise.
13418 (altivec_vsra<VI_char>): Likewise.
13419 (altivec_vsldoi_<mode>): Likewise.
13420 (altivec_vupkhsb): Likewise.
13421 (altivec_vupkhs<VU_char>): Likewise.
13422 (altivec_vupkls<VU_char>): Likewise.
13423 (altivec_vupkhsh): Likewise.
13424 (altivec_vupklsb): Likewise.
13425 (altivec_vupklsh): Likewise.
13426 (altivec_vcmpequ<VI_char>_p): Likewise.
13427 (altivec_vcmpgts<VI_char>_p): Likewise.
13428 (altivec_vcmpgtu<VI_char>_p): Likewise.
13429 (abs<mode>2): Likewise.
13430 (vec_unpacks_hi_v16qi): Likewise.
13431 (vec_unpacks_hi_v8hi): Likewise.
13432 (vec_unpacks_lo_v16qi): Likewise.
13433 (vec_unpacks_hi_<VP_small_lc>): Likewise.
13434 (vec_unpacks_lo_v8hi): Likewise.
13435 (vec_unpacks_lo_<VP_small_lc>): Likewise.
13436 (vec_pack_trunc_v8h): Likewise.
13437 (vec_pack_trunc_v4si): Likewise.
13438 (vec_pack_trunc_<mode>): Likewise.
13440 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
13442 (vec_vmaxsd): Likewise.
13443 (vec_vmaxud): Likewise.
13444 (vec_vminsd): Likewise.
13445 (vec_vminud): Likewise.
13446 (vec_vpksdss): Likewise.
13447 (vec_vpksdus): Likewise.
13448 (vec_vpkudum): Likewise.
13449 (vec_vpkudus): Likewise.
13450 (vec_vrld): Likewise.
13451 (vec_vsld): Likewise.
13452 (vec_vsrad): Likewise.
13453 (vec_vsrd): Likewise.
13454 (vec_vsubudm): Likewise.
13455 (vec_vupkhsw): Likewise.
13456 (vec_vupklsw): Likewise.
13458 2013-05-29 Jan Hubicka <jh@suse.cz>
13460 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
13461 flags; reorder rest of fields in more consistent way.
13462 (varpool_node): Remove analyzed, finalized and alias.
13463 (cgraph_ndoe): Likewise.
13464 (symtab_alias_ultimate_target): New function.
13465 (cgraph_function_node): Move offline.
13466 (cgraph_reset_node): Declare.
13467 (cgraph_comdat_can_be_unshared_p): Remove.
13468 (varpool_remove_initializer): Declare.
13469 (varpool_first_defined_variable, varpool_next_defined_variable
13470 cgraph_first_defined_function, cgraph_next_defined_function): Update.
13471 (cgraph_function_with_gimple_body_p): Update.
13472 (varpool_all_refs_explicit_p): Update.
13473 (symtab_alias_target): New function.
13474 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
13475 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
13476 (cgraph_function_or_thunk_node): Simplify using
13477 symtab_alias_ultimate_target.
13478 (varpool_variable_node): Likewise.
13479 * cgraph.c (cgraph_create_function_alias): Update.
13480 (cgraph_add_thunk): Update.
13481 (cgraph_remove_node): Update.
13482 (dump_cgraph_node): Do not dump removed flags.
13483 (cgraph_function_body_availability): Update.
13484 (cgraph_propagate_frequency): Update.
13485 (verify_cgraph_node): Check sanity of local flag.
13486 (cgraph_function_node): Move here from cgraph.h; revamp for
13487 cgraph_function_or_thunk_node.
13488 * lto-symtab.c (lto_varpool_replace_node): Update.
13489 (lto_symtab_resolve_can_prevail_p): Update.
13490 (lto_symtab_merge_cgraph_nodes): Update.
13491 * ipa-cp.c (determine_versionability, initialize_node_lattices,
13492 propagate_constants_accross_call, devirtualization_time_bonus,
13493 ipcp_propagate_stage): Update.
13494 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
13495 * ipa-inline-transform.c (clone_inlined_nodes,
13496 preserve_function_body_p): Update.
13497 * ipa-reference.c (propagate): Update.
13498 (write_node_summary_p): Update.
13499 * toplev.c (wrapup_global_declaration_2): Update.
13500 * cgraphunit.c (cgraph_analyze_function): Rename to ...
13501 (analyze_function) ... this one.
13502 (cgraph_process_new_functions): Update.
13503 (cgraph_reset_node): Export.
13504 (cgraph_finalize_function): Update.
13505 (cgraph_add_new_function): Update.
13506 (process_function_and_variable_attributes): Update.
13507 (varpool_finalize_decl): Update.
13508 (symbol_finalized): Remove.
13509 (symbol_finalized_and_needed): Rename to ...
13510 (symbol_defined_and_needed): ... update.
13511 (cgraph_analyze_functions): Update.
13512 (handle_alias_pairs): Update.
13513 (mark_functions_to_output): Update.
13514 (assemble_thunk): Update.
13515 (output_in_order): Update.
13516 (output_weakrefs): Update.
13517 (finalize_compilation_unit): Update.
13518 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
13519 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
13520 input_node, input_varpool_node): Update.
13521 * dbxout.c (dbxout_expand_expr): Update.
13522 * cgraphclones.c (cgraph_clone_node): Update.
13523 (cgraph_copy_node_for_versioning): Update.
13524 (cgraph_materialize_clone): Update.
13525 (cgraph_materialize_all_clones): Update.
13526 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
13527 propagate_pure_const, propagate_nothrow): Update.
13528 * lto-streamer-out.c (lto_output, write_symbol): Update.
13529 * ipa-utils.c (ipa_reverse_postorder): Update.
13530 * ipa-inline.c (can_inline_edge_p): Update.
13531 (update_caller_keys, ipa_inline): Update.
13532 * dwarf2out.c (reference_to_unused,
13533 premark_types_used_by_global_vars_helper): Update.
13534 * tree-eh.c (tree_could_trap_p): Update.
13535 * ipa-split.c (consider_split, execute_split_functions): Update.
13536 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
13537 has_addr_references_p): Update; move ahead in file for better
13539 (process_references): Simplify.
13540 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
13541 bodies are removed.
13542 (cgraph_comdat_can_be_unshared_p): Make static.
13543 (cgraph_externally_visible_p): Update.
13544 (varpool_externally_visible_p): Update.
13545 (function_and_variable_visibility): Update.
13546 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
13547 ipa_tm_mark_force_output_node): Update.
13548 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
13549 estimate_edge_devirt_benefit, inline_generate_summary,
13550 inline_write_summary): Update.
13551 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
13552 * ipa-prop.c (ipa_compute_jump_functions): Update.
13553 (ipa_print_node_params, ipa_prop_read_section,
13554 ipa_update_after_lto_read, read_replacements_section): Update.
13555 * varasm.c (mark_decl_referenced): Update.
13556 (assemble_alias, dump_tm_clone_pairs): Update.
13557 * tree-inline.c (copy_bb): Update.
13558 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
13560 * symtab.c (dump_symtab_base): Print new flags.
13561 (verify_symtab_base): Verify new flags.
13562 (symtab_alias_ultimate_target): New function.
13563 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
13564 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
13566 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
13568 * i386.c (ix86_get_function_versions_dispatcher,
13569 ix86_generate_version_dispatcher_body): Update.
13570 (fold_builtin_cpu): Use varpool_add_new_variable.
13571 * varpool.c (varpool_remove_initializer): Break out from ...
13572 (varpool_remove_node): ... this one.
13573 (dump_varpool_node, varpool_node_for_asm,
13574 cgraph_variable_initializer_availability, varpool_analyze_node,
13575 varpool_assemble_decl, varpool_remove_unreferenced_decls,
13576 varpool_finalize_named_section_flags, varpool_create_variable_alias):
13579 2013-05-29 Jan Hubicka <jh@suse.cz>
13581 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
13583 2013-05-29 Easwaran Raman <eraman@google.com>
13585 PR tree-optimization/57442
13586 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
13587 when control exits the main loop.
13589 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
13591 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
13593 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
13594 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
13595 * rx/t-rx: Add rx100 under multi library matches option for nofpu
13598 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13600 PR tree-optimization/57441
13601 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
13602 Don't limit size of incr_vec to number of candidates.
13604 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
13606 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
13607 and mips16 directories.
13608 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
13609 (MULTILIB_DIRNAMES): Ditto.
13610 (MULTILIB_EXCEPTIONS): Add new exceptions.
13611 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
13612 (MULTILIB_DIRNAMES): Ditto.
13613 (MULTILIB_EXCEPTIONS): Add new exceptions.
13615 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
13616 Marcus Shawcroft <marcus.shawcroft@arm.com>
13618 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
13619 SYMBOL_TINY_ABSOLUTE.
13620 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
13621 SYMBOL_TINY_ABSOLUTE.
13622 (aarch64_expand_mov_immediate): Likewise.
13623 (aarch64_classify_symbol): Likewise.
13624 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
13625 Permit SYMBOL_TINY_ABSOLUTE.
13626 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
13628 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
13629 Marcus Shawcroft <marcus.shawcroft@arm.com>
13631 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
13632 Refactor if/switch. Replace gcc_assert with if.
13634 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
13636 * config/i386/i386.c (initial_ix86_tune_features): Enable
13637 FP Reassociation for AMD bdver1 and bdver2.
13639 2013-05-29 Martin Jambor <mjambor@suse.cz>
13641 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
13642 and IMAGPART_EXPR do not occur within other handled_components.
13644 2013-05-29 Richard Biener <rguenther@suse.de>
13646 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
13647 access on whether the use is in the BB we currently try to
13649 (vect_bb_vectorization_profitable_p): Pass the BB we currently
13650 vectorize to vect_bb_slp_scalar_cost.
13652 2013-05-29 Richard Biener <rguenther@suse.de>
13654 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
13655 computing scalar cost offsetted by stmts that are kept live
13657 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
13658 for computation of scalar cost.
13660 2013-05-28 Steve Ellcey <sellcey@mips.com>
13662 * config/mips/mips-cpus.def (mips32r2): Change processor type.
13664 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
13666 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
13667 array notation built-in reduction functions.
13668 * doc/passes.texi (Passes): Added documentation about changes done
13670 * doc/invoke.texi (C Dialect Options): Added documentation about
13671 the -fcilkplus flag.
13672 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
13673 (BUILTINS_DEF): Depend on cilkplus.def.
13674 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
13675 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
13676 * cilkplus.def: New file.
13678 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
13680 PR rtl-optimization/57439
13681 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
13683 2013-05-28 Easwaran Raman <eraman@google.com>
13685 PR tree-optimization/57337
13686 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
13687 (find_insert_point): Correctly identify the insertion point
13688 when two statements with the same UID is compared.
13690 2013-05-28 Richard Biener <rguenther@suse.de>
13692 PR tree-optimization/56787
13693 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
13694 from the list of data references.
13695 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
13697 (vect_analyze_loop_operations): Likewise.
13698 (vect_transform_loop): Remove clobbers.
13700 2013-05-28 Martin Jambor <mjambor@suse.cz>
13702 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
13703 and REALPART_EXPRs have scalar type.
13705 2013-05-28 Richard Biener <rguenther@suse.de>
13707 PR tree-optimization/57411
13708 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
13710 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
13711 virtual operand propagation.
13713 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
13715 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
13716 destination register for bmasksi_vis.
13717 (vector_init_bshuffle): Likewise.
13718 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
13720 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
13722 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
13723 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
13724 mode if the instruction isn't available in the original mode.
13725 * config/sparc/sparc.opt (mfix-ut699): New option.
13726 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
13727 (divdf3): Turn into expander.
13728 (divdf3_nofix): New insn.
13729 (divdf3_fix): Likewise.
13730 (divsf3): Disable if -mfix-ut699.
13731 (sqrtdf2): Turn into expander.
13732 (sqrtdf2_nofix): New insn.
13733 (sqrtdf2_fix): Likewise.
13734 (sqrtsf2): Disable if -mfix-ut699.
13736 2013-05-27 Richard Biener <rguenther@suse.de>
13738 PR middle-end/57412
13739 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
13740 block for the new loop.
13742 2013-05-27 Richard Biener <rguenther@suse.de>
13744 PR tree-optimization/57343
13745 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
13746 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
13747 (number_of_iterations_cond): Do not build the folded tree.
13749 2013-05-27 Richard Biener <rguenther@suse.de>
13752 PR middle-end/57381
13753 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
13754 OEP_CONSTANT_ADDRESS_OF retained.
13756 PR tree-optimization/57417
13757 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
13758 for unchanged base.
13759 (set_ssa_val_to): Compare addresses using
13760 get_addr_base_and_unit_offset.
13762 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
13764 PR rtl-optimization/56833
13765 * postreload.c (move2add_record_mode): New function.
13766 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
13767 (move2add_use_add2_insn): Use move2add_record_sym_value.
13768 (move2add_use_add3_insn): Likewise.
13769 (reload_cse_move2add): Use move2add_valid_value_p and
13770 move2add_record_mode. Invalidate call-clobbered and REG_INC
13771 affected regs by setting reg_mode to VOIDmode.
13772 (move2add_note_store): Don't pretend the inside of a SUBREG is
13773 the actual destination. Invalidate single/leading registers by
13774 setting reg_mode to VOIDmode.
13775 Use move2add_record_sym_value, move2add_valid_value_p and
13776 move2add_record_mode.
13778 2013-05-27 Richard Biener <rguenther@suse.de>
13780 PR tree-optimization/57396
13781 * tree-affine.c (double_int_constant_multiple_p): Properly
13782 return false for val == 0 and div != 0.
13784 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
13786 * config/mips/mips.h: Use #elif in preprocessor conditions.
13788 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
13791 * config/mips/constraints.md (kl): New constraint.
13792 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
13793 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
13794 constraint for operand 0. Split after CSE for MIPS16. Emit a move
13795 from LO for MIPS16.
13796 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
13798 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
13801 * config/mips/mips.c (mips_can_inline_p): New function.
13802 (TARGET_CAN_INLINE_P): Define.
13804 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
13806 * sched-int.h (ds_t, dw_t): Make unsigned int.
13807 Fix documentation that describes how all the ds_t bits are used.
13808 Reserve the last bit for delayed-branch scheduling.
13809 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
13810 (BITS_PER_DEP_WEAK): Fix definition and documentation.
13811 (gen_dep_weak_1): Remove prototype.
13812 * sched-deps.c (get_dep_weak_1): Make static.
13813 * target.def (speculate_insn, needs_block_p, gen_spec_check,
13814 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
13815 * doc/tm.texi: Regenerate.
13816 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
13818 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
13821 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
13823 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
13824 Sandra Loosemore <sandra@codesourcery.com>
13826 * config.gcc (powerpc-*): Allow native for with-cpu.
13828 2013-05-24 Jeff Law <law@redhat.com>
13830 PR tree-optimization/57124
13831 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
13832 conversion feeding a condition if the range has an overflow
13833 if -fstrict-overflow. Add warnings for when we do make the
13836 2013-05-24 Dehao Chen <dehao@google.com>
13838 * tree-cfg.c (locus_discrim_map): Fix the typo.
13839 (locus_discrim_hasher): Likewise.
13840 (locus_discrim_hasher::hash): Likewise.
13841 (locus_discrim_hasher::equal): Likewise.
13843 2013-05-24 Martin Jambor <mjambor@suse.cz>
13845 PR tree-optimization/57294
13846 * cgraph.h (ipa_record_stmt_references): Declare.
13847 * cgraphbuild.c (ipa_record_stmt_references): New function.
13848 (build_cgraph_edges): Use ipa_record_stmt_references.
13849 (rebuild_cgraph_edges): Likewise.
13850 (cgraph_rebuild_references): Likewise.
13851 * ipa-prop.c (ipa_modify_call_arguments): Discard references
13852 associated with the old statement and build references from the
13853 newly built statements.
13854 * ipa-ref.c (ipa_remove_stmt_references): New function.
13855 * ipa-ref.h (ipa_remove_stmt_references): Declare.
13857 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
13859 * lra-constraints.c (emit_spill_move): Use smaller mode for
13861 (check_and_process_move): Consider mem-reg moves for secondary
13863 (curr_insn_transform): Don't lose insns emitted before for
13864 secondary memory moves.
13865 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
13866 reg set up in the current insn.
13868 2013-05-24 Dehao Chen <dehao@google.com>
13870 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
13872 (locus_descrim_hasher::equal): Likewise.
13873 (build_gimple_cfg): New discriminator assignment algorithm.
13874 (make_edges): Likewise.
13875 (next_discriminator_for_locus): Likewise.
13876 (same_line_p): Likewise.
13877 (assign_discriminators): Likewise.
13878 (make_cond_expr_edges): Likewise.
13879 (make_gimple_switch_edges): Likewise.
13880 (make_goto_expr_edges): Likewise.
13881 (make_gimple_asm_edges): Likewise.
13883 2013-05-24 Ian Bolton <ian.bolton@arm.com>
13885 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
13886 X format specifier to only display bottom 16 bits.
13887 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
13888 immediate to match for operand 2, since it will be masked.
13890 2013-05-24 Richard Biener <rguenther@suse.de>
13892 PR tree-optimization/57287
13893 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
13894 all SSA names that occur in abnormal PHIs.
13896 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
13899 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
13900 that index is not negative.
13902 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
13904 PR rtl-optimization/55177
13905 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
13906 (simplify_byte_swapping_operation): New.
13907 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
13908 (simplify_relational_operation_1): Deal with BSWAP.
13910 2013-05-23 Richard Henderson <rth@redhat.com>
13913 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
13914 (ix86_reorg): Call it.
13916 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
13919 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
13920 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
13921 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
13923 2013-05-23 Christian Bruel <christian.bruel@st.com>
13926 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
13928 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
13929 Marcus Shawcroft <marcus.shawcroft@arm.com>
13931 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
13932 * config/aarch64/constraints.md (Usa): Remove.
13933 * doc/md.texi (AArch64 Usa): Remove.
13935 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
13936 Marcus Shawcroft <marcus.shawcroft@arm.com>
13938 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
13939 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
13940 * config/aarch64/predicates.md (aarch64_const_address): Remove.
13941 (aarch64_mov_operand): Use aarch64_mov_operand_p.
13943 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
13945 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
13946 instruction (AdvSIMD).
13947 * config/aarch64/aarch64-builtins.c
13948 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
13949 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
13951 2013-05-23 Martin Jambor <mjambor@suse.cz>
13953 PR middle-end/57347
13954 * tree.h (contains_bitfld_component_ref_p): Declare.
13955 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
13956 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
13958 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
13959 not access a bit-field. Assert all final offsets are byte-aligned.
13961 2013-05-23 Richard Biener <rguenther@suse.de>
13963 PR tree-optimization/57380
13964 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
13965 least one invariant or re-used load.
13966 * passes.c (init_optimization_passes): Move pass_phiprop before
13969 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
13971 * config/aarch64/aarch64-simd.md
13972 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
13974 2013-05-23 Richard Biener <rguenther@suse.de>
13976 PR middle-end/57381
13977 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
13978 OEP_CONSTANT_ADDRESS_OF retained.
13980 2013-05-23 Jakub Jelinek <jakub@redhat.com>
13982 PR middle-end/57344
13983 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
13984 don't lower unit. Handle unit not being always BITS_PER_WORD.
13986 2013-05-23 Richard Biener <rguenther@suse.de>
13988 PR rtl-optimization/57341
13989 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
13990 instead of true_dependence.
13992 2013-05-22 David Malcolm <dmalcolm@redhat.com>
13994 * bb-reorder.c (branch_threshold): Make const.
13995 (exec_threshold): Ditto.
13997 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
13998 Pat Haugen <pthaugen@us.ibm.com>
13999 Peter Bergner <bergner@vnet.ibm.com>
14001 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
14002 documentation for the power8 crypto builtins.
14004 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
14006 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
14007 macros for defining power8 builtin functions.
14008 (BU_P8V_AV_2): Likewise.
14009 (BU_P8V_AV_P): Likewise.
14010 (BU_P8V_VSX_1): Likewise.
14011 (BU_P8V_OVERLOAD_1): Likewise.
14012 (BU_P8V_OVERLOAD_2): Likewise.
14013 (BU_CRYPTO_1): Likewise.
14014 (BU_CRYPTO_2): Likewise.
14015 (BU_CRYPTO_3): Likewise.
14016 (BU_CRYPTO_OVERLOAD_1): Likewise.
14017 (BU_CRYPTO_OVERLOAD_2): Likewise.
14018 (XSCVSPDP): Fix typo, point to the correct instruction.
14019 (VCIPHER): Add power8 crypto builtins.
14020 (VCIPHERLAST): Likewise.
14021 (VNCIPHER): Likewise.
14022 (VNCIPHERLAST): Likewise.
14023 (VPMSUMB): Likewise.
14024 (VPMSUMH): Likewise.
14025 (VPMSUMW): Likewise.
14026 (VPERMXOR_V2DI): Likewise.
14027 (VPERMXOR_V4SI: Likewise.
14028 (VPERMXOR_V8HI: Likewise.
14029 (VPERMXOR_V16QI: Likewise.
14030 (VSHASIGMAW): Likewise.
14031 (VSHASIGMAD): Likewise.
14032 (VPMSUM): Likewise.
14033 (VPERMXOR): Likewise.
14034 (VSHASIGMA): Likewise.
14036 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
14037 __CRYPTO__ if the crypto instructions are available.
14038 (altivec_overloaded_builtins): Add support for overloaded power8
14041 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
14042 support for power8 crypto builtins.
14043 (builtin_function_type): Likewise.
14044 (altivec_init_builtins): Add support for builtins that take vector
14045 long long (V2DI) arguments.
14047 * config/rs6000/crypto.md: New file, define power8 crypto
14050 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
14051 Pat Haugen <pthaugen@us.ibm.com>
14052 Peter Bergner <bergner@vnet.ibm.com>
14054 * doc/invoke.texi (Option Summary): Add power8 options.
14055 (RS/6000 and PowerPC Options): Likewise.
14057 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
14058 constraints.md instead of rs6000.h. Reorder w* constraints. Add
14059 wm, wn, wr documentation.
14061 * config/rs6000/constraints.md (wm): New constraint for VSX
14062 registers if direct move instructions are enabled.
14063 (wn): New constraint for no registers.
14064 (wq): New constraint for quad word even GPR registers.
14065 (wr): New constraint if 64-bit instructions are enabled.
14066 (wv): New constraint if power8 vector instructions are enabled.
14067 (wQ): New constraint for quad word memory locations.
14069 * config/rs6000/predicates.md (const_0_to_15_operand): New
14070 constraint for 0..15 for crypto instructions.
14071 (gpc_reg_operand): If VSX allow registers in VSX registers as well
14072 as GPR and floating point registers.
14073 (int_reg_operand): New predicate to match only GPR registers.
14074 (base_reg_operand): New predicate to match base registers.
14075 (quad_int_reg_operand): New predicate to match even GPR registers
14076 for quad memory operations.
14077 (vsx_reg_or_cint_operand): New predicate to allow vector logical
14078 operations in both GPR and VSX registers.
14079 (quad_memory_operand): New predicate for quad memory operations.
14080 (reg_or_indexed_operand): New predicate for direct move support.
14082 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
14083 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
14084 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
14085 (POWERPC_MASKS): Add power8 options.
14086 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
14089 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
14090 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
14092 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
14093 (-mpower8-fusion): New power8 options.
14094 (-mpower8-fusion-sign): Likewise.
14095 (-mpower8-vector): Likewise.
14096 (-mcrypto): Likewise.
14097 (-mdirect-move): Likewise.
14098 (-mquad-memory): Likewise.
14100 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
14101 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
14103 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
14104 (rs6000_debug_vector_unit): Add p8_vector.
14105 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
14106 definitions. Also print fusion state.
14107 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
14108 (rs6000_builtin_mask_calculate): Add power8 builtin support.
14109 (rs6000_option_override_internal): Add support for power8.
14110 (rs6000_common_init_builtins): Add debugging for skipped builtins
14111 if -mdebug=builtin.
14112 (rs6000_adjust_cost): Add power8 support.
14113 (rs6000_issue_rate): Likewise.
14114 (insn_must_be_first_in_group): Likewise.
14115 (insn_must_be_last_in_group): Likewise.
14116 (force_new_group): Likewise.
14117 (rs6000_register_move_cost): Likewise.
14118 (rs6000_opt_masks): Likewise.
14120 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
14121 power8 capable assembler, default to power7 options.
14122 (TARGET_DIRECT_MOVE): Likewise.
14123 (TARGET_CRYPTO): Likewise.
14124 (TARGET_P8_VECTOR): Likewise.
14125 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
14126 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
14127 (VECTOR_MEM_P8_VECTOR_P): Likewise.
14128 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
14129 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
14130 (TARGET_XSCVDPSPN): Likewise.
14131 (TARGET_XSCVSPDPN): Likewsie.
14132 (TARGET_SYNC_HI_QI): Likewise.
14133 (TARGET_SYNC_TI): Likewise.
14134 (MASK_CRYPTO): Likewise.
14135 (MASK_DIRECT_MOVE): Likewise.
14136 (MASK_P8_FUSION): Likewise.
14137 (MASK_P8_VECTOR): Likewise.
14138 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
14139 temporary used by some of the direct move instructions to get two FP
14140 temporary registers does not force creation of a stack frame.
14141 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
14142 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
14143 that any VSX registers are tieable, even if they are also an
14144 Altivec vector mode.
14145 (r6000_reg_class_enum): Add wm, wr, wv constraints.
14146 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
14147 (RS6000_BTM_CRYPTO): Likewise.
14148 (RS6000_BTM_COMMON): Likewise.
14150 * config/rs6000/rs6000.md (cpu attribute): Add power8.
14151 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
14152 (enum rs6000_vector): Add power8 vector support.
14154 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14158 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
14159 (any_sibcall_could_use_r3): this and handle indirect calls.
14160 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
14162 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14164 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
14166 2013-05-22 Richard Biener <rguenther@suse.de>
14168 PR middle-end/57349
14169 * profile.c (branch_prob): Do not split blocks that are
14170 abnormally receiving from ECF_RETURNS_TWICE functions.
14172 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
14174 * recog.c (offsettable_address_addr_space_p): Fix calculation of
14175 address mode. Move pointer mode initialization to the same place.
14177 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
14179 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
14180 while it has any effect.
14182 2013-05-21 Easwaran Raman <eraman@google.com>
14184 PR tree-optimization/57322
14185 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
14186 UID of the statement added to the BB to be 1.
14188 2013-05-21 Jakub Jelinek <jakub@redhat.com>
14190 PR tree-optimization/57331
14191 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
14192 of conversion from pointer type to integral type with integer.
14194 2013-05-21 Martin Jambor <mjambor@suse.cz>
14197 * ipa-prop.c (ipa_read_node_info): Process param_used and
14198 controlled_uses in the same order as when writing.
14200 2013-05-21 Magnus Granberg <baldrick@free.fr>
14203 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
14205 2013-05-21 Richard Biener <rguenther@suse.de>
14207 PR tree-optimization/57318
14208 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
14209 estimate stmts with side-effects as likely eliminated.
14211 2013-05-21 Richard Biener <rguenther@suse.de>
14213 PR tree-optimization/57330
14214 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
14215 preserve the call stmts fntype.
14217 2013-05-21 Richard Biener <rguenther@suse.de>
14219 PR tree-optimization/57303
14220 * tree-ssa-sink.c (statement_sink_location): Improve killing
14221 stmt detection and properly handle self-assignments.
14223 2013-05-21 Christian Bruel <christian.bruel@st.com>
14225 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
14226 spanning registers. LEAF_REG_REMAP is supported only for contiguous
14227 registers. Set register size out of the PARALLEL loop.
14229 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
14232 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
14233 (*fmasf4, *fmasf4_media): New insns.
14235 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
14237 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
14238 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
14239 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
14240 (mips_idiv_insns): Update the comments to say that the returned
14241 instruction counts are in units of BASE_INSN_LENGTH.
14242 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
14243 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
14244 using 2 rather than 4 as the length of indirect MIPS16 and
14245 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
14246 length of a NOP. Don't divide MIPS16 lengths by 2.
14247 (mips16_split_long_branches): Assume a branch is long if the
14248 length is greater than 4 rather than 8.
14249 * config/mips/mips.md (length): Give MIPS16 lengths directly,
14250 rather than multiplying them by 2. Multiply instruction counts
14251 by BASE_INSN_LENGTH rather than 4.
14252 (*jump_mips16, tls_get_tp_mips16_<mode>)
14253 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
14255 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
14257 * config/mips/mips.md (extended_mips16): Remove branch case.
14258 (length): Remove duplicated extended_mips16 test.
14260 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
14262 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
14264 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
14266 * recog.h (Recog_data): Rename to...
14267 (recog_data_d): ...this.
14268 (recog_data): Update accordingly.
14269 * recog.c (recog_data): Likewise.
14270 * reload.c (save_recog_data): Likewise.
14271 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
14272 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
14274 2013-05-17 Julian Brown <julian@codesourcery.com>
14276 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
14277 found in a REG_EQUAL note, invalidate it.
14279 2013-05-17 Easwaran Raman <eraman@google.com>
14281 * tree-ssa-reassoc.c (find_insert_point): New function.
14282 (insert_stmt_after): Likewise.
14283 (get_def_stmt): Likewise.
14284 (ensure_ops_are_available): Likewise.
14285 (not_dominated_by): Likewise.
14286 (rewrite_expr_tree): Do not move statements beyond what is
14287 necessary. Remove call to swap_ops_for_binary_stmt...
14288 (reassociate_bb): ... and move it here.
14289 (build_and_add_sum): Assign UIDs for new statements.
14290 (linearize_expr): Likewise.
14291 (do_reassoc): Renumber gimple statement UIDs.
14293 2013-05-17 Jan Hubicka <jh@suse.cz>
14295 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
14297 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
14298 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
14300 (output_weakrefs): Update.
14302 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
14303 Martin Jambor <mjambor@suse.cz>
14305 PR middle-end/57276
14306 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
14307 value that corresponds to the given aggval is found in values vector.
14309 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
14311 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
14312 sse, sse2, sse3, ssse3 and sse4a flags to options.
14314 2013-05-17 David Malcolm <dmalcolm@redhat.com>
14316 * gengtype-state.c: (s_expr_writer): New class, to handle
14317 prettifying of output layout of s-expressions.
14318 (state_writer): New class, to write out gtype.state.
14319 (state_written_type_count): Move this variable into member data of
14321 (s_expr_writer::s_expr_writer): New code: constructor for new class
14322 (state_writer::state_writer(): ditto
14323 (s_expr_writer::write_new_line): New function
14324 (s_expr_writer::write_any_indent): ditto
14325 (s_expr_writer::begin_s_expr): ditto
14326 (s_expr_writer::end_s_expr): ditto
14327 (write_state_fileloc): convert to method of state_writer...
14328 (state_writer:: write_state_fileloc): ...and use methods of
14329 s_expr_writer to write indentation into the gtype.state output file
14330 to visually represent the hierarchical structure of the list
14332 (write_state_fields): ditto, renaming to...
14333 (state_writer::write_state_fields)
14334 (write_state_a_string): ditto, renaming to...
14335 (state_writer::write_state_a_string)
14336 (write_state_string_option): ditto, renaming to...
14337 (state_writer::write_state_string_option)
14338 (write_state_type_option): ditto, renaming to...
14339 (state_writer::write_state_type_option)
14340 (write_state_nested_option): ditto, renaming to...
14341 (state_writer::write_state_nested_option)
14342 (write_state_option): ditto, renaming to...
14343 (state_writer::write_state_option)
14344 (write_state_options): ditto, renaming to...
14345 (state_writer::write_state_options)
14346 (write_state_lang_bitmap): ditto, renaming to...
14347 (state_writer::write_state_lang_bitmap)
14348 (write_state_version): ditto, renaming to...
14349 (state_writer::write_state_version)
14350 (write_state_scalar_type): ditto, renaming to...
14351 (state_writer::write_state_scalar_type)
14352 (write_state_string_type): ditto, renaming to...
14353 (state_writer::write_state_string_type)
14354 (write_state_undefined_type): ditto, renaming to...
14355 (state_writer::write_state_undefined_type)
14356 (write_state_struct_union_type): ditto, renaming to...
14357 (state_writer::write_state_struct_union_type)
14358 (write_state_struct_type): ditto, renaming to...
14359 (state_writer::write_state_struct_type)
14360 (write_state_user_struct_type): ditto, renaming to...
14361 (state_writer::write_state_user_struct_type)
14362 (write_state_lang_struct_type): ditto, renaming to...
14363 (state_writer::write_state_lang_struct_type)
14364 (write_state_param_struct_type): ditto, renaming to...
14365 (state_writer::write_state_param_struct_type)
14366 (write_state_pointer_type): ditto, renaming to...
14367 (state_writer::write_state_pointer_type)
14368 (write_state_array_type): ditto, renaming to...
14369 (state_writer::write_state_array_type)
14370 (write_state_gc_used): ditto, renaming to...
14371 (state_writer::write_state_gc_used)
14372 (write_state_common_type_content): ditto, renaming to...
14373 (state_writer::write_state_common_type_content)
14374 (write_state_type): ditto, renaming to...
14375 (state_writer::write_state_type)
14376 (write_state_pair_list): ditto, renaming to...
14377 (state_writer::write_state_pair_list)
14378 (write_state_pair): ditto, renaming to...
14379 (state_writer::write_state_pair)
14380 (write_state_typedefs): ditto, renaming to...
14381 (state_writer::write_state_typedefs)
14382 (write_state_structures): ditto, renaming to...
14383 (state_writer::write_state_structures)
14384 (write_state_param_structs): ditto, renaming to...
14385 (state_writer::write_state_param_structs)
14386 (write_state_variables): ditto, renaming to...
14387 (state_writer::write_state_variables)
14388 (write_state_srcdir): ditto, renaming to...
14389 (state_writer::write_state_srcdir)
14390 (write_state_files_list): ditto, renaming to...
14391 (state_writer::write_state_files_list)
14392 (write_state_languages): ditto, renaming to...
14393 (state_writer::write_state_languages)
14394 (write_state): create a state_writer instance and use it when
14395 writing out the state file
14397 2013-05-17 Mike Stump <mikestump@comcast.net>
14399 PR rtl-optimization/57304
14400 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
14401 accessing DF_REF_REAL_LOC.
14403 2013-05-17 Jakub Jelinek <jakub@redhat.com>
14405 PR rtl-optimization/57281
14406 PR rtl-optimization/57300
14407 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
14408 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
14409 what the other splitter did if the registers are dead.
14411 2013-05-17 Richard Biener <rguenther@suse.de>
14413 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
14416 2013-05-17 Jakub Jelinek <jakub@redhat.com>
14418 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
14421 2013-05-17 Marek Polacek <polacek@redhat.com>
14423 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
14424 length when doing non-zero store of storing '\0' to '\0'.
14426 2013-05-17 Jakub Jelinek <jakub@redhat.com>
14428 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
14429 vect_external_def oprnd1 with loop_vinfo, try to emit
14430 optional cast, negation and and stmts on the loop preheader
14431 edge instead of into the pattern def seq.
14433 PR tree-optimization/57051
14434 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
14435 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
14437 2013-05-16 Nick Clifton <nickc@redhat.com>
14439 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
14440 (rl78_is_naked_func): New function.
14441 (rl78_expand_prologue): Skip prologue generation for naked functions.
14442 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
14443 * doc/extend.texi (naked): Add RL78 to the list of processors
14444 that supports this attribute.
14446 2013-05-16 Jeff Law <law@redhat.com>
14448 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
14450 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
14452 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
14453 cache parameters using detect_caches_amd also for CYRIX,
14454 NSC and TM2 signatures.
14456 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
14457 Dzianis Kahanovich <mahatma@eu.by>
14461 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
14462 VIA/Centaur processors and determine their cache parameters
14463 using detect_caches_amd.
14465 2013-05-16 Teresa Johnson <tejohnson@google.com>
14467 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
14468 (rtl_verify_edges): New function.
14469 (rtl_verify_bb_insns): Ditto.
14470 (rtl_verify_bb_pointers): Ditto.
14471 (rtl_verify_bb_insn_chain): Ditto.
14472 (rtl_verify_fallthru): Ditto.
14473 (rtl_verify_bb_layout): Ditto.
14474 (rtl_verify_flow_info_1): Outline checks into new functions.
14475 (rtl_verify_flow_info): Ditto.
14477 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
14479 * cfghooks.c (copy_bbs): Add update_dominance argument.
14480 * cfghooks.h (copy_bbs): Update prototype.
14481 * tree-cfg.c (gimple_duplicate_sese_region):
14482 Add update_dominance argument.
14483 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
14484 * tree-ssa-loop-ch.c (copy_loop_headers): Update
14485 gimple_duplicate_sese_region call.
14486 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
14487 Update copy_bbs call.
14488 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
14489 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
14491 2013-05-16 Jakub Jelinek <jakub@redhat.com>
14493 * tree-vectorizer.h (NUM_PATTERNS): Increment.
14494 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
14495 vect_recog_rotate_pattern.
14496 (vect_recog_rotate_pattern): New function.
14498 2013-05-16 Jason Merrill <jason@redhat.com>
14500 * Makefile.in (LLINKER): New variable.
14501 (mostlyclean): Remove link mutex.
14502 * configure.ac: Handle --enable-link-mutex.
14503 * lock-and-run.sh: New script.
14505 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14508 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
14511 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14513 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
14515 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
14517 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
14518 * config/arm/arm.c (next_consecutive_mem): New function.
14519 (gen_movmem_ldrd_strd): Likewise.
14520 * config/arm/arm.md (movmemqi): Update condition and code.
14521 (unaligned_loaddi, unaligned_storedi): New patterns.
14523 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14525 * config.gcc: Obsolete *-*-solaris2.9*.
14526 * doc/install.texi (Specific, *-*-solaris2*): Document it.
14528 2013-05-16 Richard Biener <rguenther@suse.de>
14530 * passes.c (init_optimization_passes): Move pass_parallelize_loops
14531 earlier, after GRAPHITE transforms and IV canonicalization.
14533 2013-05-16 Jakub Jelinek <jakub@redhat.com>
14535 * omp-low.c (extract_omp_for_data): For collapsed loops,
14536 if at least one of the loops is known at compile time to
14537 iterate zero times, set count to 0.
14538 (expand_omp_regimplify_p): New function.
14539 (expand_omp_for_generic): For collapsed loops, if at least
14540 one of the loops isn't known to iterate at least once,
14541 add runtime check with setting count to 0.
14542 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
14543 For unsigned types if it isn't known at compile time that
14544 the loop will iterate at least once, add runtime check to bypass
14545 the whole loop if initial condition isn't true.
14547 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
14549 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
14551 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
14553 PR middle-end/57286
14554 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
14555 transformations to avoid an infinite loop.
14557 2013-05-16 Marek Polacek <polacek@redhat.com>
14559 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
14561 2013-05-15 Leif Ekblad <leif@rdos.net>
14563 * config/i386/i386.c (ix86_decompose_address): Use
14564 DEFAULT_TLS_SEG_REG to access TLS segment register.
14565 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
14566 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
14567 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
14569 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
14572 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
14573 sibling calls to functions that would normally be lazily bound,
14574 unless $gp is call-clobbered.
14576 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
14578 * config/i386/i386.c (ix86_option_override_internal): Update
14579 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
14580 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
14581 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
14582 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
14584 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
14586 2013-05-15 Andreas Schwab <schwab@suse.de>
14588 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
14589 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
14592 2013-05-15 Teresa Johnson <tejohnson@google.com>
14594 * loop-unroll.c (report_unroll_peel): Check decision before
14595 emitting unroll/peel message.
14597 2013-05-15 Teresa Johnson <tejohnson@google.com>
14599 * function.h (has_bb_partition): New rtl_data flag.
14600 (bb_reorder_complete): Ditto.
14601 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
14602 instead of flag_reorder_blocks_and_partition.
14603 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
14604 with some enhancements.
14605 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
14606 * bb-reorder.c (connect_traces): Check for has_bb_partition
14607 instead of flag_reorder_blocks_and_partition.
14608 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
14609 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
14610 verify_hot_cold_block_grouping.
14611 (partition_hot_cold_basic_blocks): Set has_bb_partition.
14613 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14616 * config/arm/predicates.md (call_insn_operand): New predicate.
14617 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
14618 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
14619 if insn is not a tail call.
14620 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
14622 * config/arm/arm.h (enum reg_class): New caller save register class.
14623 (REG_CLASS_NAMES): Likewise.
14624 (REG_CLASS_CONTENTS): Likewise.
14625 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
14628 2013-05-15 Richard Biener <rguenther@suse.de>
14630 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
14631 of MSG_OPTIMIZED_LOCATIONS.
14632 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
14633 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
14635 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
14636 of MSG_OPTIMIZED_LOCATIONS.
14637 (execute_vect_slp): Likewise.
14638 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
14639 (vect_create_cond_for_alias_checks): Likewise.
14640 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
14641 (vect_recog_widen_mult_pattern): Likewise.
14642 (vect_recog_widen_sum_pattern): Likewise.
14643 (vect_recog_over_widening_pattern): Likewise.
14644 (vect_recog_widen_shift_pattern): Likewise.
14645 (vect_recog_vector_vector_shift_pattern): Likewise.
14646 (vect_recog_divmod_pattern): Likewise.
14647 (vect_recog_mixed_size_cond_pattern): Likewise.
14648 (vect_recog_bool_pattern): Likewise.
14649 (vect_pattern_recog_1): Likewise.
14651 2013-05-15 Martin Jambor <mjambor@suse.cz>
14653 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
14654 non-functions to builtin_unreachable.
14655 * ipa-inline-transform.c (inline_call): Do not assert estimates were
14656 correct when new direct edges were discovered.
14658 2013-05-15 Martin Jambor <mjambor@suse.cz>
14660 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
14661 header, print symbol order instead of node uid, print more information
14662 about indirect edge targets.
14663 (ipa_make_edge_direct_to_target): Print symbol order instead of node
14665 (ipa_make_edge_direct_to_target): Likewise.
14666 (remove_described_reference): Likewise.
14667 (propagate_controlled_uses): Likewise.
14668 (ipa_print_node_params): Also print symbol order.
14669 (ipcp_transform_function): Print symbol order instead of node uids.
14670 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
14671 (cgraph_get_create_real_symbol_node): Likewise.
14672 * ipa-cp.c (print_lattice): Likewise.
14673 (print_all_lattices): Likewise.
14674 (determine_versionability): Likewise.
14675 (initialize_node_lattices): Likewise.
14676 (estimate_local_effects): Likewise.
14677 (update_profiling_info): Likewise.
14678 (create_specialized_node): Likewise.
14679 (perhaps_add_new_callers): Likewise.
14680 (decide_about_value): Likewise.
14681 (decide_whether_version_node): Likewise.
14682 (identify_dead_nodes): Likewise.
14683 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
14684 (dump_inline_summary): Likewise.
14685 (estimate_node_size_and_time): Likewise.
14686 (inline_analyze_function): Likewise.
14687 * ipa-inline.c (report_inline_failed_reason): Likewise.
14688 (want_early_inline_function_p): Likewise.
14689 (edge_badness): Likewise.
14690 (update_edge_key): Likewise.
14691 (inline_small_functions): Likewise. Add dumping of order to two other
14693 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
14694 instead of node uids.
14695 (propagate_pure_const): Likewise.
14696 (propagate_pure_const): Likewise.
14697 * ipa-utils.c (dump_cgraph_node_set): Likewise.
14698 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
14699 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
14701 * tree-pretty-print.c (dump_function_header): Likewise.
14702 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
14703 Print symbol order instead of node uids.
14705 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14707 * config/s390/s390.c (s390_register_move_cost): Don't impose the
14708 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
14710 2013-05-15 Richard Biener <rguenther@suse.de>
14712 PR tree-optimization/57275
14713 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
14714 return value for fail to do runtime alias checks for gather loads.
14716 2013-05-15 Jan Hubicka <jh@suse.cz>
14720 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
14721 weakrefs are not external.
14722 (lto_symtab_merge_decls): Fix thinko when dealing with
14723 non-lto_symtab decls.
14724 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
14725 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
14726 * varpool.c (dump_varpool_node): Dump more flags.
14728 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
14730 * config/i386/i386.c (processor_alias_table): Add instruction
14731 FSGSBASE for AMD bdver3 architecture.
14733 2013-05-14 Jakub Jelinek <jakub@redhat.com>
14735 * tree.c (warn_deprecated_use): Print file:line using locus color.
14736 * diagnostic.c (diagnostic_report_current_module): Print file:line
14737 and file:line:column using locus color.
14739 2013-05-14 Mike Stump <mikestump@comcast.net>
14741 * gdbinit.in: Add __null.
14743 2013-05-14 Mike Stump <mikestump@comcast.net>
14745 * recog.h: Rename struct recog_data to Recog_data.
14746 * recog.c: Likewise.
14747 * reload.c (can_reload_into): Likewise.
14748 * config/picochip/picochip.c: Likewise.
14750 2013-05-14 Mike Stump <mikestump@comcast.net>
14752 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
14754 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
14756 * resource.h (struct resources): Remove unch_memory member.
14757 (CLEAR_RESOURCE): Don't clear unch_memory.
14758 * resource.c (mark_referenced_resources): Don't set it.
14759 (mark_set_resources): Likewise.
14760 (mark_target_live_regs): Don't clear it.
14761 (init_resource_info): Likewise.
14762 * reorg.c (resource_conflicts_p): Don't compare it.
14763 (redundant_insn): Don't set it.
14765 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
14767 * emit-rtl.c (next_label): Remove unused function.
14768 (skip_consecutive_labels, link_cc0_insns): Move to ...
14769 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
14770 only place where these functions are used, and make them static.
14772 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
14774 * fold-const.c (fold_negate_expr): Handle vectors.
14775 (fold_truth_not_expr): Make it static.
14776 (fold_invert_truthvalue): New static function.
14777 (invert_truthvalue_loc): Handle vectors. Do not call
14778 fold_truth_not_expr directly.
14779 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
14780 <TRUTH_NOT_EXPR>: Do not cast to boolean.
14781 (fold_comparison): Handle vector constants.
14782 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
14783 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
14784 * tree.h (fold_truth_not_expr): Remove declaration.
14786 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
14788 * config/aarch64/aarch64-simd.md
14789 (aarch64_vcond_internal<mode>): Rename to...
14790 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
14791 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
14792 float modes. Clarify all iterator modes.
14793 (vcond<mode><mode>): Use new name for vcond expanders.
14794 (vcond<v_cmp_result><mode>): Likewise.
14795 (vcondu<mode><mode>: Likewise.
14796 * config/aarch64/iterators.md (VDQF_COND): New.
14798 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
14801 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
14802 variable for the shift amount. Check that we shift by non-negative
14805 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
14808 * config/arm/arm.h (EPILOGUE_USES): Only return true
14809 for LR_REGNUM after epilogue_completed.
14811 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
14813 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
14814 is error_mark_node.
14816 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14819 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
14820 and Solaris 11+/x86 with gld.
14821 * configure: Regenerate.
14823 2013-05-14 Jakub Jelinek <jakub@redhat.com>
14825 * expmed.c (expand_shift_1): Canonicalize rotates by
14826 constant bitsize / 2 to bitsize - 1.
14827 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
14828 case ROTATERT>: Likewise.
14831 2013-05-10 Jakub Jelinek <jakub@redhat.com>
14833 * config/i386/i386.md (rotateinv): New code attr.
14834 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
14835 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
14836 roll $31, %eax, etc.
14838 2013-05-14 Richard Biener <rguenther@suse.de>
14840 PR middle-end/57235
14841 * tree-eh.c (sink_clobbers): Give up for successors with
14842 multiple predecessors and no virtual uses.
14844 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
14846 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
14847 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
14849 2013-05-14 Jakub Jelinek <jakub@redhat.com>
14851 PR middle-end/57251
14852 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
14853 the case when both op0 and op1 have VOIDmode.
14855 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14857 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
14858 in multiply-accumulate mode.
14860 2013-05-13 Guozhi Wei <carrot@google.com>
14862 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
14864 2013-05-13 Kai Tietz <ktietz@redhat.com>
14867 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
14868 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
14869 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
14870 * config/i386/i386.c (ix86_option_override_internal): Likewise.
14871 (ix86_expand_prologue): Likewise.
14872 (ix86_expand_split_stack_prologue): Likewise.
14873 (legitimate_pic_address_disp_p): Likewise.
14874 (legitimize_pic_address): Likewise.
14875 (legitimize_tls_address): Likewise.
14876 (legitimize_pe_coff_symbol): Likewise.
14877 (output_pic_addr_const): Likewise.
14878 (construct_plt_address): Likewise.
14879 (ix86_expand_call): Likewise.
14880 (x86_output_mi_thunk): Likewise.
14881 (x86_function_profiler): Likewise.
14883 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
14885 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
14886 similar switch cases.
14887 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
14888 (aarch64_simd_mov_to_<mode>low): Delete.
14889 (aarch64_simd_mov_to_<mode>high): Delete.
14890 (move_lo_quad_<mode>): Add w<-r alternative.
14891 (aarch64_simd_move_hi_quad_<mode>): Likewise.
14892 (aarch64_simd_mov_from_*): Update type attribute.
14893 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
14896 2013-05-13 Jan Hubicka <jh@suse.cz>
14898 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
14899 * config/i386/i386.c (ix86_compute_frame_layout,
14900 ix86_expand_epilogue, emit_i387_cw_initialization,
14901 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
14902 ix86_local_alignment): Fix use of size/speed predicates.
14904 2013-05-13 Jakub Jelinek <jakub@redhat.com>
14906 PR tree-optimization/45216
14907 PR tree-optimization/57157
14908 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
14909 the (-Y) & (B - 1) variant if OP is |.
14910 * expmed.c (expand_shift_1): For rotations by const0_rtx just
14911 return shifted. Use (-op1) & (prec - 1) as other_amount
14912 instead of prec - op1.
14914 2013-05-13 Martin Jambor <mjambor@suse.cz>
14916 PR middle-end/42371
14917 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
14918 (ipa_constant_data): New type.
14919 (ipa_jump_func): Use ipa_constant_data to hold information about
14920 constant jump functions.
14921 (ipa_get_jf_constant): Adjust to jump function type changes.
14922 (ipa_get_jf_constant_rdesc): New function.
14923 (ipa_param_descriptor): New field controlled_uses.
14924 (ipa_get_controlled_uses): New function.
14925 (ipa_set_controlled_uses): Likewise.
14926 * ipa-ref.h (ipa_find_reference): Declare.
14927 * ipa-prop.c (ipa_cst_ref_desc): New type.
14928 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
14930 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
14931 New parameter cs. Adjust all callers.
14932 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
14933 (remove_described_reference): New function.
14934 (jfunc_rdesc_usable): Likewise.
14935 (try_make_edge_direct_simple_call): Decrement controlled use count,
14936 attempt to remove reference if it hits zero.
14937 (combine_controlled_uses_counters): New function.
14938 (propagate_controlled_uses): Likewise.
14939 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
14940 (ipa_edge_duplication_hook): Duplicate reference descriptions.
14941 (ipa_print_node_params): Print described use counter.
14942 (ipa_write_jump_function): Adjust to jump function type changes.
14943 (ipa_read_jump_function): New parameter CS, pass it to
14944 ipa_set_jf_constant. Adjust caller.
14945 (ipa_write_node_info): Stream controlled use count
14946 (ipa_read_node_info): Likewise.
14947 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
14949 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
14950 count. Remove cloning-added reference if it reaches zero.
14951 * ipa-ref.c (ipa_find_reference): New function.
14953 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
14955 * config/i386/i386.c (processor_target_table): Modified default
14956 alignment values for AMD BD and BT architectures.
14958 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
14960 * tree-vect-generic.c (uniform_vector_p): Move ...
14961 * tree.c (uniform_vector_p): ... here.
14962 * tree.h (uniform_vector_p): Declare it.
14963 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
14966 2013-05-13 Jakub Jelinek <jakub@redhat.com>
14968 PR tree-optimization/57230
14969 * tree-ssa-strlen.c (handle_char_store): Record length for
14970 array store from STRING_CST.
14972 PR tree-optimization/57230
14973 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
14976 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
14978 * config/epiphany/epiphany.c (epiphany_init): Check size of
14979 NUM_MODES_FOR_MODE_SWITCHING.
14980 (epiphany_expand_prologue):
14981 Remove CONFIG_REGNUM initial value handling code.
14982 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
14983 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
14984 (emit_set_fp_mode, epiphany_mode_after): Likewise.
14985 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
14986 Don't return 1 for FP_MODE_NONE.
14987 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
14988 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
14989 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
14990 * config/epiphany/epiphany.md (save_config): New pattern.
14992 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
14994 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
14996 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
14998 * config/i386/i386.md (memory): Handle sseishft1.
14999 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
15000 (*vec_extractv2di_1): Ditto.
15002 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
15004 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
15007 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
15009 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
15010 Add mthumb/march=armv7-a multilib.
15011 Add mthumb/march=armv7-r multilib.
15012 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
15014 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
15016 * config/v850/t-rtems: Add more multilibs.
15018 2013-05-10 Richard Biener <rguenther@suse.de>
15020 PR tree-optimization/57214
15021 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
15022 not propagate from SSA names that occur in abnormal PHI nodes.
15024 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
15026 * stor-layout.c (element_precision): New function.
15027 * machmode.h (element_precision): Declare it.
15028 * tree.c (build_minus_one_cst): New function.
15029 (element_precision): Likewise.
15030 * tree.h (build_minus_one_cst): Declare new function.
15031 (element_precision): Likewise.
15032 * fold-const.c (operand_equal_p): Use element_precision.
15033 (fold_binary_loc): Handle vector types.
15034 * convert.c (convert_to_integer): Use element_precision.
15035 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
15038 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
15040 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
15041 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
15042 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
15043 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
15044 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
15045 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
15046 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
15047 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
15048 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
15049 (Uuw8): New constraints.
15050 (Usb4): Move into alphabetical order.
15051 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
15052 (sd8_operand, ub8_operand, uw8_operand): New predicates.
15053 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
15054 previously unnamed patterns.
15055 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
15056 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
15057 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
15058 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
15059 of set_attr_alternative/if_then_else. Use extended_mips16 instead
15060 of specific lengths.
15062 2013-05-10 Jakub Jelinek <jakub@redhat.com>
15064 * config/i386/i386.md (rotateinv): New code attr.
15065 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
15066 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
15067 roll $31, %eax, etc.
15069 PR tree-optimization/45216
15070 PR tree-optimization/57157
15071 * tree-ssa-forwprop.c (simplify_rotate): New function.
15072 (ssa_forward_propagate_and_combine): Call it.
15074 2013-05-10 Richard Biener <rguenther@suse.de>
15076 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
15077 disable peeling when we version for aliasing.
15078 (vector_alignment_reachable_p): Honor explicit user alignment.
15079 (vect_supportable_dr_alignment): Likewise.
15080 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
15081 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
15082 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
15083 then peeling to arrange for the cost-model check to come first.
15085 2013-05-10 Alan Modra <amodra@gmail.com>
15087 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
15088 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
15089 * configure: Regenerate.
15091 2013-05-10 Alan Modra <amodra@gmail.com>
15094 * varasm.c (default_elf_select_section): Move !DECL_P check..
15095 (get_named_section): ..to here before calling get_section_name.
15097 (default_section_type_flags): Add DECL_P check.
15098 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
15099 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
15101 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
15103 * config/epiphany/epiphany.c (epiphany_expand_prologue):
15104 When using gen_stack_adjust_str with a register offset, add a
15105 REG_FRAME_RELATED_EXPR note.
15107 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
15109 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
15110 (*vec_extractv4si_zext_mem): Ditto.
15111 (*vec_extractv2di): Add 0->x and x->x alternatives.
15112 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
15113 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
15115 2013-05-09 Jason Merrill <jason@redhat.com>
15117 N3639 C++1y VLA support
15118 * gimplify.c (gimplify_vla_decl): Don't touch an existing
15121 * tree.c (build_constructor_va): New.
15122 * tree.h: Declare it.
15124 2013-05-09 Martin Jambor <mjambor@suse.cz>
15127 * gimple-fold.c (canonicalize_constructor_val): Call
15128 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
15130 2013-05-09 Jan Hubicka <jh@suse.cz>
15131 Richard Biener <rguenther@suse.de>
15134 * symtab.c (symtab_make_decl_local): Do not add private names.
15136 2013-05-09 Jan Hubicka <jh@suse.cz>
15139 * symtab.c (insert_to_assembler_name_hash): Handle clones.
15140 (unlink_from_assembler_name_hash): Likewise.
15141 (symtab_prevail_in_asm_name_hash, symtab_register_node,
15142 symtab_unregister_node, symtab_initialize_asm_name_hash,
15143 change_decl_assembler_name): Update.
15145 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
15147 * config/aarch64/aarch64.md: New movtf split.
15148 (*movtf_aarch64): Update.
15149 (aarch64_movdi_tilow): Handle TF modes and rename to
15150 aarch64_movdi_<mode>low.
15151 (aarch64_movdi_tihigh): Handle TF modes and rename to
15152 aarch64_movdi_<mode>high
15153 (aarch64_movtihigh_di): Handle TF modes and rename to
15154 aarch64_mov<mode>high_di
15155 (aarch64_movtilow_di): Handle TF modes and rename to
15156 aarch64_mov<mode>low_di
15157 (aarch64_movtilow_tilow): Remove spurious whitespace.
15158 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
15160 (aarch64_print_operand): Update.
15162 2013-05-09 Alan Modra <amodra@gmail.com>
15164 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
15166 * configure: Regenerate.
15168 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
15170 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
15171 splitter preparation statements.
15172 * config/i386/sse.md (*vec_extract* splitters): Ditto.
15173 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
15176 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15178 * gimple-ssa-strength-reduction.c (count_candidates): Change
15179 return value to int.
15180 (analyze_candidates_and_replace): Change type of length to int.
15182 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
15184 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
15185 (*vec_extract<mode>): Use VI12_128 mode iterator.
15186 (*vec_extract<mode>_mem): Ditto.
15187 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
15190 2013-05-08 Diego Novillo <dnovillo@google.com>
15195 2012-08-17 Diego Novillo <dnovillo@google.com>
15198 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
15199 * config.in: Regenerate.
15200 * configure: Regenerate.
15201 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
15203 2013-05-08 Jan Hubicka <jh@suse.cz>
15206 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
15207 * cgraph.h (symtab_node_base): Add unique_name.
15208 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
15209 input_overwrite_node, input_varpool_node): Stream unique_name.
15210 * cgraphclones.c (cgraph_create_virtual_clone,
15211 cgraph_function_versioning): Set unique_name.
15212 * ipa.c (function_and_variable_visibility): Set unique_name.
15214 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15216 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
15217 (alloc_cand_and_find_basis): Restrict conditional candidate
15218 processing to CAND_MULTs.
15220 2013-05-08 Jan Hubicka <jh@suse.cz>
15223 lto-symtab.c (lto_symtab_symbol_p): New function.
15224 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
15225 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
15226 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
15227 Skip static symbols.
15229 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
15231 PR tree-optimization/57200
15232 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
15233 Only call inform if the preceding warning_at returns true.
15235 2013-05-07 Han Shen <shenhan@google.com>
15237 * cfgexpand.c (record_or_union_type_has_array_p): New function.
15238 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
15239 * common.opt (fstack-protector-strong): New option.
15240 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
15241 * doc/invoke.texi (Optimization Options): Document
15242 "-fstack-protector-strong".
15243 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
15245 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
15247 * config/mips/mips.c (mips_machine_reorg2): Return 0.
15249 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
15251 * ira.c (update_equiv_regs): Add insn having equiv memory even if
15252 it is not lhs of the insn.
15253 (setup_reg_equiv): Remove insn having equiv memory which it is not
15255 * lra-constraints.c (process_address): Try to improve generation
15256 code for address base + disp.
15257 (lra_constraints): Make correct the code for checking insn setting
15258 up backward equivalence. Remove insn only if it is in the init
15260 * lra-eliminations.c (update_reg_eliminate): Change return value.
15261 (lra_eliminate): Use the result.
15263 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
15265 * config/i386/sse.md (ssescalarnummask): New mode attribute.
15266 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
15267 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
15268 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
15269 register target operands.
15270 (*vec_extractv8hi_sse2): New pattern.
15271 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
15272 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
15273 (*vec_extract<mode>_mem): New insn and split pattern.
15275 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
15277 * config/arm/arm.c (arm_asan_shadow_offset): New function.
15278 (TARGET_ASAN_SHADOW_OFFSET): Define.
15279 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
15280 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
15282 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15284 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
15285 (incr_vec_index): Return -1 if increment not found.
15286 (create_add_on_incoming_edge): Assert if increment not found.
15287 (record_increment): Limit number of increments recorded.
15288 (all_phi_incrs_profitable): Return false if an increment not found.
15289 (replace_profitable_candidates): Don't process increments that were
15291 (analyze_candidates_and_replace): Limit size of incr_vec.
15293 2013-05-07 Richard Biener <rguenther@suse.de>
15295 * calls.c (special_function_p): setjmp-like functions are leaf.
15296 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
15297 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
15299 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
15301 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
15302 (aarch64_simd_mov<mode>): New expander.
15303 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
15304 (aarch64_simd_mov_to_<mode>high): Likewise.
15305 (aarch64_simd_mov_from_<mode>low): Likewise.
15306 (aarch64_simd_mov_from_<mode>high): Likewise.
15307 (aarch64_dup_lane<mode>): Update.
15308 (aarch64_dup_lanedi): New instruction pattern.
15309 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
15310 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
15312 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15314 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
15315 (replace_mult_candidate): Remove unnecessary argument; remove
15316 unnecessary parameter from call to introduce_cast_before_cand.
15317 (replace_unconditional_candidate): Remove unnecessary parameter
15318 from call to replace_mult_candidate.
15319 (replace_conditional_candidate): Likewise.
15320 (insert_initializers): Use make_temp_ssa_name.
15321 (introduce_cast_before_cand): Remove unnecessary argument; use
15322 make_temp_ssa_name.
15323 (replace_one_candidate): Remove unnecessary argument; remove
15324 unnecessary parameter from calls to introduce_cast_before_cand.
15325 (replace_profitable_candidates): Remove unnecessary parameters
15326 from calls to replace_one_candidate.
15328 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15330 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
15331 phi def as possibly hiding a basis for a CAND_ADD whose operands
15332 have been commuted in the analysis.
15333 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
15335 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15337 * config/aarch64/aarch64.md
15338 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
15339 shift value between 0-4.
15341 2013-05-07 Richard Biener <rguenther@suse.de>
15343 * double-int.h (rshift): New overload.
15344 * double-int.c (rshift): New function.
15345 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
15346 (create_reference_ops_from_ref): Remove.
15347 (vn_reference_insert): Use shared ops for constructing the
15348 reference and copy it.
15350 2013-05-07 Richard Biener <rguenther@suse.de>
15352 PR middle-end/57190
15353 * tree-eh.c (sink_clobbers): Properly propagate
15354 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
15356 2013-05-07 Jakub Jelinek <jakub@redhat.com>
15358 PR tree-optimization/57149
15359 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
15360 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
15361 collect_phi_def_edges, execute_late_warn_uninitialized): Use
15362 uninit_undefined_value_p instead of ssa_undefined_value_p.
15365 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
15366 for modifier == EXPAND_INITIALIZER.
15368 2013-05-07 Anton Blanchard <anton@samba.org>
15370 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
15371 for powerpc64 little endian.
15372 * configure: Regenerate.
15374 2013-05-06 Graham Stott <grahams@btinternet.com>
15376 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
15377 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
15378 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
15379 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
15381 2013-05-06 Graham Stott <grahams@btinternet.com>
15383 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
15384 codes which allow non-lvalues.
15386 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
15388 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
15389 components are all 1s.
15390 (integer_minus_onep): New function.
15391 * tree.h (integer_minus_onep): Declare it.
15392 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
15393 integer_minus_onep instead of integer_all_onesp.
15395 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
15398 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
15399 variations of these patterns.
15401 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
15403 * config/i386/i386.md (isa): Add x64_sse4 member.
15404 (enabled): Handle x64_sse4.
15405 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
15406 instruction for 64bit SSE4_1 targets. Update insn attributes.
15407 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
15408 instruction for SSE4_1 targets. Update insn attributes.
15409 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
15410 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
15411 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
15413 (*vec_extractv4si): Rename from *sse4_1_pextrd.
15414 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
15415 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
15417 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
15420 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
15422 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
15424 * final.c (do_assembler_dialects): Don't handle curly braces and
15425 vertical bar escaped by % as dialect delimiters.
15426 (output_asm_insn): Print curly braces and vertical bar if escaped
15427 by % and ASSEMBLER_DIALECT defined.
15428 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
15429 * doc/tm.texi: Regenerated.
15431 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
15433 * config/mips/mips.c: Include tree-pass.h.
15434 (mips_reorg): Split in pre- and post-dbr_schedule parts.
15435 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
15436 (pass_mips_machine_reorg2): New machine specific pass.
15437 (insert_pass_mips_machine_reorg2): New pass plugin definition.
15438 (mips_option_override): Register the new pass.
15439 * rtl.h (cleanup_barriers): Remove prototype.
15440 (dbr_schedule): Likewise.
15441 * jump.c (cleanup_barriers): Make static.
15442 * reorg.c (dbr_schedule): Likewise.
15444 2013-05-06 Richard Biener <rguenther@suse.de>
15446 PR tree-optimization/57185
15447 * tree-parloops.c (add_field_for_reduction): Handle anonymous
15448 SSA names properly.
15450 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
15453 * config/i386/i386.c (add_parameter_dependencies): Add dependence
15454 between "first_arg" and "insn", not "last" and "insn".
15456 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15458 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
15459 (find_candidates_in_block): Re-enable slsr_process_phi.
15460 (create_phi_basis): Fix double counting of candidate adjustment.
15462 2013-05-06 Richard Biener <rguenther@suse.de>
15464 PR middle-end/57147
15465 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
15466 the edge is also fallthru, preserve it and just clear the
15468 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
15469 also complex, preserve that and just clear the fallthru flag.
15470 * tree-inline.c (update_ssa_across_abnormal_edges): Also
15471 update virtual operands.
15473 2013-05-06 Alan Modra <amodra@gmail.com>
15475 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
15476 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
15477 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
15478 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
15479 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
15480 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
15482 2013-05-06 Alan Modra <amodra@gmail.com>
15484 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
15485 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
15486 (DEFAULT_ASM_ENDIAN): Define.
15487 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
15488 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
15489 Update -K PIC clause from sysv4.h.
15490 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
15491 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
15493 2013-05-06 Alan Modra <amodra@gmail.com>
15495 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
15496 twice for little-endian.
15497 (ashrdi3_no_power, ashrdi3): Support little-endian.
15499 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
15502 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
15503 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
15505 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
15506 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
15507 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
15510 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
15511 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
15513 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
15514 * config/pa/pa.opt: Make mbig-switch a no-op.
15515 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
15516 (CASE_VECTOR_MODE): Always return SImode.
15517 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
15518 for the !TARGET_BIG_SWITCH case.
15519 * config/pa/pa-linux.h: Likewise.
15520 * config/pa/pa-openbsd.h: Likewise.
15521 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
15522 * config/pa/pa.md (short_jump): Remove define_insn.
15523 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
15524 (casesi0): Remove define_insn.
15525 (type): Remove btable_branch.
15526 (pa_combine_type): Likewise.
15527 (in_nullified_branch_delay): Likewise.
15528 (in_call_delay): Likewise.
15529 (define_delay): Likewise.
15530 (define_insn_reservation "Z3"): Likewise.
15531 (define_insn_reservation "Z4"): Likewise.
15532 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
15533 (pa_adjust_insn_length): Remove adjustment for btable branches.
15534 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
15537 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
15539 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
15540 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
15541 Add m->r,x alternatives.
15542 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
15543 splitters using SWI48x mode iterator.
15544 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
15545 TARGET_64BIT. Add m->x alternative.
15546 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
15547 Add o->x alternative. Enable for TARGET_SSE.
15548 (sse_storeq): Remove expander.
15549 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
15550 with memory input operand.
15551 (*vec_extractv2di_1 splitter): New.
15552 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
15553 * config/i386/i386.md (ssevecmodelower): New mode attribute.
15555 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
15557 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
15558 (INT_LOWPART): Delete.
15559 (extract_MB): Adjust.
15560 (extract_ME): Adjust.
15561 (print_operand): Adjust.
15563 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
15565 * config/rs6000/predicates.md (reg_or_add_cint_operand,
15566 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
15567 (reg_or_logical_cint_operand, easy_fp_constant,
15568 logical_const_operand): Delete "CONST_DOUBLE" case.
15569 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
15570 "HOST_BITS_PER_WIDE_INT == 64" test.
15571 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
15572 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
15573 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
15574 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
15576 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
15577 CONST_DOUBLE DImode/VOIDmode case.
15578 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
15579 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
15580 CONST_DOUBLE VOIDmode case.
15581 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
15582 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
15583 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
15584 Delete CONST_DOUBLE case.
15585 (splitters for mov FMOVE64 const_double): Delete
15586 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
15587 "HOST_BITS_PER_WIDE_INT >= 64" test.
15588 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
15590 (mov DI const_double): Delete.
15592 2013-05-04 Jakub Jelinek <jakub@redhat.com>
15594 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
15595 on op shows all bits zero in mode of a lowpart subreg, return zero.
15597 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
15600 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
15601 to save TFmode registers and DImode to save TImode registers for
15602 caller save operations.
15603 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
15604 mark being partially clobbered since they only use the first
15607 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
15608 and TDmode only use the upper 64-bits of each VSX register.
15610 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15612 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
15613 (find_candidates_in_block): Disable slsr_process_phi.
15615 2013-05-03 Guozhi Wei <carrot@google.com>
15617 * coverage.c (coverage_obj_init): Move the construction of gcov
15619 (build_init_ctor): ... here.
15621 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15623 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
15624 (slsr_cand_d): Redefine def_phi.
15625 (stride_status, phi_adjust_status, count_phis_status): New enums.
15626 (find_phi_def): New.
15627 (find_basis_for_base_expr): New.
15628 (find_basis_for_candidate): Handle hidden bases.
15629 (alloc_cand_and_find_basis): Handle phi candidates.
15630 (slsr_process_phi): New.
15631 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
15632 (create_mul_imm_cand): Likewise.
15633 (create_add_ssa_cand): Exclude phi base candidates.
15634 (create_add_imm_cand): Likewise.
15635 (slsr_process_cast): Likewise.
15636 (slsr_process_copy): Likewise.
15637 (find_candidates_in_block): Handle phi candidates.
15638 (dump_candidate): Likewise.
15639 (unconditional_cands): Delete.
15640 (unconditional_cands_with_known_stride_p): Delete.
15641 (phi_dependent_cand_p): New.
15642 (cand_increment): Handle phi-dependent candidates.
15643 (replace_dependent): Delete.
15644 (replace_mult_candidate): New.
15645 (replace_unconditional_candidate): New.
15646 (incr_vec_index): Move to avoid forward reference.
15647 (create_add_on_incoming_edge): New.
15648 (create_phi_basis): New.
15649 (replace_dependents): Delete.
15650 (replace_conditional_candidate): New.
15651 (phi_add_costs): New.
15652 (replace_uncond_cands_and_profitable_phis): New.
15653 (record_increment): Handle phi adjustments.
15654 (record_phi_increments): New.
15655 (record_increments): Handle phi adjustments.
15656 (phi_incr_cost): New.
15657 (lowest_cost_path): Handle phis.
15658 (total_savings): Likewise.
15659 (analyze_increments): Likewise.
15660 (ncd_with_phi): New.
15661 (ncd_of_cand_and_phis): New.
15662 (nearest_common_dominator_for_cands): Handle phi increments.
15663 (all_phi_incrs_profitable): New.
15664 (replace_profitable_candidates): Handle phi-dependent candidates.
15665 (analyze_candidates_and_replace): Likewise.
15667 2013-05-03 Teresa Johnson <tejohnson@google.com>
15670 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
15671 do not exceed REG_BR_PROB_BASE.
15673 2013-05-03 Jeff Law <law@redhat.com>
15675 PR tree-optimization/57144
15676 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
15677 operand of the condition will bit into the new type when eliminating
15678 a cast feeding a condition.
15680 2013-05-03 Jakub Jelinek <jakub@redhat.com>
15682 PR rtl-optimization/57130
15683 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
15684 of COMPARE as in_code to the recursive call if needed.
15686 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
15688 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
15689 (enabled): Handle new members.
15690 * config/i386/sse.md (*vec_concatv2si): Merge from
15691 *vec_concatv2si_sse2 and vec_concatv2si_sse.
15692 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
15694 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
15696 PR tree-optimization/57027
15697 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
15698 for fnms opportunity, check we got the prerequisite kind
15699 of tree / gimple before using accessor functions.
15701 2013-05-03 Richard Biener <rguenther@suse.de>
15703 * double-int.h (lshift): New overload without precision
15704 and arith argument.
15705 (operator *=, operator +=, operator -=): Move ...
15706 * double-int.c (operator *=, operator +=, operator -=): ... here
15707 and implement more efficiently.
15708 (mul_double_with_sign): Remove.
15709 (lshift_double): Adjust to take unsinged shift argument, push
15710 dispatching code to callers.
15711 (mul_double_wide_with_sign): Add early out for callers that
15712 are not interested in high parts or overflow.
15713 (lshift): New function.
15714 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
15715 dispatch code here.
15716 (lrotate, rrotate): Use logical shifts.
15717 * expr.c (get_inner_reference): Use lshift.
15718 * fixed-value.c (do_fixed_divide): Likewise.
15719 * tree-dfa.c (get_ref_base_and_extent): Likewise.
15720 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
15721 (indirect_refs_may_alias_p): Likewise.
15722 (stmt_kills_ref_p_1): Likewise.
15724 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
15726 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
15728 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
15730 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
15731 scalar form of FABD instruction.
15733 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
15735 * lra-constraints.c (process_alt_operands): Add checking alt
15736 number to choose the best alternative.
15738 2013-05-02 Richard Biener <rguenther@suse.de>
15740 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
15741 bitmap and its handling.
15742 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
15744 2013-05-02 Richard Biener <rguenther@suse.de>
15746 PR middle-end/57140
15747 * tree-inline.c (copy_loops): Properly handle removed loops.
15748 (copy_cfg_body): Mark destination loops for fixup if source
15749 loops needed fixup.
15751 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
15754 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
15755 generating simple_return for naked functions.
15757 2013-05-02 Martin Jambor <mjambor@suse.cz>
15759 PR middle-end/56988
15760 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
15761 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
15763 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
15764 ipa_agg_replacement_value structures.
15765 (known_aggs_to_agg_replacement_list): Likewise.
15766 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
15767 (read_agg_replacement_chain): Likewise.
15768 (ipcp_transform_function): Also check that by_ref flags match.
15770 2013-05-02 Richard Biener <rguenther@suse.de>
15772 * graphds.h (struct graph): Add obstack member.
15773 * graphds.c (new_graph): Initialize obstack and allocate
15775 (add_edge): Allocate edge from the obstack.
15776 (free_graph): Free the obstack instead of all edges and vertices.
15778 2013-05-02 Teresa Johnson <tejohnson@google.com>
15780 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
15782 * cfg.c (update_bb_profile_for_threading): Ditto.
15783 * tree-inline.c (copy_bb): Ditto.
15784 (copy_edges_for_bb): Ditto.
15785 (initialize_cfun): Ditto.
15786 (copy_cfg_body): Ditto.
15787 (expand_call_inline): Ditto.
15788 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
15789 (estimate_node_size_and_time): Ditto.
15790 (inline_merge_summary): Ditto.
15791 * cgraphclones.c (cgraph_clone_edge): Ditto.
15792 (cgraph_clone_node): Ditto.
15793 * sched-rgn.c (compute_dom_prob_ps): Ditto.
15794 (compute_trg_info): Ditto.
15796 2013-05-02 Ian Bolton <ian.bolton@arm.com>
15798 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
15799 S reg when fp attribute set.
15800 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
15802 2013-05-02 Ian Bolton <ian.bolton@arm.com>
15804 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
15806 (*and_one_cmplsi3_compare0_uxtw): Likewise.
15807 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
15808 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
15810 2013-05-02 Richard Biener <rguenther@suse.de>
15812 * tree-scalar-evolution.c (scev_info_hasher): Remove.
15813 (struct instantiate_cache_entry): New type.
15814 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
15815 (struct instantiate_cache_type): New type.
15816 (set_instantiated_value, get_instantiated_value): Remove.
15817 (get_instantiated_value_entry): New function.
15818 (instantiate_scev_name): Use the new cache and adjust.
15819 (instantiate_scev_poly): Adjust.
15820 (instantiate_scev_binary): Likewise.
15821 (instantiate_array_ref): Likewise.
15822 (instantiate_scev_convert): Likewise.
15823 (instantiate_scev_not): Likewise.
15824 (instantiate_scev_3): Likewise.
15825 (instantiate_scev_2): Likewise.
15826 (instantiate_scev_r): Likewise.
15827 (instantiate_scev): Likewise.
15828 (resolve_mixers): Likewise.
15830 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
15833 * lra-constraints.c (best_small_class_operands_num): Remove.
15834 (process_alt_operands): Remove small_class_operands_num. Take
15835 small classes operands into losers and only if the operand is not
15836 matched. Modify debugging output.
15837 (curr_insn_transform): Remove best_small_class_operands_num.
15840 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15842 * config/aarch64/aarch64-builtins.c
15843 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
15844 * config/aarch64/aarch64-simd-builtins.def
15845 (reduc_splus_): Add new modes.
15846 (reduc_uplus_): New.
15847 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
15848 (reduc_uplus_v4sf): Likewise.
15849 (reduc_splus_v4sf): Likewise.
15850 (aarch64_addv<mode>): Likewise.
15851 (reduc_uplus_<mode>): Likewise.
15852 (reduc_splus_<mode>): Likewise.
15853 (aarch64_addvv2di): Likewise.
15854 (reduc_uplus_v2di): Likewise.
15855 (reduc_splus_v2di): Likewise.
15856 (aarch64_addvv2si): Likewise.
15857 (reduc_uplus_v2si): Likewise.
15858 (reduc_splus_v2si): Likewise.
15859 (reduc_<sur>plus_<mode>): New.
15860 (reduc_<sur>plus_v2di): Likewise.
15861 (reduc_<sur>plus_v2si): Likewise.
15862 (reduc_<sur>plus_v4sf): Likewise.
15863 (aarch64_addpv4sf): Likewise.
15864 * config/aarch64/arm_neon.h
15865 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
15866 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
15867 add UNSPEC_SADDV, UNSPEC_UADDV.
15869 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
15871 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15873 * config/aarch64/arm_neon.h
15874 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
15876 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15878 * config/aarch64/aarch64-builtins
15879 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
15881 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15883 * config/aarch64/aarch64-simd-builtins.def
15884 (reduc_smax_): New.
15885 (reduc_smin_): Likewise.
15886 (reduc_umax_): Likewise.
15887 (reduc_umin_): Likewise.
15888 (reduc_smax_nan_): Likewise.
15889 (reduc_smin_nan_): Likewise.
15892 (smax): Update for V2SF, V4SF and V2DF modes.
15895 (smin_nan): Likewise.
15896 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
15897 (<su><maxmin><mode>3): ...This, refactor.
15898 (s<maxmin><mode>3): New.
15899 (<maxmin_uns><mode>3): Likewise.
15900 (reduc_<maxmin_uns>_<mode>): Refactor.
15901 (reduc_<maxmin_uns>_v4sf): Likewise.
15902 (reduc_<maxmin_uns>_v2si): Likewise.
15903 (aarch64_<fmaxmin><mode>: Remove.
15904 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
15906 (vmin<q>_f<32,64>): Likewise.
15907 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
15909 (su): Add mappings for smax, smin, umax, umin.
15911 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
15912 (FMAXMIN): Rename as...
15913 (FMAXMIN_UNS): ...This.
15915 (fmaxminv): Likewise.
15916 (fmaxmin): Likewise.
15918 (maxmin_uns_op): Likewise.
15920 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15922 * config/aarch64/arm_neon.h
15923 (vac<ge, gt><sd>_f<32, 64>): Rename to...
15924 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
15925 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
15927 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15929 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
15930 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
15932 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15934 * config/aarch64/aarch64-simd.md
15935 (vcond<mode>_internal): Handle special cases for constant masks.
15936 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
15937 (vcondu<mode><mode>): Likewise.
15938 (vcond<v_cmp_result><mode>): New.
15940 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15942 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
15943 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
15944 * config/aarch64/aarch64-simd-builtins.def
15945 (cmeq): Update to BUILTIN_VALLDI.
15950 * config/aarch64/arm_neon.h
15951 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
15952 to builtins or C as appropriate.
15954 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15956 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
15958 (cmhi): Rename to...
15960 * config/aarch64/aarch64-simd.md
15961 (simd_mode): Add SF.
15962 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
15963 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
15964 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
15965 (cstore<mode>_neg): ...This.
15966 * config/aarch64/iterators.md
15968 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
15969 (COMPARISONS): New.
15970 (UCOMPARISONS): Likewise.
15971 (optab): Add missing comparisons.
15977 (VCMP_S): Likewise.
15978 (VCMP_U): Likewise.
15979 (V_cmp_result): Add DF, SF modes.
15980 (v_cmp_result): Likewise.
15982 (vmtype): Likewise.
15983 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
15985 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
15987 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
15988 define_insn to define_insn_and_split.
15989 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
15990 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
15991 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
15992 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
15993 (thumb2_negscc): Likewise.
15995 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
15997 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
15999 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
16001 * config/arm/thumb2.md: Remove trailing whitespaces.
16003 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
16005 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
16006 Use gen_int_mode rather than GEN_INT.
16008 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
16010 * value-prof.c (stream_in_histogram_value): Remove the strayed
16013 2013-04-30 Richard Biener <rguenther@suse.de>
16015 PR middle-end/57122
16016 * cfghooks.c (split_edge): Properly check for the loop latch edge.
16018 2013-04-30 Richard Biener <rguenther@suse.de>
16020 PR middle-end/57107
16021 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
16023 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
16025 PR rtl-optimization/56957
16026 PR rtl-optimization/57105
16027 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
16028 variable. Use just INSN_UID for determining whether an insn
16029 should be only disconnected from the insn stream.
16030 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
16032 2013-04-30 Jakub Jelinek <jakub@redhat.com>
16034 PR tree-optimization/57104
16035 * tsan.c (instrument_expr): Don't instrument accesses to
16036 DECL_HARD_REGISTER VAR_DECLs.
16038 2013-04-30 Richard Biener <rguenther@suse.de>
16040 * function.h (loops_for_fn): New inline function.
16041 (set_loops_for_fn): Likewise.
16042 * cfgloop.h (place_new_loop): Add struct function parameter.
16043 (get_loop): Likewise.
16044 (get_loops): Likewise.
16045 (number_of_loops): Likewise.
16046 (fel_next): Adjust.
16047 (fel_init): Likewise.
16048 * cfg.c (get_loop_copy): Adjust.
16049 * cfgloop.c (flow_loops_dump): Likewise.
16050 (record_loop_exits): Likewise.
16051 (verify_loop_structure): Likewise.
16052 * cfgloopanal.c (mark_irreducible_loops): Likewise.
16053 (estimate_reg_pressure_cost): Likewise.
16054 (mark_loop_exit_edges): Likewise.
16055 * cfgloopmanip.c (place_new_loop): Likewise.
16056 (add_loop): Likewise.
16057 (duplicate_loop): Likewise.
16058 * graph.c (draw_cfg_nodes): Likewise.
16059 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
16060 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
16061 (extract_affine_chrec): Likewise.
16062 (build_scop_iteration_domain): Likewise.
16063 * graphite.c (graphite_initialize): Likewise.
16064 * ira-build.c (create_loop_tree_nodes): Likewise.
16065 (more_one_region_p): Likewise.
16066 (rebuild_regno_allocno_maps): Likewise.
16067 (mark_loops_for_removal): Likewise.
16068 (mark_all_loops_for_removal): Likewise.
16069 (remove_unnecessary_regions): Likewise.
16070 (ira_build): Likewise.
16071 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
16072 * loop-init.c (fix_loop_structure): Likewise.
16073 (gate_rtl_move_loop_invariants): Likewise.
16074 (gate_rtl_unswitch): Likewise.
16075 (gate_rtl_unroll_and_peel_loops): Likewise.
16076 (rtl_doloop): Likewise.
16077 * lto-streamer-in.c (input_cfg): Likewise.
16078 * lto-streamer-out.c (output_cfg): Likewise.
16079 * modulo-sched.c (sms_schedule): Likewise.
16080 * predict.c (tree_estimate_probability): Likewise.
16081 (tree_estimate_probability_driver): Likewise.
16082 (estimate_loops): Likewise.
16083 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
16084 (move_sese_region_to_fn): Likewise.
16085 (debug_loop_num): Likewise.
16086 * tree-chrec.c (chrec_evaluate): Likewise.
16087 (hide_evolution_in_other_loops_than_loop): Likewise.
16088 (chrec_component_in_loop_num): Likewise.
16089 (reset_evolution_in_loop): Likewise.
16090 (evolution_function_is_invariant_rec_p): Likewise.
16091 * tree-if-conv.c (main_tree_if_conversion): Likewise.
16092 * tree-inline.c (copy_loops): Likewise.
16093 (copy_cfg_body): Likewise.
16094 (tree_function_versioning): Likewise.
16095 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
16096 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
16098 (add_to_evolution_1): Likewise.
16099 (scev_const_prop): Likewise.
16100 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
16101 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
16102 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
16103 (tree_ssa_lim_initialize): Likewise.
16104 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
16105 (verify_loop_closed_ssa): Likewise.
16106 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
16107 (tree_ssa_loop_im): Likewise.
16108 (tree_ssa_loop_unswitch): Likewise.
16109 (tree_vectorize): Likewise.
16110 (check_data_deps): Likewise.
16111 (tree_ssa_loop_ivcanon): Likewise.
16112 (tree_ssa_loop_bounds): Likewise.
16113 (tree_complete_unroll): Likewise.
16114 (tree_complete_unroll_inner): Likewise.
16115 (tree_parallelize_loops): Likewise.
16116 (tree_ssa_loop_prefetch): Likewise.
16117 (tree_ssa_loop_ivopts): Likewise.
16118 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
16119 * tree-vectorizer.c (vectorize_loops): Likewise.
16121 2013-04-29 Mike Frysinger <vapier@gentoo.org>
16123 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
16124 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
16125 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
16126 with EABI_LINK_SPEC.
16128 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
16131 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
16134 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
16137 * lra-constraints.c (process_alt_operands): Discourage a bit more
16138 using memory for pseudos. Print cost dump for alternatives.
16139 Modify cost values for conflicts with early clobbers.
16140 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
16142 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
16145 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
16147 2013-04-29 Ian Bolton <ian.bolton@arm.com>
16149 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
16150 from/to S register.
16151 (movdi_aarch64): Support LDR/STR from/to D register.
16153 2013-04-29 Ian Bolton <ian.bolton@arm.com>
16155 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
16156 or higher by default.
16158 2013-04-29 Richard Biener <rguenther@suse.de>
16160 PR middle-end/57075
16161 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
16162 even if not adding abnormal edges for calls that can make
16165 2013-04-29 Richard Biener <rguenther@suse.de>
16167 PR middle-end/57103
16168 * tree-cfg.c (move_stmt_op): Fix condition under which to update
16170 (move_stmt_r): Remove redundant checking.
16172 2013-04-29 Teresa Johnson <tejohnson@google.com>
16175 * basic-block.h (apply_scale): New function.
16176 (apply_probability): Use apply_scale.
16177 * gimple-streamer-in.c (input_bb): Ditto.
16178 * lto-streamer-in.c (input_cfg): Ditto.
16179 * lto-cgraph.c (merge_profile_summaries): Ditto.
16180 * tree-optimize.c (execute_fixup_cfg): Ditto.
16181 * tree-inline.c (copy_bb): Update comment to use apply_scale.
16182 (copy_edges_for_bb): Ditto.
16183 (copy_cfg_body): Ditto.
16185 2013-04-29 Tom de Vries <tom@codesourcery.com>
16187 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
16188 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
16189 (tail_merge_optimize): Handle current_loops == NULL.
16191 2013-04-26 Jeff Law <law@redhat.com>
16193 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
16194 (simplify_cond_using_ranges): Generalize code to simplify
16195 COND_EXPRs where one argument is a constant and the other
16196 is an SSA_NAME created by an integral type conversion.
16198 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16200 * config/arm/arm.md (store_minmaxsi): Use only when
16201 optimize_insn_for_size_p.
16203 2013-04-29 Christian Bruel <christian.bruel@st.com>
16206 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
16208 2013-04-29 Richard Biener <rguenther@suse.de>
16210 PR middle-end/57089
16211 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
16212 loop tree make sure to schedule a fixup for the child as well.
16213 (expand_omp_for_generic): Properly add loops.
16214 (expand_omp_for_static_nochunk): Likewise.
16215 (expand_omp_for_static_chunk): Likewise.
16216 (expand_omp_for): For the degenerate case fixup loops.
16217 (expand_omp_sections): Fix default bb placement in loops.
16218 (expand_omp_atomic_pipeline): Properly add loops.
16220 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16222 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
16224 2013-04-29 Tom de Vries <tom@codesourcery.com>
16226 * tree-ssa-tail-merge.c: Update header comment.
16228 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16230 * config/aarch64/arm_neon.h
16231 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
16232 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
16233 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
16234 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
16235 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
16236 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
16237 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
16238 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
16240 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16242 * config/aarch64/aarch64-simd.md
16243 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
16244 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
16245 fix_trunc, fixuns_trunc.
16246 (ftrunc<VDQF:mode>2): New.
16247 * config/aarch64/iterators.md (optab): Add fix, fixuns.
16248 (fix_trunc_optab): New.
16250 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16252 * config/aarch64/aarch64-builtins.c
16253 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
16254 iceilf, lround, iroundf.
16256 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
16259 * config/i386/i386.h (enum ix86_tune_indices)
16260 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
16261 New, split from X86_TUNE_INTER_UNIT_MOVES.
16262 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
16263 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
16264 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
16265 (TARGET_INTER_UNIT_MOVES): Remove.
16266 * config/i386/i386.c (initial_ix86_tune_features): Update.
16267 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
16268 (ix86_expand_convert_uns_didf_sse): Use
16269 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
16270 (ix86_expand_vector_init_one_nonzero): Ditto.
16271 (ix86_expand_vector_init_interleave): Ditto.
16272 (inline_secondary_memory_needed): Return true for moves from SSE class
16273 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
16274 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
16275 * config/i386/constraints.md (Yi, Ym): Depend on
16276 TARGET_INTER_UNIT_MOVES_TO_VEC.
16277 (Yj, Yn): New constraints.
16278 * config/i386/i386.md (*movdi_internal): Change constraints of
16279 operand 1 from Yi to Yj and from Ym to Yn.
16280 (*movsi_internal): Ditto.
16281 (*movdf_internal): Ditto.
16282 (*movsf_internal): Ditto.
16283 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
16284 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
16285 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
16286 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
16287 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
16288 * config/i386/sse.md (movdi_to_sse): Ditto.
16289 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
16290 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
16291 TARGET_INTER_UNIT_MOVES.
16292 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
16293 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
16294 instead of TARGET_INTER_UNIT_MOVES.
16295 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
16296 operand 1 from Yi to Yj and from Ym to Yn.
16298 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16300 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
16301 (float_truncate_hi_): Likewise.
16302 (float_extend_lo_): Likewise.
16303 (float_truncate_lo_): Likewise.
16304 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
16305 (aarch64_float_extend_lo_v2df): Likewise.
16306 (vec_unpacks_hi_v4sf): Likewise.
16307 (aarch64_float_truncate_lo_v2sf): Likewise.
16308 (aarch64_float_truncate_hi_v4sf): Likewise.
16309 (vec_pack_trunc_v2df): Likewise.
16310 (vec_pack_trunc_df): Likewise.
16312 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16314 * config/aarch64/aarch64-builtins.c
16315 (aarch64_fold_builtin): Fold float conversions.
16316 * config/aarch64/aarch64-simd-builtins.def
16317 (floatv2si, floatv4si, floatv2di): New.
16318 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
16319 * config/aarch64/aarch64-simd.md
16320 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
16321 * config/aarch64/iterators.md (FLOATUORS): New.
16322 (optab): Add float, floatuns.
16323 (su_optab): Likewise.
16325 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16327 * config/aarch64/aarch64-builtins.c
16328 (aarch64_builtin_vectorized_function): Use new names for
16330 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
16331 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
16332 (fcvtzu): Split as...
16333 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
16334 (fcvtas): Split as...
16335 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
16336 (fcvtau): Split as...
16337 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
16338 (fcvtps): Split as...
16339 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
16340 (fcvtpu): Split as...
16341 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
16342 (fcvtms): Split as...
16343 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
16344 (fcvtmu): Split as...
16345 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
16346 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
16347 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
16348 (lfrintnusf, lfrintnudf): Likewise.
16349 * config/aarch64/aarch64-simd.md
16350 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
16352 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
16353 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
16354 (fcvt_pattern): Likewise.
16356 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16358 * config/aarch64/aarch64-simd.md
16359 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
16360 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
16362 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16364 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
16365 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
16366 (vrnd<a,m,n,p>_f32): Implement using builtins.
16367 (vrnd<i,x><q>_f<32, 64>): New.
16369 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16371 * config/aarch64/aarch64-builtins.c
16372 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
16373 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
16374 (frintz): Rename to...
16376 (frintp): Rename to...
16378 (frintm): Rename to...
16380 (frinti): Rename to...
16381 (nearbyint): ...this.
16382 (frintx): Rename to...
16384 (frinta): Rename to...
16386 * config/aarch64/aarch64-simd.md
16387 (aarch64_frint<frint_suffix><mode>): Delete.
16388 (<frint_pattern><mode>2): Convert to insn.
16389 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
16390 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
16391 (frint_pattern): Likewise.
16392 (frint_suffix): Likewise.
16394 2013-04-29 Richard Biener <rguenther@suse.de>
16396 PR tree-optimization/57081
16397 * loop-init.c: Include tree-flow.h.
16398 (loop_optimizer_finalize): Free number of iteration estimates.
16399 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
16401 2013-04-29 Jakub Jelinek <jakub@redhat.com>
16403 PR tree-optimization/57083
16404 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
16405 non-singleton shift count range, zero extend low_bound for uns case.
16407 * config/i386/predicates.md (general_vector_operand): New predicate.
16408 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
16409 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
16410 if they aren't nonimmediate operands. If their original values
16411 satisfy const_vector_equal_evenodd_p, don't shift them.
16412 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
16413 predicates. For the SSE4.1 case force operands[{1,2}] into registers
16414 if not nonimmediate_operand.
16415 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
16416 instead of register_operand.
16417 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
16419 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
16421 * stor-layout.c (finalize_size_functions): Allocate a structure and
16422 reset cfun before dumping the functions.
16424 2013-04-27 Jakub Jelinek <jakub@redhat.com>
16426 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
16429 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
16430 use xop_pmacsdqh if uns_p.
16431 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
16432 the immediate rotate count.
16434 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
16436 * rtl.h (struct rtx_def): Add comment for field jump.
16437 (LRA_SUBREG_P): New macro.
16438 * recog.c (register_operand): Check LRA_SUBREG_P.
16439 * lra.c (lra): Add note at the end of RTL code. Align non-empty
16441 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
16442 (lra_final_code_change): Skip subreg change for operators.
16443 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
16444 if there are no operand changes.
16445 * lra-constraints.c (curr_insn_set): New.
16446 (match_reload): Set LRA_SUBREG_P.
16447 (emit_spill_move): Ditto.
16448 (check_and_process_move): Use curr_insn_set. Process only single
16449 set insns. Don't initialize sec_mem_p and change_p.
16450 (simplify_operand_subreg): Use LRA_SUBREG_P.
16451 (reg_in_class_p): New function.
16452 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
16453 of #ifdef. Add code to remove cycling.
16454 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
16455 non-null disp. Reload inner instead of disp when base and index
16456 are null. Try to put lo_sum into register.
16457 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
16458 (check_and_process_move): Move code for move cost check to
16459 simple_move_p. Remove equiv_substitution.
16460 (simple_move_p): New function.
16461 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
16462 curr_insn_set. Call check_and_process_move only for single set
16463 insns. Use the new function. Move call of check_and_process_move
16464 after operand equiv substitution and address process.
16466 2013-04-26 Jakub Jelinek <jakub@redhat.com>
16469 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
16470 with nonlocal goto receivers or returns twice calls, ignore
16471 unininitialized values from abnormal edges to nl goto receiver
16472 or returns twice call.
16474 2013-04-26 Jakub Jelinek <jakub@redhat.com>
16476 PR tree-optimization/57051
16477 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
16478 and VEC_RSHIFT_EXPR if shift count is a multiple of element
16481 2013-04-26 Richard Biener <rguenther@suse.de>
16483 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
16484 (expand_omp_taskreg): Likewise. Mark loops for fixup.
16485 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
16486 (fixup_loop_arrays_after_move): New function.
16487 (move_sese_region_to_fn): Properly outline the loop tree parts
16488 of the SESE region.
16490 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
16492 * config/i386/i386.md (type, unit): Fix long lines.
16494 2013-04-26 Richard Biener <rguenther@suse.de>
16496 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
16497 (lto-streamer-out.o): Likewise.
16498 * cfgloop.c (init_loops_structure): Export, add struct function
16499 argument and adjust.
16500 (flow_loops_find): Adjust.
16501 * cfgloop.h (enum loop_estimation): Add EST_LAST.
16502 (init_loops_structure): Declare.
16503 * lto-streamer-in.c: Include cfgloop.h.
16504 (input_cfg): Input the loop tree.
16505 * lto-streamer-out.c: Include cfgloop.h.
16506 (output_cfg): Output the loop tree.
16507 (output_struct_function_base): Do not drop PROP_loops.
16509 2013-03-26 Richard Biener <rguenther@suse.de>
16511 * tree-cfg.c (execute_build_cfg): Build the loop tree.
16512 (pass_build_cfg): Provide PROP_loops.
16513 (move_sese_region_to_fn): Remove loops that are outlined into fn
16515 * tree-inline.c: Include cfgloop.h.
16516 (initialize_cfun): Do not drop PROP_loops.
16517 (copy_loops): New function.
16518 (copy_cfg_body): Copy loop structure.
16519 (tree_function_versioning): Initialize destination loop tree.
16520 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
16521 (pass_parallelize_loops): Do IL verification.
16522 * loop-init.c (loop_optimizer_init): Fixup loops if required.
16523 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
16524 the CFG make sure we fixup loops as well.
16525 * tree-ssa-tail-merge.c: Include cfgloop.h.
16526 (replace_block_by): When merging loop latches mark loops for fixup.
16527 * lto-streamer-out.c (output_struct_function_base): Drop
16528 PROP_loops for now.
16529 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
16530 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
16531 * ipa-split.c: Include cfgloop.h.
16532 (split_function): Add the new return block to the loop tree root.
16533 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
16534 whether we have removed the forwarder block.
16535 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
16536 * cfgloop.h (place_new_loop): Declare.
16537 * cfgloopmanip.c (place_new_loop): Export.
16538 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
16539 (tree-switch-conversion.o): Likewise.
16540 (tree-complex.o): Likewise.
16541 (tree-inline.o): Likewise.
16542 (tree-ssa-tailmerge.o): Likewise.
16543 (ipa-split.o): Likewise.
16544 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
16545 (tree-ssa-copy.o): Likewise.
16546 * tree-switch-conversion.c: Include cfgloop.h
16547 (process_switch): If we emit a bit-test cascade, schedule loops
16549 * tree-complex.c: Include cfgloop.h.
16550 (expand_complex_div_wide): Properly add new basic-blocks to loops.
16551 * asan.c: Include cfgloop.h.
16552 (create_cond_insert_point): Properly add new basic-blocks to
16553 loops, schedule loop fixup.
16554 * cfgloop.c (verify_loop_structure): Check that looks are not
16556 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
16558 (expand_omp_for_generic): Likewise.
16559 (expand_omp_sections): Likewise.
16560 (expand_omp_atomic_pipeline): Schedule loops for fixup.
16561 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
16562 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
16563 is initialized, not when loops are present.
16564 * tree-parloops.c (parallelize_loops): Remove checking here.
16565 * passes.c (init_optimization_passes): Schedule a copy-propagation
16566 pass before complete unrolling of inner loops.
16568 2013-04-26 Jakub Jelinek <jakub@redhat.com>
16570 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
16571 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
16572 (colorize_init): Add argument to _WIN32 version.
16573 * toplev.c: Include diagnostic-color.h.
16574 (process_options): Default to -fdiagnostics-color=auto if
16575 GCC_COLORS env var is in the environment.
16576 * common.opt (fdiagnostics-color=): Add Var and Init.
16577 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
16578 env var is in the environment, the default is auto rather than never.
16580 * diagnostic.h (file_name_as_prefix): Add context argument.
16581 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
16582 the string as locus.
16583 * langhooks.c (lhd_print_error_function): Adjust caller.
16585 2013-04-25 Lawrence Crowl <crowl@google.com>
16587 * var-tracking.c (shared_hash_def::htab):
16588 Change type to hash_table. Update dependent calls and types.
16590 2013-04-25 Lawrence Crowl <crowl@google.com>
16592 * Makefile.in: Update as needed below.
16594 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
16595 Move declaration to after the type's method definitons.
16597 * attribs.c (htab_t scoped_attributes::attribute_hash):
16598 Change type to hash_table. Update dependent calls and types.
16600 * bitmap.c (htab_t bitmap_desc_hash):
16601 Change type to hash_table. Update dependent calls and types.
16603 * cselib.c (htab_t cselib_hash_table):
16604 Change type to hash_table. Update dependent calls and types.
16606 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
16607 (hash_string_slot_node): Move implementation into lto-streamer.h
16608 struct string_slot_hasher.
16609 (eq_string_slot_node): Likewise.
16611 * data-streamer-out.c: Update output_block::string_hash_table
16612 dependent calls and types.
16614 * dwarf2cfi.c (htab_t trace_index):
16615 Change type to hash_table. Update dependent calls and types.
16617 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
16618 Change type to hash_table. Update dependent calls and types.
16619 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
16620 (htab_t optimize_external_refs::map): Likewise.
16621 (htab_t output_comp_unit::extern_map): Likewise.
16622 (htab_t output_comdat_type_unit::extern_map): Likewise.
16623 (htab_t output_macinfo::macinfo_htab): Likewise.
16624 (htab_t optimize_location_lists::htab): Likewise.
16625 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
16627 * except.c (htab_t ehspec_hash_type):
16628 Change type to hash_table. Update dependent calls and types.
16629 (assign_filter_values::ttypes): Likewise.
16630 (assign_filter_values::ehspec): Likewise.
16631 (sjlj_assign_call_site_values::ar_hash): Likewise.
16632 (convert_to_eh_region_ranges::ar_hash): Likewise.
16634 * gcse.c (htab_t pre_ldst_table):
16635 Change type to hash_table. Update dependent calls and types.
16637 * ggc-common.c (htab_t saving_htab):
16638 Change type to hash_table. Update dependent calls and types.
16639 (htab_t loc_hash): Likewise.
16640 (htab_t ptr_hash): Likewise.
16641 (call_count): Rename ggc_call_count.
16642 (call_alloc): Rename ggc_call_alloc.
16643 (loc_descriptor): Rename make_loc_descriptor.
16644 (add_statistics): Rename ggc_add_statistics.
16646 * ggc-common.c (saving_htab):
16647 Change type to hash_table. Update dependent calls and types.
16649 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
16650 (push_gimplify_context): Likewise.
16651 (pop_gimplify_context): Likewise.
16652 (struct gimple_temp_hash_elt): Added.
16653 (struct gimplify_hasher): Likewise.
16654 (struct gimplify_ctx.temp_htab):
16655 Change type to hash_table. Update dependent calls and types.
16657 * gimple-fold.c: Include gimplify-ctx.h.
16659 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
16660 Change type to hash_table. Update dependent calls and types.
16661 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
16662 avoid potential global name collision.
16664 * gimplify.c: Include gimplify-ctx.h.
16665 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
16666 (htab_t gimplify_ctx::temp_htab):
16667 Update dependent calls and types for new type hash_table.
16668 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
16669 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
16671 * gimplify-ctx.h: New.
16672 (struct gimple_temp_hash_elt): Move from gimplify.c.
16673 (class gimplify_hasher): New.
16674 (struct gimplify_ctx): Move from gimple.h.
16675 (htab_t gimplify_ctx::temp_htab):
16676 Change type to hash_table. Update dependent calls and types.
16678 * graphite-clast-to-gimple.c: Include graphite-htab.h.
16679 (htab_t ivs_params::newivs_index):
16680 Change type to hash_table. Update dependent calls and types.
16681 (htab_t ivs_params::params_index): Likewise.
16682 (htab_t print_generated_program::params_index): Likewise.
16683 (htab_t gloog::newivs_index): Likewise.
16684 (htab_t gloog::params_index): Likewise.
16686 * graphite.c: Include graphite-htab.h.
16687 4htab_t graphite_transform_loops::bb_pbb_mapping):
16688 Change type to hash_table. Update dependent calls and types.
16690 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
16691 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
16692 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
16694 * graphite-dependences.c: Include graphite-htab.h.
16695 (loop_is_parallel_p): Change hash table type of parameter.
16697 * graphite-htab.h: New.
16698 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
16699 (extern find_pbb_via_hash): Move from graphite-poly.h.
16700 (extern loop_is_parallel_p): Move from graphite-poly.h.
16701 (extern get_loop_body_pbbs): Move from graphite-poly.h.
16703 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
16704 (extern loop_is_parallel_p): Move to graphite-htab.h.
16705 (extern get_loop_body_pbbs): Move to graphite-htab.h.
16707 * haifa-sched.c (htab_t delay_htab):
16708 Change type to hash_table. Update dependent calls and types.
16709 (htab_t delay_htab_i2): Likewise.
16711 * ira-color.c (htab_t allocno_hard_regs_htab):
16712 Change type to hash_table. Update dependent calls and types.
16714 * ira-costs.c (htab_t cost_classes_htab):
16715 Change type to hash_table. Update dependent calls and types.
16717 * loop-invariant.c (htab_t merge_identical_invariants::eq):
16718 Change type to hash_table. Update dependent calls and types.
16720 * loop-iv.c (htab_t bivs):
16721 Change type to hash_table. Update dependent calls and types.
16723 * loop-unroll.c (htab_t opt_info::insns_to_split):
16724 Change type to hash_table. Update dependent calls and types.
16725 (htab_t opt_info::insns_with_var_to_expand): Likewise.
16727 * lto-streamer.h (struct string_slot): Move from data-streamer.h
16728 (struct string_slot_hasher): New.
16729 (htab_t output_block::string_hash_table):
16730 Change type to hash_table. Update dependent calls and types.
16732 * lto-streamer-in.c (freeing_string_slot_hasher): New.
16733 (htab_t file_name_hash_table):
16734 Change type to hash_table. Update dependent calls and types.
16736 * lto-streamer-out.c: Update output_block::string_hash_table dependent
16739 * lto-streamer.c (htab_t tree_htab):
16740 Change type to hash_table. Update dependent calls and types.
16742 * omp-low.c: Include gimplify-ctx.h.
16744 * passes.c (htab_t name_to_pass_map):
16745 Change type to hash_table. Update dependent calls and types.
16746 (pass_traverse): Rename to passes_pass_traverse.
16748 * plugin.c (htab_t event_tab):
16749 Change type to hash_table. Update dependent calls and types.
16751 * postreload-gcse.c (htab_t expr_table):
16752 Change type to hash_table. Update dependent calls and types.
16753 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
16755 * sese.c (debug_rename_map_1): Make extern.
16756 (htab_t copy_bb_and_scalar_dependences::rename_map):
16757 Change type to hash_table. Update dependent calls and types.
16759 * sese.h (extern debug_rename_map): Move to .c file.
16761 * store-motion.c (htab_t store_motion_mems_table):
16762 Change type to hash_table. Update dependent calls and types.
16764 * trans-mem.c (htab_t tm_new_mem_hash):
16765 Change type to hash_table. Update dependent calls and types.
16767 * tree-browser.c (htab_t TB_up_ht):
16768 Change type to hash_table. Update dependent calls and types.
16770 * tree-cfg.c (htab_t discriminator_per_locus):
16771 Change type to hash_table. Update dependent calls and types.
16773 * tree-complex.c: Include tree-hasher.h
16774 (htab_t complex_variable_components):
16775 Change type to hash_table. Update dependent calls and types.
16777 * tree-eh.c (htab_t finally_tree):
16778 Change type to hash_table. Update dependent calls and types.
16780 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
16781 struct int_tree_hasher.
16782 (extern int_tree_map_eq): Likewise.
16783 (uid_decl_map_hash): Removed.
16784 (extern decl_tree_map_eq): Likewise.
16786 * tree-hasher.h: New.
16787 (struct int_tree_hasher): New.
16788 (typedef int_tree_htab_type): New.
16790 * tree-inline.c: Include gimplify-ctx.h.
16792 * tree-mudflap.c: Include gimplify-ctx.h.
16794 * tree-parloops.c: Include tree-hasher.h.
16795 (htab_t eliminate_local_variables_stmt::decl_address):
16796 Change type to hash_table. Update dependent calls and types.
16797 (htab_t separate_decls_in_region::decl_copies): Likewise.
16799 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
16800 Change type to hash_table. Update dependent calls and types.
16802 * tree-sra.c (candidates):
16803 Change type to hash_table. Update dependent calls and types.
16805 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
16807 (int_tree_map_hash): Likewise.
16809 * tree-ssa-dom.c (htab_t avail_exprs):
16810 Change type to hash_table. Update dependent calls and types.
16812 * tree-ssa-live.c (var_map_base_init::tree_to_index):
16813 Change type to hash_table. Update dependent calls and types.
16815 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
16816 Change type to hash_table. Update dependent calls and types.
16818 * tree-ssa-phiopt.c (seen_ssa_names):
16819 Change type to hash_table. Update dependent calls and types.
16821 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
16822 Change type to hash_table. Update dependent calls and types.
16824 * tree-ssa-uncprop.c (equiv):
16825 Change type to hash_table. Update dependent calls and types.
16827 2013-04-25 Jakub Jelinek <jakub@redhat.com>
16829 PR rtl-optimization/57003
16830 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
16831 call note_stores with kill_clobbered_value callback again after
16832 killing regs_invalidated_by_call.
16834 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
16836 * config/aarch64/aarch64-simd.md
16837 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
16838 (aarch64_simd_bsl<mode>): Likewise.
16839 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
16841 2013-04-25 Marek Polacek <polacek@redhat.com>
16843 PR tree-optimization/57066
16844 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
16846 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
16848 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
16850 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
16852 * config/aarch64/aarch64-builtins.c
16853 (aarch64_fold_builtin): New.
16854 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
16855 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
16856 * config/aarch64/aarch64-simd-builtins.def (abs): New.
16857 * config/aarch64/arm_neon.h
16858 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
16860 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
16861 Tejas Belagod <tejas.belagod@arm.com>
16863 * config/aarch64/aarch64-builtins.c
16864 (aarch64_gimple_fold_builtin): New.
16865 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
16866 * config/aarch64/aarch64-simd-builtins.def (addv): New.
16867 * config/aarch64/aarch64-simd.md (addpv4sf): New.
16868 (addvv4sf): Update.
16869 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
16871 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16873 * config/aarch64/aarch64.md
16874 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
16876 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16878 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
16879 (*ngcsi_uxtw): New pattern.
16881 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16882 Julian Brown <julian@codesourcery.com>
16884 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
16885 (TB_DREG): Add T_V4HF.
16886 (v4hf_UP): New macro.
16887 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
16888 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
16889 Handle initialisation of V4HF. Adjust initialisation of reinterpret
16891 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
16892 (arm_vector_mode_supported_p): Handle V4HF.
16893 (arm_mangle_map): Handle V4HFmode.
16894 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
16895 * config/arm/arm_neon_builtins.def: Add entries for
16896 vcvtv4hfv4sf, vcvtv4sfv4hf.
16897 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
16898 (neon_vcvtv4hfv4sf): Likewise.
16899 * config/arm/neon-gen.ml: Handle half-precision floating point
16901 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
16902 * config/arm/arm_neon.h: Regenerate.
16903 * config/arm/neon.ml (type elts): Add F16.
16904 (type vectype): Add T_float16x4, T_floatHF.
16905 (type vecmode): Add V4HF.
16906 (type features): Add Requires_FP_bit feature.
16907 (elt_width): Handle F16.
16908 (elt_class): Likewise.
16909 (elt_of_class_width): Likewise.
16910 (mode_of_elt): Refactor.
16911 (type_for_elt): Handle F16, fix error messages.
16912 (vectype_size): Handle T_float16x4.
16913 (vcvt_sh): New function.
16914 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
16915 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
16916 (string_of_mode): Handle V4HF.
16917 * doc/arm-neon-intrinsics.texi: Regenerate.
16919 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
16921 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
16922 format specifier in 'X' case.
16924 2013-04-25 Alan Modra <amodra@gmail.com>
16927 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
16928 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
16929 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
16930 Repeat for many other rotate/shift and mask patterns using subregs.
16931 Name lshiftrt insns.
16932 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
16933 on WORDS_BIG_ENDIAN.
16935 2013-04-25 Alan Modra <amodra@gmail.com>
16937 * config.gcc: Support little-endian powerpc-linux targets.
16938 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
16939 (LINK_OS_LINUX_SPEC): Define.
16940 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
16941 Preserve MASK_LITTLE_ENDIAN.
16942 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
16943 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
16944 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
16945 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
16946 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
16947 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
16948 Correct fp word order for little-endian. Don't shift toc entries
16949 smaller than a word for little-endian.
16950 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
16951 (bswapdi2 splits): Correct low-part subreg for little-endian.
16952 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
16953 low/high where such is correct only for be.
16954 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
16955 little-endian for -mcall-aixdesc.
16957 2013-04-25 Alan Modra <amodra@gmail.com>
16959 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
16960 replace_equiv_address_nv.
16962 2013-04-25 Alan Modra <amodra@gmail.com>
16964 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
16966 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
16969 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
16970 * rtl.h (struct rtx_def): ...
16972 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
16974 PR rtl-optimizations/57046
16975 * lra-constraints (split_reg): Set up lra_risky_transformations_p
16976 for multi-reg splits.
16978 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
16980 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
16982 2013-04-24 Sterling Augustine <saugustine@google.com>
16984 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
16985 (comp_dir_string, debug_str_dwo_section): New.
16986 (DEBUG_STR_DWO_SECTION): Rename to ...
16987 (DEBUG_DWO_STR_SECTION): ... this.
16988 (DEBUG_NORM_STR_SECTION): Delete.
16989 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
16990 (DEBUG_STR_DWO_SECTION_FLAGS): New.
16991 (find_AT_string): Move most logic to ...
16992 (find_AT_string_in_table): ... here. New.
16993 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
16994 add_skeleton_AT_string. Delete logic.
16995 (output_skeleton_debug_sections): Remove call to
16996 add_top_level_skeleton_die_attrs.
16997 (add_comp_dir_attribute): Move logic to comp_dir_string.
16998 (dwarf2out_init): Initialize debug_str_dwo_section.
16999 (output_indirect_string): Call find_string_form.
17000 (output_indirect_strings): Rewrite.
17001 (prune_unused_types): Empty skeleton_debug_str_hash.
17002 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
17003 (dwarf2out_finish): Call output_indirect_strings.
17005 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
17007 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
17009 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
17011 * rtl.h (struct rtx_def): Add comment for field jump.
17012 (LRA_SUBREG_P): New macro.
17013 * recog.c (register_operand): Check LRA_SUBREG_P.
17014 * lra.c (lra): Add note at the end of RTL code. Align non-empty
17016 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
17017 (lra_final_code_change): Skip subreg change for operators.
17018 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
17019 if there are no operand changes.
17020 * lra-constraints.c (curr_insn_set): New.
17021 (match_reload): Set LRA_SUBREG_P.
17022 (emit_spill_move): Ditto.
17023 (check_and_process_move): Use curr_insn_set. Process only single
17024 set insns. Don't initialize sec_mem_p and change_p.
17025 (simplify_operand_subreg): Use LRA_SUBREG_P.
17026 (reg_in_class_p): New function.
17027 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
17028 of #ifdef. Add code to remove cycling.
17029 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
17030 non-null disp. Reload inner instead of disp when base and index
17031 are null. Try to put lo_sum into register.
17032 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
17033 (check_and_process_move): Move code for move cost check to
17034 simple_move_p. Remove equiv_substitution.
17035 (simple_move_p): New function.
17036 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
17037 curr_insn_set. Call check_and_process_move only for single set
17038 insns. Use the new function. Move call of check_and_process_move
17039 after operand equiv substitution and address process.
17041 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
17043 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
17044 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
17045 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
17047 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
17049 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
17051 2013-04-24 Marek Polacek <polacek@redhat.com>
17053 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
17054 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
17055 (select_loops_exit_conditions): Likewise.
17056 (number_of_iterations_for_all_loops): Likewise.
17057 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
17058 (scev_analysis): Likewise.
17060 2013-04-02 Catherine Moore <clm@codesourcery.com>
17061 Chao-ying Fu <fu@mips.com>
17063 * config/mips/micromips.md (jraddiusp): New pattern.
17064 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
17065 instruction if possible.
17067 2013-04-24 Alan Modra <amodra@gmail.com>
17069 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
17071 2013-04-24 Julian Brown <julian@codesourcery.com>
17072 Chung-Lin Tang <cltang@codesourcery.com>
17074 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
17075 dependency behavior in enumeration type DIE generation. Add TODO note
17076 to comments about future DW_FORM_sdata/udata re-work of related code.
17078 2013-04-23 Lawrence Crowl <crowl@google.com>
17080 * Makefile.in: Update as needed below.
17082 * hash-table.h (class hash_table):
17083 Correct many methods with parameter types compare_type to the correct
17084 value_type. (Correct code was unlikely to notice the change.)
17085 (hash_table::elements_with_deleted) New.
17086 (class hashtable::iterator): New.
17087 (hashtable::begin()): New.
17088 (hashtable::end()): New.
17089 (FOR_EACH_HASH_TABLE_ELEMENT): New.
17091 * statistics.c (statistics_hashes):
17092 Change type to hash_table. Update dependent calls and types.
17094 * tree-into-ssa.c (var_infos):
17095 Change type to hash_table. Update dependent calls and types.
17097 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
17098 Change type to hash_table. Update dependent calls and types.
17100 * tree-ssa-loop-im.c (struct mem_ref.refs):
17101 Change type to hash_table. Update dependent calls and types.
17103 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
17104 Change type to hash_table. Update dependent calls and types.
17106 * tree-ssa-sccvn.c (vn_tables_s::nary):
17107 Change type to hash_table. Update dependent calls and types.
17108 (vn_tables_s::phis): Likewise.
17109 (vn_tables_s::references): Likewise.
17111 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
17112 (vn_reference_eq): Update parameter and return types.
17114 * tree-ssa-structalias.c (pointer_equiv_class_table):
17115 Change type to hash_table. Update dependent calls and types.
17116 (location_equiv_class_table): Likewise.
17118 * tree-vect-data-refs.c: Consequential changes for making
17119 peeling a hash_table.
17121 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
17122 (destroy_loop_vec_info): Dependent hash_table update.
17124 * tree-vectorizer.h (peeling_htab):
17125 Change type to hash_table. Update dependent calls and types.
17127 2013-04-23 Shiva Chen <shiva0217@gmail.com>
17129 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
17130 to check the register content is equal or not.
17131 * lra-constraints.c (match_reload): Use lra_assign_reg_val
17132 to assign register content record.
17133 * lra-eliminations.c (update_reg_eliminate): Use
17134 lra_update_reg_val_offset to update register content offset.
17135 * lra-int.h (struct lra_reg): Add offset member.
17136 (lra_reg_val_equal_p): New static inline function.
17137 (lra_update_reg_val_offset): New static inline function.
17138 (lra_assign_reg_val): New static inline function.
17139 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
17140 to assign register content record.
17141 (initialize_lra_reg_info_element): Initial offset to zero.
17143 2013-04-23 Catherine Moore <clm@codesourcery.com>
17145 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
17146 operands. Record compression.
17148 2013-04-23 Xinliang David Li <davidxl@google.com>
17150 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
17152 2013-04-23 Richard Biener <rguenther@suse.de>
17154 PR middle-end/57036
17155 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
17156 parameter, only add abnormal goto edges from the copied body
17157 if the call could perform abnormal gotos.
17158 (copy_cfg_body): Adjust.
17160 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
17162 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
17164 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
17166 * coretypes.h (gimple_stmt_iterator): Add struct to make
17169 2013-04-23 Richard Biener <rguenther@suse.de>
17171 PR tree-optimization/57026
17172 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
17173 from SSA names occuring in abnormal PHI nodes.
17175 2013-04-22 Andi Kleen <ak@linux.intel.com>
17177 * lto/lto.c (print_lto_report_1): Fix LTO report names.
17179 2013-04-22 Andi Kleen <ak@linux.intel.com>
17181 * lto/lto.c (print_lto_report_1): Declare early.
17182 (read_cgraph_and_symbols): Call print_lto_report_1 early.
17184 2013-04-22 Andi Kleen <ak@linux.intel.com>
17186 * common.opt (-flto-report-wpa): Add.
17187 * doc/invoke.texi (-flto-report-wpa): Add.
17188 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
17191 2013-04-22 Xinliang David Li <davidxl@google.com>
17193 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
17194 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
17195 * Makefile.in: New dependency
17197 David Daney <ddaney.cavm@gmail.com>
17199 * configure.ac (gcc_cv_as_micromips_support): Use the
17200 --fatal-warnings option.
17201 * configure: Regenerate.
17203 2013-04-22 Marek Polacek <polacek@redhat.com>
17206 * tsan.c (instrument_expr): Don't instrument expression
17207 in case its size is zero.
17209 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
17213 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
17215 * config/alpha/alpha.c (TARGET_LRA_P): New define.
17217 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
17219 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
17220 (gimple_stmt_iterator): New typedef.
17221 * gimple.h (gimple_stmt_iterator): Rename to...
17222 (gimple_stmt_iterator_d): ... This.
17223 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
17224 trees be valid for GIMPLE and GENERIC.
17225 (TARGET_GIMPLE_FOLD_BUILTIN): New.
17226 * gimple-fold.c (gimple_fold_call): Call target hook
17227 gimple_fold_builtin.
17228 * hooks.c (hook_bool_gsiptr_false): New.
17229 * hooks.h (hook_bool_gsiptr_false): New.
17230 * target.def (fold_stmt): New.
17231 * doc/tm.texi: Regenerate.
17233 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
17236 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
17237 a set sp if no stack realignment.
17239 2013-04-22 Nick Clifton <nickc@redhat.com>
17241 * config.gcc (tilegx-linux): Extend extra_objs rather than
17243 (tilepro-linux): Likewise.
17245 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
17247 * config/aarch64/aarch64-builtins.c
17249 (CF0, CF1, CF2, CF3, CF4, CF10): New.
17250 (VAR<1-12>): Add MAP parameter.
17251 (BUILTIN_*): Likewise.
17252 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
17253 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
17254 (aarch64_ushl_n<mode>): Likewise.
17255 (aarch64_sshr_n<mode>): Likewise.
17256 (aarch64_ushr_n<mode>): Likewise.
17257 (aarch64_<maxmin><mode>): Likewise.
17258 (aarch64_sqrt<mode>): Likewise.
17259 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
17260 (vshr<q>_n_*): Likewise.
17262 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
17264 * config/aarch64/aarch64-builtins.c
17265 (aarch64_simd_builtin_type_mode): Handle SF types.
17267 (BUILTIN_GPF): Define.
17268 (aarch64_init_simd_builtins): Handle SF types.
17269 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
17270 (frecps): Likewise.
17271 (frecpx): Likewise.
17272 * config/aarch64/aarch64-simd.md
17273 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
17274 (aarch64_frecpe<mode>): New.
17275 (aarch64_frecps<mode>): Likewise.
17276 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
17277 (v8type): Add frecp<esx>.
17278 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
17279 (aarch64_frecps<mode>): Likewise.
17280 * config/aarch64/iterators.md (FRECP): New.
17281 (frecp_suffix): Likewise.
17282 * config/aarch64/arm_neon.h
17283 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
17285 2013-04-22 Christian Bruel <christian.bruel@st.com>
17288 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
17289 (REG_CLASS_NAMES): Idem.
17290 (REG_CLASS_CONTENTS): Idem.
17291 (REGCLASS_HAS_FP_REG): Idem.
17292 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
17293 (sh_conditional_register_usage): Idem.
17295 2013-04-21 Jeff Law <law@redhat.com>
17297 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
17298 (ssa_forward_propagate_and_combine): Use it.
17300 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
17302 * lra.c: Update the flow chart diagram.
17304 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
17306 PR rtl-optimization/56847
17307 * lra-constraints.c (process_alt_operands): Discourage alternative
17308 with non-matche doffsettable memory constraint fro memory with
17311 2013-04-19 Richard Biener <rguenther@suse.de>
17313 PR tree-optimization/56982
17314 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
17316 * gimplify.c (gimplify_call_expr): Notice special calls.
17317 (gimplify_modify_expr): Likewise.
17318 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
17319 abnormal control flow receivers.
17320 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
17321 in the same way as cfun->has_nonlocal_labels.
17322 (gimple_purge_dead_abnormal_call_edges): Likewise.
17323 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
17324 receivers start a basic-block.
17326 2013-04-19 Richard Biener <rguenther@suse.de>
17328 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
17330 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
17331 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
17332 (SLP_TREE_LOAD_PERMUTATION): Add.
17333 (vect_transform_slp_perm_load): Adjust prototype.
17334 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
17335 (vect_free_slp_instance): Likewise.
17336 (vect_create_new_slp_node): Likewise.
17337 (vect_supported_slp_permutation_p): Remove.
17338 (vect_slp_rearrange_stmts): Adjust.
17339 (vect_supported_load_permutation_p): Likewise. Inline
17340 vect_supported_slp_permutation_p here.
17341 (vect_analyze_slp_instance): Compute load permutations per
17342 slp node instead of per instance.
17343 (vect_get_slp_defs): Adjust.
17344 (vect_transform_slp_perm_load): Likewise.
17345 (vect_schedule_slp_instance): Remove redundant code.
17346 (vect_schedule_slp): Remove hack for PR56270, add it ...
17347 * tree-vect-stmts.c (vectorizable_load): ... here, do not
17348 CSE loads for SLP. Adjust.
17350 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
17352 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
17353 spelling in two comments.
17355 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
17358 * config/arm/arm.c (load_multiple_sequence): Require SP
17359 as base register for loads if SP is in the register list.
17361 2013-04-19 Martin Jambor <mjambor@suse.cz>
17363 PR tree-optimization/56718
17364 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
17365 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
17366 and made public. Adjusted all callers.
17367 (ipa_intraprocedural_devirtualization): New function.
17368 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
17369 (ipa_intraprocedural_devirtualization): Likewise.
17370 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
17372 2013-04-19 Richard Biener <rguenther@suse.de>
17374 PR tree-optimization/57000
17375 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
17377 2013-04-19 Terry Guo <terry.guo@arm.com>
17379 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
17381 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
17382 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
17383 (cortex_m4_fmacs): Use new reservations.
17384 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
17386 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
17388 PR rtl-optimization/56999
17389 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
17391 (lra_coalesce): Remove split_origin_bitmap and related code.
17392 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
17393 ranges if necessary.
17395 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
17397 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
17399 (ix86_expand_call): Remove clobbered_registers array and use
17400 x86_64_ms_sysv_extra_clobbered_registers instead.
17401 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
17403 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
17405 * config/i386/i386.md (*call_rex64_ms_sysv): Use
17406 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
17407 (*call_value_rex64_ms_sysv): Ditto.
17409 2013-04-18 Cary Coutant <ccoutant@google.com>
17411 * dwarf2out.c (output_pubnames): Check die_perennial_p of
17412 parent instead of die_mark.
17414 2013-04-18 Diego Novillo <dnovillo@google.com>
17416 * gimple.c (create_gimple_tmp): New.
17417 (get_expr_type): New.
17418 (build_assign): New.
17419 (build_type_cast): New.
17420 * gimple.h (enum ssa_mode): Define.
17421 (gimple_seq_set_location): New.
17422 * asan.c (build_check_stmt): Change some gimple_build_* calls
17423 to use build_assign and build_type_cast.
17425 2013-04-18 Richard Biener <rguenther@suse.de>
17427 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
17428 handle negative step. Remove redundant checks.
17429 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
17430 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
17431 for negative step and grouped loads fail to vectorize.
17433 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
17435 * emit-rtl.c (reset_insn_used_flags): New function.
17436 (reset_all_used_flags): Use it.
17437 (verify_insn_sharing): New function.
17438 (verify_rtl_sharing): Fix verification for SEQUENCEs.
17440 2013-04-18 Jakub Jelinek <jakub@redhat.com>
17442 PR tree-optimization/56984
17443 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
17444 and (x >> M) >= N don't register any assertion if N << M is the
17447 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
17449 * lower-subreg.c (resolve_simple_move): If called self-recursive,
17450 do not delete_insn insns that have not yet been emitted, only
17451 unlink them with remove_insn.
17452 * df-scan.c (df_insn_delete): Revert r197492.
17454 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
17456 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
17457 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
17459 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
17461 * config/arm/arm.md (movsicc_insn): Convert define_insn into
17462 define_insn_and_split.
17463 (and_scc,ior_scc,negscc): Likewise.
17464 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
17466 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
17468 * config/arm/arm.c (use_return_insn): Return 0 for targets that
17469 can benefit from using a sequence of LDRD instructions in epilogue
17470 instead of a single LDM instruction.
17472 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
17475 * doc/extend.texi: Fix typo.
17477 2013-04-17 Richard Biener <rguenther@suse.de>
17479 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
17480 (vect_build_slp_tree): ... here.
17481 (vect_build_slp_tree_1): Compute which stmts of the SLP group
17482 match. Remove special-casing of mismatched complex loads.
17483 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
17484 re-try the match with swapped commutative operands.
17485 (vect_supported_load_permutation_p): Remove special-casing of
17486 mismatched complex loads.
17487 (vect_analyze_slp_instance): Adjust.
17489 2013-04-17 Richard Biener <rguenther@suse.de>
17491 PR rtl-optimization/56921
17492 * cfgloop.h (struct loop): Add simple_loop_desc member.
17493 (struct niter_desc): Mark with GTY(()).
17494 (simple_loop_desc): Do not use aux field but simple_loop_desc.
17495 * loop-iv.c (get_simple_loop_desc): Likewise.
17496 (free_simple_loop_desc): Likewise.
17499 2013-04-16 Richard Biener <rguenther@suse.de>
17501 PR rtl-optimization/56921
17502 * loop-init.c (pass_rtl_move_loop_invariants): Add
17503 TODO_do_not_ggc_collect to todo_flags_finish.
17504 (pass_rtl_unswitch): Same.
17505 (pass_rtl_unroll_and_peel_loops): Same.
17506 (pass_rtl_doloop): Same.
17508 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
17510 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
17511 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
17512 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
17514 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
17515 * tree-streamer.c (record_common_node): Adjust reference in comment.
17517 2013-04-17 Terry Guo <terry.guo@arm.com>
17519 * config/arm/cortex-m4.md: Add a new bypass.
17521 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17523 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
17525 (*subs_<optab><mode>_multp2): New pattern.
17526 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
17527 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
17529 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17531 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
17532 (*subs_mul_imm_<mode>): New pattern.
17534 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
17537 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
17538 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
17539 (vsx_movti_32bit): Change j->wa to O->wa.
17541 2013-04-16 Richard Biener <rguenther@suse.de>
17543 PR rtl-optimization/56921
17544 * loop-init.c (pass_rtl_move_loop_invariants): Add
17545 TODO_do_not_ggc_collect to todo_flags_finish.
17546 (pass_rtl_unswitch): Same.
17547 (pass_rtl_unroll_and_peel_loops): Same.
17548 (pass_rtl_doloop): Same.
17550 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
17552 * config/arm/arm.c (emit_multi_reg_push): New declaration
17553 for an existing function.
17554 (arm_emit_strd_push): New function.
17555 (arm_expand_prologue): Used here.
17556 (arm_emit_ldrd_pop): New function.
17557 (arm_expand_epilogue): Used here.
17558 (arm_get_frame_offsets): Update condition.
17559 (arm_emit_multi_reg_pop): Add a special case for load of a single
17560 register with writeback.
17562 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
17564 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
17567 2013-04-16 Richard Biener <rguenther@suse.de>
17569 PR tree-optimization/56756
17570 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
17571 (first_mem_ref_loc): New.
17572 (execute_sm): Place the load temporarily before a previous
17573 access instead of in the latch edge to ensure its SSA dependencies
17574 are defined at points dominating the load.
17576 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
17578 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
17579 correct fix by moving header and footer insn to the footer of
17580 the merged basic block. Clear BB_END of the merged-away block.
17582 PR middle-end/43631
17583 * emit-rtl.c (make_note_raw): New function.
17584 (link_insn_into_chain): New static inline function.
17585 (add_insn): Use it.
17586 (add_insn_before, add_insn_after): Factor insn chain linking code...
17587 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
17588 using link_insn_into_chain.
17589 (note_outside_basic_block_p): New helper function for emit_note_after
17590 and emit_note_before.
17591 (emit_note_after): Use nobb variant of add_insn_after if the note
17592 should not be contained in a basic block.
17593 (emit_note_before): Use nobb variant of add_insn_before if the note
17594 should not be contained in a basic block.
17595 (emit_note_copy): Use make_note_raw.
17596 (emit_note): Likewise.
17597 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
17598 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
17599 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
17600 the moved barrier the tail of the basic block it follows.
17601 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
17603 2013-04-15 Jakub Jelinek <jakub@redhat.com>
17605 PR tree-optimization/56962
17606 * gimple-ssa-strength-reduction.c (record_increment): Only set
17607 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
17608 either rhs1 or rhs2 is equal to c->base_expr.
17610 2013-04-15 Richard Biener <rguenther@suse.de>
17612 PR tree-optimization/56933
17613 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
17615 (GROUP_READ_WRITE_DEPENDENCE): Remove.
17616 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
17617 * tree-vect-data-refs.c (vect_analyze_group_access): Move
17618 dependence check ...
17619 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
17621 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
17622 GROUP_READ_WRITE_DEPENDENCE.
17624 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17626 * emit-rtl.c (reset_all_used_flags): New function.
17627 (verify_rtl_sharing): Call reset_all_used_flags before and after
17628 performing the checks.
17630 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17632 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
17633 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
17634 * config/arm/constraints.md (De): New constraint.
17635 * config/arm/neon.md (anddi3_neon): Delete.
17636 (neon_vand<mode>): Expand to standard anddi3 pattern.
17637 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
17638 Move earlier in the file.
17639 (neon_inv_logic_op2): Likewise.
17640 (arm_anddi_operand_neon): New predicate.
17642 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17644 * configure.ac (gcc_cv_ld_as_needed): Set
17645 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
17646 Use -z ignore, -z record on *-*-solaris2*.
17647 (HAVE_LD_AS_NEEDED): Update comment.
17648 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
17649 * configure: Regenerate.
17650 * config.in: Regenerate.
17651 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
17652 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
17653 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
17654 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
17655 equivalents. Fix markup.
17656 * doc/tm.texi: Regenerate.
17658 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
17660 * config/i386/i386.opt: New option mstack-protector-guard=.
17661 * config/i386/i386-opts.h: Add enum stack_protector_guard.
17662 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
17663 TARGET_SSP_TLS_GUARD.
17664 * config/i386/i386.c (ix86_option_override_internal): Set
17665 ix86_stack_protector_guard.
17666 * config/i386/i386.md (stack_protect_set): Enable for
17667 TARGET_SSP_TLS_GUARD only.
17668 (stack_protect_set_<mode>): Ditto.
17669 (stack_protect_test): Ditto.
17670 (stack_protect_test_<mode>): Ditto.
17671 * doc/invoke.texi (i386 Option): Document.
17673 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
17676 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
17677 (S_MODES): Set H_MODE bit.
17678 (SF_MODES): Set only S_MODE and SF_MODE bits.
17679 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
17680 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
17681 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
17682 <MODE_FLOAT>: Likewise.
17684 2013-04-15 Joey Ye <joey.ye@arm.com>
17686 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
17688 2013-04-15 Joey Ye <joey.ye@arm.com>
17690 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
17692 (thumb_far_jump_used_p): Count instruction size and set
17695 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
17697 * reorg.c (fill_simple_delay_slots): Reindent block of code.
17698 * resource.c (mark_target_live_regs): Reformat conditional block.
17700 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
17702 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
17703 notes, they are emitted only just before final.
17704 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
17706 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
17708 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
17709 * cfgrtl.c (delete_insn): Call it here instead.
17710 * lra-spills.c (lra_final_code_change): Use delete_insn.
17711 * haifa-sched.c (sched_remove_insn): Likewise.
17712 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
17713 returning to the nop pool.
17714 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
17715 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
17717 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
17719 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
17720 * doc/tm.texi: Regenerated.
17722 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
17724 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
17727 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
17729 * df-core.c (df_find_def): Compare register numbers.
17730 (df_find_use): Likewise.
17732 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
17735 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
17736 lra_in_progress for return.
17738 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
17740 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
17741 define_insn into define_insn_and_split and emit movsicc patterns.
17743 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
17745 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
17747 2013-04-12 Richard Biener <rguenther@suse.de>
17749 * tree-pass.h (TODO_do_not_ggc_collect): New.
17750 * passes.c (execute_one_ipa_transform_pass): Honor
17751 TODO_do_not_ggc_collect.
17752 (execute_one_pass): Likewise.
17755 2013-04-10 Richard Biener <rguenther@suse.de>
17757 * passes.c (init_optimization_passes): Remove reload pass.
17758 * ira.c (do_reload): Merge into ...
17760 (rest_of_handle_reload): Remove.
17761 (pass_reload): Likewise.
17762 * config/i386/i386.c (ix86_option_override): Refer to ira instead
17763 of reload for vzeroupper pass placement.
17765 2013-04-12 Jakub Jelinek <jakub@redhat.com>
17767 PR tree-optimization/56918
17768 PR tree-optimization/56920
17769 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
17770 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
17771 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
17772 use wide_mul_with_sign method.
17774 2013-04-12 Richard Biener <rguenther@suse.de>
17776 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
17777 not be considered a gimple constant.
17779 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
17781 * fold-const.c (const_binop): Handle vector shifts by a scalar.
17782 (fold_binary_loc): Call const_binop also for mixed vector-scalar
17785 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
17786 Jakub Jelinek <jakub@redhat.com>
17788 * opts.c: Include diagnostic-color.h.
17789 (common_handle_option): Handle OPT_fdiagnostics_color_.
17790 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
17791 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
17792 (diagnostic-color.o): New.
17793 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
17794 (diagnostic_color_rule): New enum.
17795 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
17796 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
17797 the location string.
17798 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
17799 either NULL, or color kind.
17800 * diagnostic-color.c: New file.
17801 * diagnostic-color.h: New file.
17802 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
17804 * doc/invoke.texi (-fdiagnostics-color): Document.
17805 * pretty-print.h (pp_show_color): Define.
17806 (struct pretty_print_info): Add show_color field.
17807 * diagnostic.c: Include diagnostic-color.h.
17808 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
17809 macros. Colorize error:, warning: etc. strings and also the location
17811 (diagnostic_show_locus): Colorize the caret line.
17812 * pretty-print.c: Include diagnostic-color.h.
17813 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
17814 inside of %< %> quotes or quoted through q format modifier.
17816 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17818 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
17820 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17822 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
17823 code in CC_NZ mode.
17824 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
17827 2013-04-11 Marek Polacek <polacek@redhat.com>
17829 PR tree-optimization/48184
17830 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
17832 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
17834 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
17835 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
17836 (skip_simple_arithmetic): Tidy up.
17837 * tree.h (skip_simple_constant_arithmetic): Declare.
17839 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17841 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
17843 2013-04-11 Richard Biener <rguenther@suse.de>
17845 * tree-vect-loop.c (get_initial_def_for_induction): Properly
17846 generate vector constants.
17848 2013-04-11 Richard Biener <rguenther@suse.de>
17850 PR tree-optimization/56878
17851 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
17852 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
17854 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
17855 Prefer to align the DR with the most invariant base address.
17857 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
17859 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
17862 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
17864 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
17865 floating-point vector comparisons against 0.
17867 2013-04-11 Jakub Jelinek <jakub@redhat.com>
17869 PR tree-optimization/56899
17870 * fold-const.c (extract_muldiv_1): Apply distributive law
17871 only if TYPE_OVERFLOW_WRAPS (ctype).
17873 2013-04-11 Bin Cheng <bin.cheng@arm.com>
17876 * ira-costs.c (scan_one_insn): Check whether the source rtx of
17877 loading has side effect.
17879 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
17881 * config/sparc/sparc.c: Include tree-pass.h.
17882 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
17883 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
17884 head of file. Change return type. Split off gate function.
17885 (sparc_gate_work_around_errata): New function.
17886 (pass_work_around_errata): New pass definition.
17887 (insert_pass_work_around_errata) New pass insert definition to
17888 insert pass_work_around_errata just after delayed-branch scheduling.
17889 (sparc_option_override): Insert the pass.
17890 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
17892 2013-04-10 David S. Miller <davem@davemloft.net>
17894 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
17895 or -mcpu=hypersparc.
17897 * target.def (cstore_mode): New hook.
17898 * target.h: Include insn-codes.h
17899 * targhooks.c: Likewise.
17900 (default_cstore_mode): New function.
17901 * targhooks.h: Declare it.
17902 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
17903 * doc/tm.texi: Rebuild.
17904 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
17905 target hook, rather than inspecting the insn_data.
17906 * config/sparc/sparc.c (sparc_cstore_mode): New function.
17907 (TARGET_CSTORE_MODE): Redefine.
17908 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
17910 * config/sparc/predicates.md (cstore_result_operand): New special
17912 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
17913 Use it for operand 0.
17914 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
17915 (*snesi_special): Likewise.
17916 (*snesi_zero): Likewise.
17917 (*seqsi_zero): Likewise.
17918 (*sltu_insn): Likewise.
17919 (*sgeu_insn): Likewise.
17920 (*seqdi_special): Make operand 0 and comparison operation be of
17922 (*snedi_special): Likewise.
17923 (*snedi_special_vis3): Likewise.
17924 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
17925 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
17926 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
17927 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
17928 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
17929 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
17930 (*sltu_extend_sp64): Likewise.
17931 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
17932 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
17933 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
17934 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
17935 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
17937 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
17939 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
17940 (aarch64_start_file): Use the new function.
17942 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
17943 Jason Merrill <jason@redhat.com>
17945 * common.opt: Add -gdwarf.
17946 * opts.c (common_handle_option): Handle it.
17947 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
17949 2013-04-10 Richard Biener <rguenther@suse.de>
17951 * passes.c (execute_todo): Do not call ggc_collect conditional here.
17952 (execute_one_ipa_transform_pass): But unconditionally here.
17953 (execute_one_pass): And here.
17954 (init_optimization_passes): Remove reload pass.
17955 * tree-pass.h (TODO_ggc_collect): Remove.
17956 (pass_reload): Likewise.
17957 * ira.c (do_reload): Merge into ...
17959 (rest_of_handle_reload): Remove.
17960 (pass_reload): Likewise.
17961 * config/i386/i386.c (ix86_option_override): Refer to ira instead
17962 of reload for vzeroupper pass placement.
17963 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
17964 and todo_flags_finish of all passes.
17966 2013-04-10 Richard Biener <rguenther@suse.de>
17968 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
17969 first_const_oprnd field, rename first_def_type to first_op_type.
17970 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
17971 (vect_get_and_check_slp_defs): Always use the type of the
17972 operand. Allow mixed vect_external_def, vect_constant_def types.
17973 (vect_get_constant_vectors): Handle mixed vect_external_def,
17974 vect_constant_def types.
17976 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
17978 PR tree-optimization/55524
17979 * tree-ssa-math-opts.c
17980 (convert_mult_to_fma): Don't use an fms construct
17981 when we don't have an fms operation, but fnma, and it looks
17982 likely that we'll be able to use the latter.
17984 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
17986 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
17988 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
17989 inline fail caused by overwritable functions.
17991 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
17993 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
17994 unnecessary bits in the constant power of two case.
17996 2013-04-10 Richard Biener <rguenther@suse.de>
17998 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
17999 broken code swapping operands.
18000 (vect_build_slp_tree): Do not compute load permutations here.
18001 (vect_analyze_slp_instance): Compute load permutations here,
18002 after building the SLP tree.
18004 2013-04-09 Christian Bruel <christian.bruel@st.com>
18006 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
18007 of next/prev_real_insn.
18009 2013-04-09 Jan Hubicka <jh@suse.cz>
18011 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
18012 Drop aliased parameter.
18013 (function_and_variable_visibility): Do not handle alias pairs.
18014 * cgraph.c (varpool_externally_visible_p): Update prototype.
18015 * varpool.c (varpool_add_new_variable): Update.
18017 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18019 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
18021 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
18023 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
18025 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
18027 2013-04-09 Marek Polacek <polacek@redhat.com>
18029 PR tree-optimization/48762
18030 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
18032 2013-04-09 Richard Biener <rguenther@suse.de>
18034 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
18036 (vect_build_slp_tree): Likewise.
18037 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
18038 calculating the cost of a SLP instance.
18039 (vect_analyze_slp_instance): Use it from here, after building
18042 2013-04-09 Jakub Jelinek <jakub@redhat.com>
18044 PR middle-end/56883
18045 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
18046 expand_omp_for_static_chunk): Use simple_p = true in
18047 force_gimple_operand_gsi calls when assigning to addressable decls.
18049 2013-04-09 Jeff Law <law@redhat.com>
18051 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
18052 when the boolean was created by converting a wider object which
18053 had a boolean range.
18055 2013-04-09 Richard Biener <rguenther@suse.de>
18057 * tree-vectorizer.h (slp_void_p): Remove.
18058 (slp_tree): Typedef before _slp_tree declaration.
18059 (struct _slp_tree): Use a vector of slp_tree as children.
18060 (vect_get_place_in_interleaving_chain): Remove.
18061 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
18063 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
18065 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
18066 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
18067 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
18068 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
18069 Use slp_node instead of slp_void_p and adjust.
18071 2013-04-09 Richard Biener <rguenther@suse.de>
18073 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
18074 work that is not necessary.
18076 2013-04-09 Jakub Jelinek <jakub@redhat.com>
18078 PR tree-optimization/56854
18079 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
18080 forward into clobber stmts if it would change MEM_REF lhs into
18083 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
18085 * tree.c (type_hash_lookup, type_hash_add): Make static.
18086 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
18088 2013-04-09 Richard Biener <rguenther@suse.de>
18090 * tree.h (unsave_expr_now): Remove.
18091 * tree-inline.c (mark_local_for_remap_r): Remove.
18092 (unsave_expr_1): Likewise.
18093 (unsave_r): Likewise.
18094 (unsave_expr_now): Likewise.
18095 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
18096 (propagate_tree_value): Likewise.
18098 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
18100 * doc/rtl.texi (sequence): Rewrite documentation to match the
18101 current use of SEQUENCE rtl objects.
18102 * rtl.def (SEQUENCE): Likewise.
18104 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
18105 Update documentation.
18106 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
18107 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
18109 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
18111 2013-04-08 Teresa Johnson <tejohnson@google.com>
18113 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
18114 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
18116 (estimate_edge_size_and_time): Add comment to suggest using rounding
18118 (estimate_node_size_and_time): Ditto.
18119 (remap_edge_change_prob): Use helper rounding divide methods.
18120 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
18121 (gimple_mod_pow2_value_transform): Ditto.
18122 (gimple_mod_subtract_transform): Ditto.
18123 (gimple_ic_transform): Ditto.
18124 (gimple_stringops_transform): Ditto.
18125 * stmt.c (conditional_probability): Ditto.
18126 (emit_case_dispatch_table): Ditto.
18127 * lto-cgraph.c (merge_profile_summaries): Ditto.
18128 * tree-optimize.c (execute_fixup_cfg): Ditto.
18129 * cfgcleanup.c (try_forward_edges): Ditto.
18130 * cfgloopmanip.c (scale_loop_profile): Ditto.
18132 (duplicate_loop_to_header_edge): Ditto.
18133 (lv_adjust_loop_entry_edge): Ditto.
18134 * tree-vect-loop.c (vect_transform_loop): Ditto.
18135 * profile.c (compute_branch_probabilities): Ditto.
18136 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
18137 * lto-streamer-in.c (input_cfg): Ditto.
18138 * gimple-streamer-in.c (input_bb): Ditto.
18139 * ipa-cp.c (update_profiling_info): Ditto.
18140 (update_specialized_profile): Ditto.
18141 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
18142 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
18144 * sched-rgn.c (compute_dom_prob_ps): Ditto.
18145 (compute_trg_info): Ditto.
18146 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
18147 (purge_dead_edges): Ditto.
18148 * loop-unswitch.c (unswitch_loop): Ditto.
18149 * cgraphclones.c (cgraph_clone_edge): Ditto.
18150 (cgraph_clone_node): Ditto.
18151 * tree-inline.c (copy_bb): Ditto.
18152 (copy_edges_for_bb): Ditto.
18153 (initialize_cfun): Ditto.
18154 (copy_cfg_body): Ditto.
18155 (expand_call_inline): Ditto.
18157 2013-04-08 Kai Tietz <ktietz@redhat.com>
18159 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
18160 TARGET_CYGWIN64 by TARGET_64BIT.
18162 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
18164 * config/epiphany/epiphany.md (GPR_1): New constant.
18165 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
18166 * config/epiphany/epiphany.c (gen_compare_reg):
18167 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
18168 is already in place.
18169 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
18170 Don't require being called during rtl expansion; If y operlaps r0,
18172 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
18173 (epiphany_expand_epilogue): Likewise.
18175 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
18176 Don't use CC_FPmode for ORDERED / UNORDERED.
18177 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
18179 * config/epiphany/constraints.md (CnL): New constraint.
18180 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
18181 * config/epiphany/predicates.md (add_operand): Allow 1024.
18183 * config/epiphany/epiphany.md (logical_op): New code iterator.
18184 (op_mnc): New code attribute.
18185 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
18186 (mov_f+1, mov_f+2): New peephole2 patterns.
18188 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
18189 (cstoresi4): Also allow re-use of zero result when doing a NE
18190 comparison to a non-zero operand.
18191 Use (clobber (scratch)) for first insn if the gpr output is not needed.
18193 * config/epiphany/epiphany.md (<insn_opname>v2si3):
18194 Use gen_addsi3_i / gen_subsi3_i.
18196 2013-04-08 Jakub Jelinek <jakub@redhat.com>
18200 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
18201 contain anything but clobbers, at most one __builtin_stack_restore,
18202 optionally debug stmts and final resx, and if it has at least one
18203 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
18204 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
18205 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
18206 which isn't defaut definition, remove them.
18207 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
18208 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
18209 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
18210 with MEM_REF LHS with SSA_NAME address.
18212 2013-04-08 Jeff Law <law@redhat.com>
18214 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
18216 2013-04-08 Richard Biener <rguenther@suse.de>
18218 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
18220 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
18221 determined vector type.
18222 (vect_analyze_data_refs): Likewise.
18223 (vect_get_new_vect_var): Adjust.
18224 (vect_create_destination_var): Preserve SSA name versions.
18225 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
18226 not dump anything here.
18228 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
18230 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
18231 Add member lr_slot_known.
18232 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
18234 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
18235 Remove code that sets lr_slot_offset according to what a previous
18236 version of epiphany_emit_save_restore used to do.
18237 (epiphany_emit_save_restore): When doing an lr save or restore,
18238 set/verify lr_slot_known and lr_slot_offset.
18240 2013-04-08 Xinyu Qi <xyqi@marvell.com>
18243 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
18246 2013-04-08 Richard Biener <rguenther@suse.de>
18248 * alias.c (find_base_term): Fix thinko in previous change.
18250 2013-04-08 Jakub Jelinek <jakub@redhat.com>
18252 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
18253 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
18254 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
18255 if possible to compute val.
18256 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
18257 For QImode integers don't require anything about precision. Use
18258 const_with_all_bytes_same to find out if the constant doesn't have
18259 repeated bytes in it.
18261 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18263 * config/s390/s390.c (s390_expand_insv): Only accept insertions
18266 2013-04-08 Marek Polacek <polacek@redhat.com>
18268 PR rtl-optimization/48182
18269 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
18272 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18275 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
18276 nuses, make sure we have a label.
18278 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18281 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
18282 (rs6000_emit_swdiv_low_precision): Remove.
18283 (rs6000_emit_swdiv): Rewrite to handle between one and four
18284 iterations of Newton-Raphson generally; modify required number of
18285 iterations for some cases.
18286 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
18288 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
18290 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
18291 set-but-unused variable.
18293 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
18294 basic blocks of released function bodies garbage-collectable.
18296 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
18297 (struct rtl_opt_pass): Add TODO_df_finish.
18299 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
18301 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18303 * config/arm/constraints.md (q): New constraint.
18304 * config/arm/ldrdstrd.md: New file.
18305 * config/arm/arm.md (ldrdstrd.md) New include.
18306 (arm_movdi): Use "q" instead of "r" constraint
18307 for double-word memory access.
18308 (movdf_soft_insn): Likewise.
18309 * config/arm/vfp.md (movdi_vfp): Likewise.
18310 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
18311 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
18312 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
18313 (mem_ok_for_ldrd_strd): Likewise.
18314 (output_move_double): Update assertion.
18316 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18318 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
18320 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18322 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
18323 define_insn_and_split.
18324 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
18326 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18328 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
18329 define_insn_and_split.
18330 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
18331 (shiftsi3_compare): New pattern.
18332 (rrx): New pattern.
18333 * config/arm/unspecs.md (UNSPEC_RRX): New.
18335 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18337 * config/arm/arm.md (negdi_extendsidi): New pattern.
18338 (negdi_zero_extendsidi): Likewise.
18340 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18342 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
18343 define_insn_and_split.
18344 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
18345 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
18347 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18349 * config/arm/arm.md (arm_subdi3): Convert define_insn into
18350 define_insn_and_split.
18351 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
18352 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
18354 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18356 * config/arm/arm.md (subsi3_carryin): New pattern.
18357 (subsi3_carryin_const): Likewise.
18358 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
18359 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
18361 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18363 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
18365 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18367 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
18368 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
18370 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18372 * config/arm/arm.c (arm_expand_builtin): Change fcode
18373 type to unsigned int.
18375 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18377 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
18379 2013-04-04 Ian Lance Taylor <iant@google.com>
18381 * doc/standards.texi (Standards): The Go frontend supports the Go 1
18384 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
18386 PR middle-end/56729
18387 * df-scan.c (df_insn_delete): Disable failing assert.
18389 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18391 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
18392 New function prototype.
18393 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
18394 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
18395 (arm_builtin_vectorized_function): New function.
18397 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18399 * config/arm/arm_neon_builtins.def: New file.
18400 * config/arm/arm.c (neon_builtin_data): Move contents to
18401 arm_neon_builtins.def.
18402 (enum arm_builtins): Include neon builtin definitions.
18403 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
18404 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
18406 2013-04-04 Marek Polacek <polacek@redhat.com>
18408 PR tree-optimization/48186
18409 * predict.c (maybe_hot_frequency_p): Return false if
18410 HOT_BB_FREQUENCY_FRACTION is 0.
18411 (cgraph_maybe_hot_edge_p): Likewise.
18413 2013-04-04 Richard Biener <rguenther@suse.de>
18415 PR tree-optimization/56826
18416 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
18419 2013-04-04 Richard Biener <rguenther@suse.de>
18421 PR tree-optimization/56213
18422 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
18423 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
18425 2013-04-04 Richard Biener <rguenther@suse.de>
18427 PR tree-optimization/56837
18428 * tree-loop-distribution.c (classify_partition): For non-zero
18429 values require that the value has the same precision as its
18430 mode to be useful as memset value.
18432 2013-04-03 Nick Clifton <nickc@redhat.com>
18434 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
18435 (fmssf4): Use fmsf.s on E3V5 architectures.
18436 (fnmasf4): Use fnmaf.s on E3V5 architectures.
18437 (fnmssf4): Use fnmsf.s on E3V5 architectures.
18439 2013-04-03 Jeff Law <law@redhat.com>
18441 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
18442 (lra-eliminations.o): Likewise.
18444 2013-04-03 Teresa Johnson <tejohnson@google.com>
18446 * gcov-io.c (compute_working_sets): Moved most of body of old
18447 compute_working_sets here from profile.c.
18448 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
18449 (gcov_working_set_t): Moved typedef here from basic-block.h
18450 (compute_working_set): Declare.
18451 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
18452 (get_working_sets): Renamed from compute_working_set,
18453 replace most of body with call to new compute_working_sets.
18454 (get_exec_counts): Replace call to compute_working_sets
18455 to get_working_sets.
18456 * profile.h (get_working_sets): Renamed from compute_working_set.
18457 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
18458 to get_working_sets.
18459 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
18460 * gcov-dump.c (dump_working_sets): New function.
18462 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
18464 * hwint.c (sext_hwi, zext_hwi): New functions.
18465 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
18466 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
18467 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
18468 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
18469 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
18470 (sext_hwi, zext_hwi): New functions.
18472 2013-04-03 Jeff Law <law@redhat.com>
18474 PR tree-optimization/56799
18475 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
18476 back test for widening conversion erroneously dropped in prior change.
18478 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18481 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
18482 instead of next_real_insn.
18484 2013-04-03 Marek Polacek <polacek@redhat.com>
18487 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
18489 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18492 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
18494 (thumb1_output_casesi): Likewise.
18495 (thumb2_output_casesi): Likewise.
18497 2013-04-03 Richard Biener <rguenther@suse.de>
18499 PR tree-optimization/56817
18500 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
18502 (tree_unroll_loops_completely_1): ... new function to manually
18503 walk the loop tree, properly defering outer loops of unrolled
18504 loops to later iterations.
18506 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
18508 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
18509 (vectorizable_load): Likewise.
18510 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
18511 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
18513 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
18515 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
18518 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
18520 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
18522 2013-04-03 Bin Cheng <bin.cheng@arm.com>
18524 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
18526 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
18528 PR tree-optimization/56790
18529 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
18532 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
18534 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
18536 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
18537 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
18540 2013-04-03 Jakub Jelinek <jakub@redhat.com>
18543 * tree.h (force_folding_builtin_constant_p): New decl.
18544 * builtins.c (force_folding_builtin_constant_p): New variable.
18545 (fold_builtin_constant_p): Fold immediately also if
18546 force_folding_builtin_constant_p.
18548 2013-04-03 Richard Biener <rguenther@suse.de>
18550 PR tree-optimization/56812
18551 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
18552 DRs of the same interleaving chain are independent.
18554 2013-04-02 Jason Merrill <jason@redhat.com>
18556 * gdbinit.in (pbb): Use debug fn.
18558 2013-04-02 Lawrence Crowl <crowl@google.com>
18560 * sese.h (struct ivtype_map_elt_s): Remove unused.
18561 (extern debug_ivtype_map): Remove unused.
18562 (extern eq_ivtype_map_elts): Remove unused.
18563 * sese.c (debug_ivtype_map): Removed unused.
18564 (debug_ivtype_map_1): Removed unused.
18565 (debug_ivtype_elt): Remove unused.
18566 (eq_ivtype_map_elts): Remove unused.
18569 2013-04-02 Kai Tietz <ktietz@redhat.com>
18572 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
18573 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
18574 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
18576 (legitimize_pe_coff_symbol): Likewise.
18577 (is_imported_p): New helper-function.
18578 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
18579 for Windows x64 targets.
18580 (ix86_expand_prologue): Optimize for pe-coff targets.
18581 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
18582 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
18583 medium/large code-model.
18584 (legitimize_pic_address): Likewise.
18585 (legitimize_tls_address): Likewise.
18586 (ix86_expand_call): Likewise.
18587 (x86_output_mi_thunk): Likewise.
18588 (get_dllimport_decl): Add new beimport argument.
18589 (construct_plt_address): Don't assert for x64 pe-coff targets.
18590 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
18592 (SYMBOL_FLAG_STUBVAR): New macro.
18593 (SYMBOL_REF_STUBVAR_P): Likewise.
18594 * config/i386/winnt.c (stub_list): New structure.
18595 (stub_head): New local variable.
18596 (i386_pe_record_stub): New function.
18597 (i386_pe_file_end): Emit refptr-stubs.
18599 2013-04-02 Jakub Jelinek <jakub@redhat.com>
18601 PR rtl-optimization/56745
18602 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
18603 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
18606 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
18607 and both of them are MEM_REFs, just compare first argument for
18608 equality and attempt to deal even with differing offsets.
18611 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
18612 of gimple_clobber_p to be MEM_REF.
18613 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
18614 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
18615 after gimplification.
18616 * asan.c (get_mem_ref_of_assignment): Don't instrument
18617 gimple_clobber_p stmts.
18618 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
18619 gimple_clobber_p stmt if they have MEM_REF lhs and
18620 are dead because of another gimple_clobber_p stmt.
18621 * tree-ssa-live.c (clear_unused_block_pointer): Treat
18622 gimple_clobber_p stmts like debug stmts.
18623 (remove_unused_locals): Remove clobbers with MEM_REF lhs
18624 that refer to unused VAR_DECLs or uninitialized values.
18625 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
18626 gimple_clobber_p stmts if they refer to removed parameters.
18627 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
18630 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
18632 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
18633 using SWI48 mode attribute.
18635 2013-04-02 Wei Mi <wmi@google.com>
18637 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
18638 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
18639 *<rotate_insn><mode>3_mask in i386.md.
18641 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
18643 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
18645 2013-04-02 Richard Biener <rguenther@suse.de>
18647 PR tree-optimization/56778
18648 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
18649 Runtime alias tests are not supported for gather loads.
18650 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
18651 stmts referenced from SSA operands before updating SSA form.
18653 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
18654 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18656 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
18657 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
18658 * config/arm/cortex-a53.md: New file.
18659 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
18660 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
18661 * config/arm/arm.c (arm_issue_rate): Likewise.
18662 * config/arm/arm-tune.md: Regenerate
18663 * config/arm/arm-tables.opt: Regenerate.
18664 * config/arm/arm-cores.def: Add cortex-a53.
18666 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
18668 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
18671 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
18673 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
18674 scalar load/store operations using B/H registers.
18675 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
18677 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
18679 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
18681 * config/aarch64/aarch64.c
18682 (aarch64_simd_scalar_immediate_valid_for_move): New.
18683 * config/aarch64/aarch64-protos.h
18684 (aarch64_simd_scalar_immediate_valid_for_move): New.
18685 * config/aarch64/constraints.md (Dh, Dq): New.
18686 * config/aarch64/iterators.md (hq): New.
18688 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
18690 * reorg.c (get_branch_condition): Deal with conditional returns.
18691 (fill_simple_delay_slots): Remove dead code dealing with jumps.
18693 2013-04-01 Wei Mi <wmi@google.com>
18695 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
18696 Truncate operand 2 using %b asm operand modifier.
18697 (*<shift_insn><mode>3_mask): Ditto.
18698 (*<rotate_insn><mode>3_mask): Ditto.
18700 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
18702 PR middle-end/56798
18703 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
18705 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
18707 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
18709 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
18711 2013-03-30 Lawrence Crowl <crowl@google.com>
18713 * dse.c (clear_alias_sets): Remove never set.
18714 (disqualified_clear_alias_sets): Remove never set.
18715 (clear_alias_mode_pool): Remove never set.
18716 (dse_step0): Remove condition that is never true.
18717 (canon_address): Remove condition that is never true.
18718 (dse_step7): Remove condition that is never true.
18719 (rest_of_handle_dse): Remove condition that is never true.
18720 (rest_of_handle_dse::did_global): Remove never read from above.
18721 (dse_step2_spill): Remove never called from above.
18722 (dse_step5_spill): Remove never called from above.
18724 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
18726 * doc/md.texi (Standard Names) <casesi>: Update documentation for
18727 JUMP_TABLE_DATA changes.
18728 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
18729 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
18730 (Insns) <jump_table_data>: New entry.
18731 * doc/tm.texi: Regenerate.
18733 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
18735 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
18736 for table jump at the end of a basic block using tablejump_p.
18737 * targhooks.c (default_invalid_within_doloop): Likewise.
18738 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
18739 target hook implementation that is identical to the default hook.
18740 (rs6000_invalid_within_doloop): Remove.
18742 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
18743 unused variable from tablejump_p call.
18745 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
18746 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
18747 (INSN_DELETED_P): Likewise.
18748 (emit_jump_table_data): New prototype.
18749 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
18750 after 4th as unused.
18751 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
18752 * sched-vis.c (print_insn): Likewise.
18753 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
18754 insn for compatibility with back ends that use next_active_insn to
18755 identify jump table data.
18756 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
18757 (remove_insn): Likewise.
18758 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
18760 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
18761 (emit_jump_table_data): New function.
18763 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
18764 basic block, a JUMP_TABLE_DATA never is.
18765 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
18766 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
18767 off from code handling real insns.
18768 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
18769 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
18770 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
18771 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
18772 is not a NONDEBUG_INSN_P.
18773 * ira-costs.c (scan_one_insn): Likewise.
18774 * jump.c (mark_all_labels): Likewise.
18775 (mark_jump_label_1): Likewise.
18776 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
18777 * lra.c (get_insn_freq): Expect all insns reaching here to be in
18779 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
18780 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
18781 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
18782 JUMP_TABLE_DATA_P insns.
18783 (calculate_elim_costs_all_insns): Likewise.
18784 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
18785 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
18786 (delete_output_reload): Code style fixups.
18787 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
18788 insn flags on this non-insn.
18789 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
18790 as scheduling barriers, for pre-change compatibility.
18791 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
18792 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
18794 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
18795 redundant JUMP_TABLE_DATA_P test.
18796 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
18797 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
18798 (frv_for_each_packet): Likewise.
18799 * config/i386/i386.c (min_insn_size): Likewise.
18800 (ix86_avoid_jump_mispredicts): Likewise.
18801 * config/m32r/m32r.c (m32r_is_insn): Likewise.
18802 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
18803 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
18804 (mips16_insn_length): Robustify.
18805 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
18806 (mips16_split_long_branches): Likewise.
18807 * config/pa/pa.c (pa_combine_instructions): Likewise.
18808 * config/rs6000/rs6000.c (get_next_active_insn): Treat
18809 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
18810 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
18811 as contributing to pool range lengths.
18812 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
18813 Remove redundant JUMP_TABLE_DATA_P test.
18814 (sh_loop_align): Likewise.
18815 (split_branches): Likewise.
18816 (sh_insn_length_adjustment): Likewise.
18817 * config/spu/spu.c (get_branch_target): Likewise.
18819 2013-03-29 Jan Hubicka <jh@suse.cz>
18821 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
18822 gcov streaming; stream hot bb threshold to ltrans.
18823 * predict.c (get_hot_bb_threshold): Break out from ....
18824 (maybe_hot_count_p): ... here.
18825 (set_hot_bb_threshold): New function.
18826 * lto-section-in.c (lto_section_name): Add profile.
18827 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
18828 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
18829 and data-streamer.h
18830 (histogram_entry): New structure.
18831 (histogram, histogram_pool): New global vars.
18832 (histogram_hash): New structure.
18833 (histogram_hash::hash): New method.
18834 (histogram_hash::equal): Likewise.
18835 (account_time_size): New function.
18836 (cmp_counts): New function.
18837 (dump_histogram): New function.
18838 (ipa_profile_generate_summary): New function.
18839 (ipa_profile_write_summary): New function.
18840 (ipa_profile_read_summary): New function.
18841 (ipa_profile): Decide on threshold.
18842 (pass_ipa_profile): Add ipa_profile_write_summary and
18843 ipa_profile_read_summary.
18844 * Makefile.in (ipa.o): Update dependencies.
18845 * lto-streamer.h (LTO_section_ipa_profile): New section.
18847 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
18849 * tree.h (VAR_P): New.
18851 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
18854 * doc/invoke.texi ([-fwhole-program]): Fix typo.
18856 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
18858 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
18859 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
18860 (control_flow_insn_p): Likewise.
18861 * cfgrtl.c (duplicate_insn_chain): Likewise.
18862 * final.c (get_attr_length_1): Likewise.
18863 (shorten_branches): Likewise.
18864 (final_scan_insn): Likewise.
18865 * function.c (instantiate_virtual_regs): Likewise.
18866 * gcse.c (insert_insn_end_basic_block): Likewise.
18867 * ira-costs.c (scan_one_insn): Likewise.
18868 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
18869 * lra.c (check_rtl): Likewise.
18870 * reload1.c (elimination_costs_in_insn): Likewise.
18871 * reorg.c (follow_jumps): Likewise.
18873 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
18874 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
18875 (thumb_far_jump_used_p): Likewise.
18876 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
18877 (workaround_speculation): Likewise.
18878 (add_sched_insns_for_speculation): Likewise.
18879 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
18880 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
18881 (frv_for_each_packet): Likewise.
18882 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
18883 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
18884 (final_emit_insn_group_barriers): Likewise.
18885 * config/m32r/m32r.c (m32r_is_insn): Likewise.
18886 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
18887 (mips16_insn_length): Likewise.
18888 * config/pa/pa.c (pa_reorg): Likewise.
18889 (pa_combine_instructions): Likewise.
18890 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
18891 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
18892 (sh_reorg): Likewise.
18893 (split_branches): Likewise.
18894 * config/spu/spu.c (get_branch_target): Likewise.
18896 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
18899 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
18901 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
18902 Fix declaration name.
18904 2013-03-28 Lawrence Crowl <crowl@google.com>
18906 * graphds.h (struct graph.indicies): Remove unused.
18907 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
18908 (SCOP_ORIGINAL_PDDRS): Remove unused.
18909 * sese.h (extern insert_loop_close_phis): Removed unused.
18910 (extern insert_guard_phis): Removed unused.
18911 (extern ivtype_map_elt_info): Removed unused.
18912 (new_ivtype_map_elt): Removed unused.
18913 * sese.c (ivtype_map_elt_info): Removed unused.
18915 2013-03-28 Lawrence Crowl <crowl@google.com>
18917 * Makefile.in: Add several missing include dependences.
18919 (test-dump.o): New. This object is not added to any executable,
18920 but is present for ad-hoc testing.
18922 (debug (const bitmap_head_def &)): New.
18923 (debug (const bitmap_head_def *)): New.
18925 (extern debug (const bitmap_head_def &)): New.
18926 (extern debug (const bitmap_head_def *)): New.
18928 (debug (edge_def &)): New.
18929 (debug (edge_def *)): New.
18931 (debug (basic_block_def &)): New.
18932 (debug (basic_block_def *)): New.
18934 (dump_node (const_tree, int, FILE *)): Correct source file.
18936 (debug (die_struct &)): New.
18937 (debug (die_struct *)): New.
18939 (extern debug (die_struct &)): New.
18940 (extern debug (die_struct *)): New.
18941 * gimple-pretty-print.c
18942 (debug (gimple_statement_d &)): New.
18943 (debug (gimple_statement_d *)): New.
18944 * gimple-pretty-print.h
18945 (extern debug (gimple_statement_d &)): New.
18946 (extern debug (gimple_statement_d *)): New.
18948 (debug (ira_allocno_copy &)): New.
18949 (debug (ira_allocno_copy *)): New.
18950 (debug (ira_allocno &)): New.
18951 (debug (ira_allocno *)): New.
18953 (extern debug (ira_allocno_copy &)): New.
18954 (extern debug (ira_allocno_copy *)): New.
18955 (extern debug (ira_allocno &)): New.
18956 (extern debug (ira_allocno *)): New.
18958 (debug (live_range &)): New.
18959 (debug (live_range *)): New.
18961 (debug (lra_live_range &)): New.
18962 (debug (lra_live_range *)): New.
18964 (debug (lra_live_range &)): New.
18965 (debug (lra_live_range *)): New.
18967 (debug (omega_pb_d &)): New.
18968 (debug (omega_pb_d *)): New.
18970 (extern debug (omega_pb_d &)): New.
18971 (extern debug (omega_pb_d *)): New.
18973 (debug (const rtx_def &)): New.
18974 (debug (const rtx_def *)): New.
18976 (debug_tree (tree): Move within file.
18977 (debug_raw (const tree_node &)): New.
18978 (debug_raw (const tree_node *)): New.
18979 (dump_tree_via_hooks (const tree_node *, int)): New.
18980 (debug (const tree_node &)): New.
18981 (debug (const tree_node *)): New.
18982 (debug_verbose (const tree_node &)): New.
18983 (debug_verbose (const tree_node *)): New.
18984 (debug_head (const tree_node &)): New.
18985 (debug_head (const tree_node *)): New.
18986 (debug_body (const tree_node &)): New.
18987 (debug_body (const tree_node *)): New.
18988 (debug_vec_tree (tree): Move and reimplement in terms of dump.
18989 (debug (vec<tree, va_gc> &)): New.
18990 (debug (vec<tree, va_gc> *)): New.
18992 (extern debug (const rtx_def &)): New.
18993 (extern debug (const rtx_def *)): New.
18995 (debug_raw (simple_bitmap_def &)): New.
18996 (debug_raw (simple_bitmap_def *)): New.
18997 (debug (simple_bitmap_def &)): New.
18998 (debug (simple_bitmap_def *)): New.
19000 (extern debug (simple_bitmap_def &)): New.
19001 (extern debug (simple_bitmap_def *)): New.
19002 (extern debug_raw (simple_bitmap_def &)): New.
19003 (extern debug_raw (simple_bitmap_def *)): New.
19005 (debug (vinsn_def &)): New.
19006 (debug (vinsn_def *)): New.
19007 (debug_verbose (vinsn_def &)): New.
19008 (debug_verbose (vinsn_def *)): New.
19009 (debug (expr_def &)): New.
19010 (debug (expr_def *)): New.
19011 (debug_verbose (expr_def &)): New.
19012 (debug_verbose (expr_def *)): New.
19013 (debug (vec<rtx> &)): New.
19014 (debug (vec<rtx> *)): New.
19016 (extern debug (vinsn_def &)): New.
19017 (extern debug (vinsn_def *)): New.
19018 (extern debug_verbose (vinsn_def &)): New.
19019 (extern debug_verbose (vinsn_def *)): New.
19020 (extern debug (expr_def &)): New.
19021 (extern debug (expr_def *)): New.
19022 (extern debug_verbose (expr_def &)): New.
19023 (extern debug_verbose (expr_def *)): New.
19024 (extern debug (vec<rtx> &)): New.
19025 (extern debug (vec<rtx> *)): New.
19027 (_list_iter_cond_expr): Make inline instead of static.
19029 (debug (sreal &)): New.
19030 (debug (sreal *)): New.
19032 (extern debug (sreal &)): New.
19033 (extern debug (sreal *)): New.
19035 (extern debug_raw (const tree_node &)): New.
19036 (extern debug_raw (const tree_node *)): New.
19037 (extern debug (const tree_node &)): New.
19038 (extern debug (const tree_node *)): New.
19039 (extern debug_verbose (const tree_node &)): New.
19040 (extern debug_verbose (const tree_node *)): New.
19041 (extern debug_head (const tree_node &)): New.
19042 (extern debug_head (const tree_node *)): New.
19043 (extern debug_body (const tree_node &)): New.
19044 (extern debug_body (const tree_node *)): New.
19045 (extern debug (vec<tree, va_gc> &)): New.
19046 (extern debug (vec<tree, va_gc> *)): New.
19048 (debug (struct loop &)): New.
19049 (debug (struct loop *)): New.
19050 (debug_verbose (struct loop &)): New.
19051 (debug_verbose (struct loop *)): New.
19052 * tree-dump.c: Add header dependence.
19054 (extern debug (struct loop &)): New.
19055 (extern debug (struct loop *)): New.
19056 (extern debug_verbose (struct loop &)): New.
19057 (extern debug_verbose (struct loop *)): New.
19059 (debug (data_reference &)): New.
19060 (debug (data_reference *)): New.
19061 (debug (vec<data_reference_p> &)): New.
19062 (debug (vec<data_reference_p> *)): New.
19063 (debug (vec<ddr_p> &)): New.
19064 (debug (vec<ddr_p> *)): New.
19066 (extern debug (data_reference &)): New.
19067 (extern debug (data_reference *)): New.
19068 (extern debug (vec<data_reference_p> &)): New.
19069 (extern debug (vec<data_reference_p> *)): New.
19070 (extern debug (vec<ddr_p> &)): New.
19071 (extern debug (vec<ddr_p> *)): New.
19073 (debug (pt_solution &)): New.
19074 (debug (pt_solution *)): New.
19076 (extern debug (pt_solution &)): New.
19077 (extern debug (pt_solution *)): New.
19079 (debug (_var_map &)): New.
19080 (debug (_var_map *)): New.
19081 (debug (tree_live_info_d &)): New.
19082 (debug (tree_live_info_d *)): New.
19084 (extern debug (_var_map &)): New.
19085 (extern debug (_var_map *)): New.
19086 (extern debug (tree_live_info_d &)): New.
19087 (extern debug (tree_live_info_d *)): New.
19089 2013-03-28 Jan Hubicka <jh@suse.cz>
19091 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
19093 2013-03-28 Ian Bolton <ian.bolton@arm.com>
19095 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
19096 record only when desired or required.
19098 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
19100 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
19101 *vec_extractv2di_1_rex64. Use x64 isa attribute.
19103 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19105 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
19106 (*andsi3_compare0_uxtw): New pattern.
19107 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
19108 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
19110 2013-03-28 Jan Hubicka <jh@suse.cz>
19112 * data-streamer-in.c (streamer_read_gcov_count): New function.
19113 * gimple-streamer-out.c: Include value-prof.h.
19114 (output_gimple_stmt): Output histogram.
19115 (output_bb): Use streamer_write_gcov_count.
19116 * value-prof.c: Include data-streamer.h
19117 (dump_histogram_value): Add HIST_TYPE_MAX.
19118 (stream_out_histogram_value): New function.
19119 (stream_in_histogram_value): New function.
19120 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
19121 (stream_out_histogram_value, stream_in_histogram_value): Declare.
19122 * data-streamer-out.c (streamer_write_gcov_count): New function.
19123 (streamer_write_gcov_count_stream): New function.
19124 * lto-cgraph.c (lto_output_edge): Update counter streaming.
19125 (lto_output_node): Likewise.
19126 (input_node, input_edge): Likewise.
19127 * lto-streamer-out.c (output_cfg): Update streaming.
19128 * lto-streamer-in.c (input_cfg): Likewise.
19129 * data-streamer.h (streamer_write_gcov_count,
19130 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
19131 * gimple-streamer-in.c: Include value-prof.h
19132 (input_gimple_stmt): Input histograms.
19133 (input_bb): Update profile streaming.
19135 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
19137 * genmodes.c (emit_max_int): New function.
19138 (emit_insn_modes_h): Added call to emit_max_function.
19139 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
19141 * machmode.def: Fixed comment.
19143 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
19145 * combine.c (try_combine): Removed useless assert.
19146 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
19148 2013-03-28 Marek Polacek <polacek@redhat.com>
19149 Richard Biener <rguenther@suse.de>
19151 PR tree-optimization/56695
19152 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
19153 build signed result of a vector comparison.
19154 * tree-cfg.c (verify_gimple_comparison): Check that a result
19155 of a vector comparison has signed type.
19157 2013-03-28 Richard Biener <rguenther@suse.de>
19159 PR tree-optimization/37021
19160 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
19161 do not restrict gaps between groups.
19162 * tree-vect-stmts.c (vectorizable_load): Properly account for
19163 a gap between groups.
19165 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
19167 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
19168 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
19171 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
19173 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
19174 * doc/extend.texi (Named Address Spaces): Ditto.
19175 (Variable Attributes): Ditto.
19177 2013-03-27 Kai Tietz <ktietz@redhat.com>
19179 * config.build: Add support for cygwin x64 target.
19180 * config.gcc: Likewise.
19181 * config.host: Likewise.
19182 * configure.ac: Likewise
19183 * configure: Regenerated.
19185 2013-03-27 Kai Tietz <ktietz@redhat.com>
19187 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
19188 * config/i386/t-cygwin-w64: New file.
19189 * config/i386/cygwin-w64.h: New file.
19190 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
19191 and add support for x64-cygwin target.
19192 (CPP_SPEC): Likewise.
19193 (CXX_WRAP_SPEC_LIST): Undefine before define.
19194 (LIBGCJ_SONAME): Use 15 as version.
19196 2013-03-27 Richard Biener <rguenther@suse.de>
19198 PR tree-optimization/56716
19199 * tree-ssa-structalias.c (perform_var_substitution): Adjust
19200 dumping for ref nodes.
19202 2013-03-27 Martin Jambor <mjambor@suse.cz>
19204 PR tree-optimization/55334
19205 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
19206 restricted pointers to arrays.
19208 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
19210 * Makefile.in (.SUFFIXES): Add .cc.
19211 (.c.o): Apply same recipe for implicit rule .cc.o.
19213 2013-03-27 Richard Biener <rguenther@suse.de>
19215 PR tree-optimization/37021
19216 * tree-vect-data-refs.c (vect_check_strided_load): Allow
19217 REALPART/IMAGPART_EXPRs around the supported refs.
19218 * tree-ssa-structalias.c (find_func_aliases): Assume that
19219 floating-point values are not used to transfer pointers.
19221 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
19223 * target.def (TARGET_HAS_IFUNC_P): New target hook.
19224 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
19225 * doc/tm.texi: Regenerate.
19226 * targhooks.h (default_has_ifunc_p): New.
19227 * targhooks.c (default_has_ifunc_p): Ditto.
19228 * config/linux-protos.h: New file.
19229 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
19230 hook for linux which disables support of indirect functions in android.
19231 * config/linux-android.c: New file.
19232 * config/t-linux-android.c: Ditto.
19233 * config.gcc: Added new object file linux-android.o.
19234 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
19235 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
19236 * varasm.c (do_assemble_alias): Likewise.
19237 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
19238 doesn't support indirect functions.
19239 * configure: Regenerate.
19241 2013-03-27 Bin Cheng <bin.cheng@arm.com>
19244 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
19245 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
19248 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19250 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
19252 2013-03-27 Terry Guo <terry.guo@arm.com>
19254 * config/arm/arm-cores.def: Added core cortex-r7.
19255 * config/arm/arm-tune.md: Regenerated.
19256 * config/arm/arm-tables.opt: Regenerated.
19257 * doc/invoke.texi: Added entry for core cortex-r7.
19259 2013-03-27 Walter Lee <walt@tilera.com>
19261 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
19262 double-decrement of next_scratch_regno.
19264 2013-03-27 Walter Lee <walt@tilera.com>
19266 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
19268 (insn_v1mulus): Ditto.
19269 (insn_v2muls): Ditto.
19271 2013-03-27 Walter Lee <walt@tilera.com>
19273 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
19274 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
19276 2013-03-27 Walter Lee <walt@tilera.com>
19278 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
19279 (*sibcall_value): Ditto.
19281 2013-03-27 Walter Lee <walt@tilera.com>
19283 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
19284 (insn_mnz_v8qi): ... this ...
19285 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
19287 (insn_v<n>mnz): Replaced by ...
19288 (insn_v1mnz): ... this ...
19289 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
19291 (insn_mz_<mode>): Replaced by ...
19292 (insn_mz_v8qi): ... this ...
19293 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
19295 (insn_v<n>mz): Replaced by ...
19296 (insn_v1mz): ... this ...
19297 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
19300 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
19302 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
19304 2013-03-26 Roland McGrath <mcgrathr@google.com>
19306 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
19307 than fprintf with a non-constant, non-format string.
19309 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
19311 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
19312 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
19313 operand 0 predicate.
19314 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
19315 attribute. Use general_x64nomem_operand as operand 1 predicate.
19316 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
19317 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
19318 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
19319 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
19320 (mov<mode>_insv_1): Remove expander. Merge insn with
19321 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
19322 Use general_x64nomem_operand as operand 1 predicate.
19323 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
19324 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
19325 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
19326 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
19327 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
19328 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
19329 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
19330 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
19331 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
19332 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
19333 (general_x64nomem_operand): Ditto.
19335 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
19337 * config/rtems.opt: Add -pthread option.
19339 2013-03-26 Richard Biener <rguenther@suse.de>
19341 * alias.c (find_base_term): Avoid redundant and not used recursion.
19342 (base_alias_check): Get the initial base term from the caller.
19343 (true_dependence_1): Compute and pass base terms to base_alias_check.
19344 (write_dependence_p): Likewise.
19345 (may_alias_p): Likewise.
19347 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
19349 * config/aarch64/aarch64.c (aarch64_classify_address): Support
19350 PC-relative load in SI modes and above only.
19352 2013-03-26 Xinyu Qi <xyqi@marvell.com>
19354 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
19355 * config/arm/iwmmxt.md (WCGR0): Update.
19356 (WCGR1, WCGR2, WCGR3): Likewise.
19358 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
19360 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
19361 Use x64 and nox64 isa attributes.
19363 2013-03-26 Richard Biener <rguenther@suse.de>
19365 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
19366 alignment computations and rely on get_object_alignment_1
19367 for the !TYPE_P case.
19368 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
19370 2013-03-26 Walter Lee <walt@tilera.com>
19372 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
19373 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
19375 2013-03-25 Jeff Law <law@redhat.com>
19377 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
19378 check for INTEGRAL_TYPE_P that was missing due to checking in
19379 wrong version of prior patch.
19381 2013-03-25 Walter Lee <walt@tilera.com>
19383 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
19384 TILEGX_INSN_SHUFFLEBYTES1.
19385 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
19387 (tilegx_builtins): Ditto.
19388 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
19390 2013-03-25 Walter Lee <walt@tilera.com>
19392 * config/tilegx/tilegx.md (floatsisf2): New pattern.
19393 (floatunssisf2): New pattern.
19394 (floatsidf2): New pattern.
19395 (floatunssidf2): New pattern.
19397 2013-03-25 Walter Lee <walt@tilera.com>
19399 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
19400 tests for constraint J, K, N, P.
19402 2013-03-25 Walter Lee <walt@tilera.com>
19404 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
19405 Use indirect/pcrel encoding.
19406 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
19409 2013-03-25 Steve Ellcey <sellcey@mips.com>
19411 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
19412 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
19413 * config/mips/mips.c (mips_option_override): Set IMADD default.
19414 * config/mips/mips.h (PTF_AVOID_IMADD): New.
19415 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
19416 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
19417 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
19418 * doc/invoke.texi (-mimadd/-mno-imadd): New.
19420 2013-03-25 Jeff Law <law@redhat.com>
19422 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
19423 slightly to avoid creating and folding useless trees. Simplify
19424 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
19426 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
19428 * config/i386/i386.md (*zero_extendsidi2): Merge with
19429 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
19430 * config/i386/predicates.md (x86_64_zext_operand): Rename from
19431 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
19432 targets. Clarify comment.
19434 2013-03-25 Martin Jambor <mjambor@suse.cz>
19436 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
19437 pass-through jump functions differently.
19438 (ipa_read_jump_function): Likewise. Also use setter functions to set
19441 2013-03-25 Martin Jambor <mjambor@suse.cz>
19443 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
19444 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
19446 (ipa_get_indirect_edge_target): New function.
19447 (devirtualization_time_bonus): New parameter known_aggs, pass it to
19448 ipa_get_indirect_edge_target. Update all callers.
19449 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
19450 ipa_get_indirect_edge_target_1 instead of calling
19451 ipa_get_indirect_edge_target.
19452 (create_specialized_node): Pass aggvlas to
19453 ipcp_discover_new_direct_edges.
19455 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19457 * config/arm/arm.md (f_sels, f_seld): New types.
19458 (*cmov<mode>): New pattern.
19459 * config/arm/predicates.md (arm_vsel_comparison_operator): New
19462 2013-03-25 Kai Tietz <ktietz@redhat.com>
19464 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
19465 POSIX-printf for mingw-hosted builds.
19467 2013-03-25 Richard Biener <rguenther@suse.de>
19469 PR middle-end/56694
19470 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
19471 must-not-throw stmt location.
19473 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19475 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
19476 Emit load-acquire versions when acq is true.
19477 (arm_emit_store_exclusive): Add rel parameter.
19478 Emit store-release versions when rel is true.
19479 (arm_split_compare_and_swap): Use acquire-release instructions
19481 of barriers when appropriate.
19482 (arm_split_atomic_op): Likewise.
19483 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
19484 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
19485 (VUNSPEC_SLX): Likewise.
19486 (VUNSPEC_LDA): Likewise.
19487 (VUNSPEC_STL): Likewise.
19488 * config/arm/sync.md (atomic_load<mode>): New pattern.
19489 (atomic_store<mode>): Likewise.
19490 (arm_load_acquire_exclusive<mode>): Likewise.
19491 (arm_load_acquire_exclusivesi): Likewise.
19492 (arm_load_acquire_exclusivedi): Likewise.
19493 (arm_store_release_exclusive<mode>): Likewise.
19495 2013-03-25 Catherine Moore <clm@codesourcery.com>
19497 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
19498 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
19499 * config/mip/predicates.md (lwsp_swsp_operand,
19500 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
19501 sb16_operand, db4_operand, db7_operand, ib3_operand,
19502 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
19503 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
19504 andi16_operand): New predicates.
19505 * config/mips/mips.md (compression): New attribute.
19506 (enabled): New attribute.
19507 (length): Consider compression in computing length.
19508 (shift_compression): New code attribute.
19509 (*add<mode>3): New operands. Record compression.
19510 (sub<mode>3): Likewise.
19511 (one_cmpl<mode>2): Likewise.
19512 (*and<mode>3): Likewise.
19513 (*ior<mode>3): Likewise.
19514 (unnamed pattern for xor): Likewise.
19515 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
19516 (*<optab><mode>3): Likewise.
19517 (*mov<mode>_internal: Likewise.
19518 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
19519 (mips_unsigned_immediate_p): New.
19520 (umips_lwsp_swsp_address_p): New.
19521 (m16_based_address_p): New.
19522 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
19523 (mips_unsigned_immediate_p): New prototype.
19524 (lwsp_swsp_address_p): New prototype.
19525 (m16_based_address_p): New prototype.
19526 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
19527 (mips_signed_immediate_p): New function.
19528 (m16_based_address_p): New function.
19529 (lwsp_swsp_address_p): New function.
19530 (mips_print_operand_punctuation): Recognize short delay slot insns
19531 for microMIPS.add<mode>3"
19533 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19536 * config/arm/iterators.md (v_cmp_result): New mode attribute.
19537 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
19539 2013-03-25 Richard Biener <rguenther@suse.de>
19541 PR tree-optimization/56689
19542 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
19545 2013-03-25 Richard Biener <rguenther@suse.de>
19547 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
19549 (memory_references): Likewise.
19550 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
19551 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
19552 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
19553 (gather_mem_refs_in_loops): Fold into ...
19554 (analyze_memory_references): ... this. Move initialization
19555 to tree_ssa_lim_initialize.
19556 (fill_always_executed_in): Rename to ...
19557 (fill_always_executed_in_1): ... this.
19558 (fill_always_executed_in): Move contains_call computation to
19559 this new function from ...
19560 (tree_ssa_lim_initialize): ... here.
19561 (tree_ssa_lim): Call fill_always_executed_in.
19563 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
19565 * postreload.c (reload_combine): Fix code detecting returns.
19567 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
19569 * function.c (emit_use_return_register_into_block): On cc0 targets,
19570 do not emit the sequence between cc0 setter and user.
19572 2013-03-25 Kai Tietz <ktietz@redhat.com>
19574 * config/i386/predicates.md (local_symbolic_operand): Interpret
19575 dll-imported symbols as none-local.
19577 2013-03-25 Richard Biener <rguenther@suse.de>
19579 * tree-ssa-loop-im.c (struct depend): Remove.
19580 (struct lim_aux_data): Make depends a vec of gimples.
19581 (free_lim_aux_data): Adjust.
19582 (add_dependency): Likewise.
19583 (set_level): Likewise.
19585 2013-03-25 Richard Biener <rguenther@suse.de>
19587 PR middle-end/56434
19588 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
19589 the pointer returned by calls with ECF_MALLOC set.
19591 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
19593 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
19595 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
19597 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
19598 using MMXMODE mode iterator.
19599 (*move<mode>_internal): Merge with *movv2sf_internal and
19600 *movv2sf_internal_rex64 using MMXMODE mode iterator.
19602 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
19604 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
19605 (record_last_mem_set_info): Likewise.
19607 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
19608 of XNEWVEC followed by memset.
19609 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
19611 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
19613 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
19614 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
19615 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
19616 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
19617 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
19618 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
19619 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
19620 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
19621 BARRIER_P instead of GET_CODE.
19623 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
19625 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
19626 inaccuracy in the probing code.
19628 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
19629 (ctrapdi4): Likewise.
19631 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
19633 * calls.c (expand_call): Add missing guard to code handling return
19634 of non-BLKmode structures in MSB.
19635 * function.c (expand_function_end): Likewise.
19637 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
19639 * combine.c (try_combine): Adjust comment. Do not add the set of
19640 insn #0 if the destination indirectly is set or dies in insn #2.
19641 Tidy up code to distribute a new note.
19643 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
19645 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
19646 also for alternatives 16 and 17.
19648 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
19650 * config/i386/sse.md (*mov<mode>_internal): Merge with
19651 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
19652 Emit insn template depending on type attribute. Use
19653 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
19654 movd instead of movq mnemonic for interunit moves. Rewrite mode
19655 attribute calculation. Remove unit attribute calculation.
19656 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
19657 Set prefix_data16 attribute for DImode ssemov types.
19658 Use Ym instead of y for SSE-MMX conversion alternatives.
19659 Reorder operand constraints.
19661 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
19663 * df.h (df_insn_delete): Adjust prototype.
19664 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
19665 and let it decide whether mark the basic block dirty.
19666 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
19667 * df-scan.c (df_insn_info_delete): New helper function, split
19668 off from df_insn_delete.
19669 (df_scan_free_bb_info): Use it.
19670 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
19672 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
19673 that the insn is actually an insn and it has a non-NULL basic block.
19674 Do not mark basic block dirty if only deleting a DEBUG_INSN.
19676 2013-03-22 Richard Biener <rguenther@suse.de>
19678 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
19680 (mem_ref_alloc): Do not allocate them.
19681 (refs_independent_p): Do not query or maintain a cache.
19683 2013-03-22 Richard Biener <rguenther@suse.de>
19685 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
19686 (gather_mem_refs_in_loops): Do not compute it.
19687 (analyze_memory_references): Do not allocate it.
19688 (tree_ssa_lim_finalize): Do not free it.
19689 (for_all_locs_in_loop): Do not query all_refs_in_loop.
19691 2013-03-22 Richard Biener <rguenther@suse.de>
19693 * is-a.h (as_a): Use gcc_checking_assert.
19695 2013-03-22 Ian Bolton <ian.bolton@arm.com>
19697 * config/aarch64/aarch64.c (aarch64_print_operand): New
19698 format specifier for printing a constant in hex.
19699 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
19700 format specifier for printing second operand.
19702 2013-03-22 Richard Biener <rguenther@suse.de>
19704 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
19706 (gather_mem_refs_in_loops): Perform store accumulation here.
19707 (create_vop_ref_mapping_loop): Remove.
19708 (create_vop_ref_mapping): Likewise.
19709 (analyze_memory_references): Initialize refs_stored_in_loop.
19710 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
19711 (record_indep_loop): Remove.
19712 (record_dep_loop): New function.
19713 (ref_indep_loop_p_1): Adjust to only walk over references
19714 in the loop, not its subloops.
19715 (ref_indep_loop_p): Rename to ...
19716 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
19717 maintaining a more fine-grained cache.
19718 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
19719 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
19721 2013-03-22 Richard Biener <rguenther@suse.de>
19723 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
19724 (struct mem_ref): Make accesses_in_loop a vec of a vec of
19725 aggregate mem_ref_loc.
19726 (free_mem_ref_locs): Inline into ...
19727 (memref_free): ... this and adjust.
19728 (mem_ref_alloc): Adjust.
19729 (mem_ref_locs_alloc): Remove.
19730 (record_mem_ref_loc): Adjust.
19731 (get_all_locs_in_loop): Rewrite into ...
19732 (for_all_locs_in_loop): ... this iterator.
19733 (rewrite_mem_ref_loc): New functor.
19734 (rewrite_mem_refs): Use for_all_locs_in_loop.
19735 (sm_set_flag_if_changed): New functor.
19736 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
19737 (ref_always_accessed): New functor.
19738 (ref_always_accessed_p): Use for_all_locs_in_loop.
19740 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
19742 * tree-pass.h (PROP_gimple_lvec): New.
19743 * passes.c (dump_properties): Handle PROP_gimple_lvec.
19744 (init_optimization_passes): Move pass_lower_vector.
19745 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
19747 (pass_lower_vector): Provide PROP_gimple_lvec.
19748 (pass_lower_vector_ssa): Likewise.
19749 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
19751 2013-03-21 Mark Wielaard <mjw@redhat.com>
19753 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
19755 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
19757 * config/i386/i386.md (*movdi_internal): Disparage slightly
19758 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
19759 conversion alternatives.
19761 2013-03-21 Jakub Jelinek <jakub@redhat.com>
19763 PR middle-end/48087
19764 * diagnostic.def (DK_WERROR): New kind.
19765 * diagnostic.h (werrorcount): Define.
19766 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
19767 promoted to DK_ERROR, increment DK_WERROR counter instead of
19769 * toplev.c (toplev_main): Call print_ignored_options even if
19770 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
19771 even if just werrorcount is non-zero.
19774 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
19776 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
19777 (string_cst_pool_decl): New function.
19778 (optimize_one_addr_into_implicit_ptr): New function.
19779 (resolve_addr_in_expr): Optimize DWARF location expression
19780 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
19781 which doesn't live in memory, but has DW_AT_location or
19782 DW_AT_const_value, or refers to a string literal, into
19783 DW_OP_GNU_implicit_pointer.
19784 (optimize_location_into_implicit_ptr): New function.
19785 (resolve_addr): If removing DW_AT_location of a variable because
19786 it was DW_OP_addr of address of the variable, but the variable doesn't
19787 live in memory, try to emit const value attribute for the initializer.
19789 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
19791 * tree.h (VECTOR_TYPE_P): New macro.
19792 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
19793 TYPE_MODE): Use it.
19794 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
19795 VEC_COND_EXPR cannot be lvalues.
19796 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
19798 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
19800 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
19801 Restrict the transformation to equal modes.
19803 2013-03-21 Richard Biener <rguenther@suse.de>
19805 PR tree-optimization/39326
19806 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
19807 (MEM_ANALYZABLE): Adjust.
19808 (record_mem_ref_loc): Move bitmap ops ...
19809 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
19810 unanalyzable refs, do not record locations for it.
19811 (analyze_memory_references): Allocate ref zero as shared
19813 (refs_independent_p): Do not test for unanalyzed mems here.
19814 (ref_indep_loop_p_1): Special-case disambiguation against
19815 the unanalyzed ref.
19816 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
19818 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
19820 * config/arm/arm-protos.h (tune_params): Add
19821 prefer_neon_for_64bits field.
19822 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
19823 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
19824 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
19825 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
19826 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
19827 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
19828 (arm_option_override): Handle -mneon-for-64bits new option.
19829 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
19830 (prefer_neon_for_64bits): Declare new variable.
19831 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
19832 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
19833 (arch_enabled): Handle new arch types. Remove support for onlya8
19835 (one_cmpldi2): Use new arch names.
19836 (zero_extend<mode>di2, extend<mode>di2): Ditto.
19837 * config/arm/arm.opt (mneon-for-64bits): Add option.
19838 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
19839 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
19840 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
19842 * doc/invoke.texi (-mneon-for-64bits): Document.
19844 2013-03-21 Richard Biener <rguenther@suse.de>
19846 PR tree-optimization/39326
19847 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
19848 (sort_bbs_in_loop_postorder_cmp): New function.
19849 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
19851 2013-03-21 Richard Biener <rguenther@suse.de>
19853 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
19854 (vect_insert_into_interleaving_chain): Likewise.
19855 (vect_drs_dependent_in_basic_block): Inline ...
19856 (vect_slp_analyze_data_ref_dependence): ... here. New function,
19858 (vect_analyze_data_ref_dependence): ... here. Simplify.
19859 (vect_check_interleaving): Simplify.
19860 (vect_analyze_data_ref_dependences): Likewise. Split out ...
19861 (vect_slp_analyze_data_ref_dependences): ... this new function.
19862 (dr_group_sort_cmp): New function.
19863 (vect_analyze_data_ref_accesses): Compute data-reference groups
19864 here instead of in vect_analyze_data_ref_dependence. Use
19865 a more efficient algorithm.
19866 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
19867 vect_slp_analyze_data_ref_dependences. Call
19868 vect_analyze_data_ref_accesses earlier.
19869 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
19870 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
19871 (vect_slp_analyze_data_ref_dependences): New prototype.
19873 2013-03-21 Richard Biener <rguenther@suse.de>
19875 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
19876 ref is stored in the loop.
19877 (find_refs_for_sm): Walk only over all stores.
19878 (store_motion_loop): Allocate from lim_bitmap_obstack.
19879 (store_motion): Likewise.
19881 2013-03-21 Richard Biener <rguenther@suse.de>
19883 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
19884 Update virtual SSA form.
19886 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19888 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
19889 * configure: Regenerate.
19890 * config.in: Regenerate.
19891 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
19892 if !HAVE_LD_EH_FRAME_CIEV3.
19894 2013-03-21 Richard Biener <rguenther@suse.de>
19896 * tree-cfg.c (verify_expr_no_block): New function.
19897 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
19898 nor DECL_VALUE_EXPR have locations with associated blocks.
19899 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
19900 (clear_unused_block_pointer): Remove code dealing with
19901 blocks in DECL_DEBUG_EXPR locations.
19903 2013-03-21 Richard Biener <rguenther@suse.de>
19905 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
19906 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
19907 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
19908 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
19909 instead of DECL_DEBUG_EXPR_IS_FROM.
19910 * gimplify.c (gimplify_modify_expr): Likewise.
19911 * tree-cfg.c (verify_expr_location_1): Likewise.
19912 * tree-complex.c (create_one_component_var): Likewise.
19913 * tree-sra.c (create_access_replacement): Likewise.
19914 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
19915 (clear_unused_block_pointer): Likewise.
19916 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
19917 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
19918 * var-tracking.c (var_debug_decl): Likewise.
19919 (track_expr_p): Likewise.
19920 * tree-inline.c (add_local_variables): Likewise. Set
19921 DECL_HAS_DEBUG_EXPR_P after copying it.
19922 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
19923 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
19925 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
19928 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
19929 * configure: Regenerate.
19930 * config.in: Regenerate.
19931 * config/i386/i386.md (*movdf_internal): Use
19932 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
19933 movd instead of movq mnemonic for interunit moves.
19934 (*movdi_internal): Ditto.
19936 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19938 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
19939 (abd<mode>_3): New pattern.
19940 (aba<mode>_3): New pattern.
19941 (fabd<mode>_3): New pattern.
19943 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19945 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
19946 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
19947 occurrence of REGISTER_PREFIX as its empty string.
19949 2013-03-20 Jeff Law <law@redhat.com>
19951 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
19952 addititional equivalences for equality comparisons between an SSA_NAME
19953 and a constant where the SSA_NAME was set from a widening conversion.
19955 2013-03-20 Walter Lee <walt@tilera.com>
19957 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
19959 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
19961 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
19962 depending on type attribute.
19963 (*movti_internal): Ditto.
19964 (*movtf_internal): Ditto.
19965 (*movxf_internal): Ditto.
19966 (*movdf_internal): Ditto.
19967 (*movsf_internal): Ditto.
19969 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
19971 * config/i386/i386.md (*movti_internal): Set prefix attribute to
19972 maybe_vex for sselog1 and ssemov types.
19973 (*movdi_internal): Reorder operand constraints.
19974 (*movsi_internal): Ditto. Set prefix attribute to
19975 maybe_vex for sselog1 and ssemov types.
19976 (*movtf_internal): Set prefix attribute to maybe_vex
19977 for sselog1 and ssemov types.
19978 (*movdf_internal): Ditto. Set prefix_data16 attribute for
19979 DImode ssemov types. Reorder operand constraints.
19980 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
19981 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
19982 attribute for SImode ssemov types. Reorder operand constraints.
19984 2013-03-20 Martin Jambor <mjambor@suse.cz>
19986 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
19987 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
19989 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
19991 * config/rs6000/predicates.md (indexed_address, update_address_mem
19992 update_indexed_address_mem): New predicates.
19993 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
19994 attribute for load/store instructions.
19995 * config/rs6000/dfp.md (movsd_store): Likewise.
19996 (movsd_load): Likewise.
19997 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
19998 (unnamed HI->DI extend define_insn): Likewise.
19999 (unnamed SI->DI extend define_insn): Likewise.
20000 (unnamed QI->SI extend define_insn): Likewise.
20001 (unnamed QI->HI extend define_insn): Likewise.
20002 (unnamed HI->SI extend define_insn): Likewise.
20003 (unnamed HI->SI extend define_insn): Likewise.
20004 (extendsfdf2_fpr): Likewise.
20005 (movsi_internal1): Likewise.
20006 (movsi_internal1_single): Likewise.
20007 (movhi_internal): Likewise.
20008 (movqi_internal): Likewise.
20009 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
20010 attribute for load/store instructions.
20011 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
20013 (mov<mode>_softfloat): Likewise.
20014 (mov<mode>_hardfloat32): Likewise.
20015 (mov<mode>_hardfloat64): Likewise.
20016 (mov<mode>_softfloat64): Likewise.
20017 (movdi_internal32): Likewise.
20018 (movdi_internal64): Likewise.
20019 (probe_stack_<mode>): Likewise.
20021 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
20023 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
20024 floating point, and decimal floating point to reload iterator.
20026 * config/rs6000/constraints.md (wl constraint): New constraints to
20027 return FLOAT_REGS if certain options are used to reduce the number
20028 of separate patterns that exist in the file.
20029 (wx constraint): Likewise.
20030 (wz constraint): Likewise.
20032 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
20033 -mdebug=reg, print wg, wl, wx, and wz constraints.
20034 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
20035 Initialize the reload functions for 64-bit binary/decimal floating
20037 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
20038 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
20039 create the buffer on the stack to overcome not having a 32-bit
20041 (rs6000_emit_move): Likewise.
20042 (rs6000_secondary_memory_needed_rtx): Likewise.
20043 (rs6000_alloc_sdmode_stack_slot): Likewise.
20044 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
20045 via xxlxor, just like DFmode 0.0.
20047 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
20048 define as 1 if we are running on a power7 or newer.
20049 (enum r6000_reg_class_enum): Add new constraints.
20051 * config/rs6000/dfp.md (movsd): Delete, combine with binary
20052 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
20053 with other moves by using conditional constraits (wg). Use LFIWZX
20054 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
20055 (movsd splitter): Likewise.
20056 (movsd_hardfloat): Likewise.
20057 (movsd_softfloat): Likewise.
20059 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
20060 binary and decimal floating point moves.
20061 (fmove_ok): New attributes to combine binary and decimal floating
20062 point moves, and to combine power6x (mfpgpr) moves along normal
20064 (real_value_to_target): Likewise.
20065 (f32_lr): Likewise.
20066 (f32_lm): Likewise.
20067 (f32_li): Likewise.
20068 (f32_sr): Likewise.
20069 (f32_sm): Likewise.
20070 (f32_si): Likewise.
20071 (movsf): Combine binary and decimal floating point moves. Combine
20072 power6x (mfpgpr) moves with other moves by using conditional
20073 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
20074 (mov<mode> for SFmode/SDmode); Likewise.
20075 (SFmode/SDmode splitters): Likewise.
20076 (movsf_hardfloat): Likewise.
20077 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
20078 (movsf_softfloat): Likewise.
20079 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
20081 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
20082 wx and wz constraints.
20084 * config/rs6000/constraints.md (wg constraint): New constraint to
20085 return FLOAT_REGS if -mmfpgpr (power6x) was used.
20087 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
20090 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
20091 -mdebug=reg, print wg, wl, wx, and wz constraints.
20092 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
20093 Initialize the reload functions for 64-bit binary/decimal floating
20095 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
20096 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
20097 create the buffer on the stack to overcome not having a 32-bit
20099 (rs6000_emit_move): Likewise.
20100 (rs6000_secondary_memory_needed_rtx): Likewise.
20101 (rs6000_alloc_sdmode_stack_slot): Likewise.
20102 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
20103 via xxlxor, just like DFmode 0.0.
20105 * config/rs6000/dfp.md (movdd): Delete, combine with binary
20106 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
20107 with other moves by using conditional constraits (wg). Use LFIWZX
20108 and STFIWX for loading SDmode on power7.
20109 (movdd splitters): Likewise.
20110 (movdd_hardfloat32): Likewise.
20111 (movdd_softfloat32): Likewise.
20112 (movdd_hardfloat64_mfpgpr): Likewise.
20113 (movdd_hardfloat64): Likewise.
20114 (movdd_softfloat64): Likewise.
20116 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
20117 64-bit binary and decimal floating point moves.
20118 (FMOVE64X): Likewise.
20119 (movdf): Combine 64-bit binary and decimal floating point moves.
20120 Combine power6x (mfpgpr) moves with other moves by using
20121 conditional constraits (wg).
20122 (mov<mode> for DFmode/DDmode): Likewise.
20123 (DFmode/DDmode splitters): Likewise.
20124 (movdf_hardfloat32): Likewise.
20125 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
20126 (movdf_softfloat32): Likewise.
20127 (movdf_hardfloat64_mfpgpr): Likewise.
20128 (movdf_hardfloat64): Likewise.
20129 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
20130 (movdf_softfloat64): Likewise.
20131 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
20132 (reload_<mode>_load): Move to later in the file so they aren't in
20133 the middle of the floating point move insns.
20134 (reload_<mode>_store): Likewise.
20136 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
20139 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
20140 constraint if -mdebug=reg.
20141 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
20142 Enable using dd reload support if needed.
20144 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
20145 binary and decimal floating point moves in rs6000.md.
20146 (movtd_internal): Likewise.
20148 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
20149 decimal floating point moves.
20151 (movtf_internal): Likewise.
20152 (mov<mode>_internal, TDmode/TFmode): Likewise.
20153 (movtf_softfloat): Likewise.
20154 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
20156 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
20157 movdi_internal64, using wg constraint for move direct operations.
20158 (movdi_internal64): Likewise.
20160 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
20161 MODES_TIEABLE_P for selected modes. Print the numerical value of
20162 the various virtual registers. Use GPR/FPR first/last values,
20163 instead of hard coding the register numbers. Print which modes
20164 have reload functions registered.
20165 (rs6000_option_override_internal): If -mdebug=reg, trace the options
20166 settings before/after setting cpu, target and subtarget settings.
20167 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
20168 and for secondary reload failures in rs6000_secondary_reload_inner.
20169 (rs6000_secondary_reload_fail): Likewise.
20170 (rs6000_secondary_reload_inner): Likewise.
20172 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
20173 macros for first/last GPR and FPR registers.
20174 (LAST_GPR_REGNO): Likewise.
20175 (FIRST_FPR_REGNO): Likewise.
20176 (LAST_FPR_REGNO): Likewise.
20178 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
20179 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
20180 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
20181 (vcond<mode><mode>): Likewise.
20182 (vcondu<mode><mode>): Likewise.
20183 (vector_gtu<mode>): Likewise.
20184 (vector_gte<mode>): Likewise.
20185 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
20186 to prevent the compiler from converting DImode operations to TImode.
20187 (ior<mode>3): Likewise.
20188 (and<mode>3): Likewise.
20189 (one_cmpl<mode>2): Likewise.
20190 (nor<mode>3): Likewise.
20191 (andc<mode>3): Likewise.
20193 * config/rs6000/constraints.md (wt constraint): New constraint
20194 that returns VSX_REGS if TImode is allowed in VSX registers.
20196 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
20197 constant under VSX.
20199 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
20200 similar to TImode, but it is restricted to being in the GPRs.
20202 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
20203 TImode to occupy a single VSX register.
20205 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
20206 -mvsx-timode for power7/power8.
20207 (power7 cpu): Likewise.
20208 (power8 cpu): Likewise.
20210 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
20211 sure that TFmode/TDmode take up two registers if they are ever
20212 allowed in the upper VSX registers.
20213 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
20215 (rs6000_init_hard_regno_mode_ok): Likewise.
20216 (rs6000_debug_reg_global): Add debugging for PTImode and wt
20217 constraint. Print if LRA is turned on.
20218 (rs6000_option_override_internal): Give an error if -mvsx-timode
20219 and VSX is not enabled.
20220 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
20221 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
20222 to reg+offset addressing. Use PTImode when checking offset
20223 addresses for validity.
20224 (reg_offset_addressing_ok_p): Likewise.
20225 (rs6000_legitimate_offset_address_p): Likewise.
20226 (rs6000_legitimize_address): Likewise.
20227 (rs6000_legitimize_reload_address): Likewise.
20228 (rs6000_legitimate_address_p): Likewise.
20229 (rs6000_eliminate_indexed_memrefs): Likewise.
20230 (rs6000_emit_move): Likewise.
20231 (rs6000_secondary_reload): Likewise.
20232 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
20233 reloads to fpr registers to continue to use reg+offset addressing,
20234 but 64-bit reloads to altivec registers need reg+reg addressing.
20235 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
20236 it. Treat LO_SUM like a PLUS operation.
20237 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
20238 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
20239 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
20240 registers to share a register with a smaller sized type, since VSX
20241 puts scalars in the upper 64-bits.
20242 (print_operand): Add support for PTImode.
20243 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
20244 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
20245 registers, but don't have arithmetic support.
20246 (rs6000_memory_move_cost): Add test for VSX.
20247 (rs6000_opt_masks): Add -mvsx-timode.
20249 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
20252 (VSr): Use wt constraint for TImode.
20253 (VSv): Drop TImode support.
20254 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
20255 (vsx_movti_64bit): Likewise.
20256 (vsx_movti_32bit): Likewise.
20257 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
20258 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
20259 one '?' on the appropriate output constraint. Do not allow TImode
20260 logical operations on 32-bit systems.
20261 (vsx_ior<mode>3): Likewise.
20262 (vsx_xor<mode>3): Likewise.
20263 (vsx_one_cmpl<mode>2): Likewise.
20264 (vsx_nor<mode>3): Likewise.
20265 (vsx_andc<mode>3): Likewise.
20266 (vsx_concat_<mode>): Likewise.
20267 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
20269 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
20270 OPTION_MASK_VSX_TIMODE.
20271 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
20272 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
20274 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
20275 (TI2 iterator): New iterator for TImode, PTImode.
20276 (wd mode attribute): Add values for vector types.
20277 (movti_string): Replace TI move operations with operations for TImode
20278 and PTImode. Add support for TImode being allowed in VSX registers.
20279 (mov<mode>_string, TImode/PTImode): Likewise.
20280 (movti_ppc64): Likewise.
20281 (mov<mode>_ppc64, TImode/PTImode): Likewise.
20282 (TI mode splitters): Likewise.
20284 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
20287 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
20289 PR tree-optimization/56355
20290 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
20291 Also handle integers with undefined overflow.
20293 2013-03-20 Catherine Moore <clm@codesourcery.com>
20294 Maciej W. Rozycki <macro@codesourcery.com>
20295 Tom de Vries <tom@codesourcery.com>
20296 Nathan Sidwell <nathan@codesourcery.com>
20297 Iain Sandoe <iain@codesourcery.com>
20298 Nathan Froyd <froydnj@codesourcery.com>
20299 Chao-ying Fu <fu@mips.com>
20301 * doc/extend.texi: (micromips, nomicromips, nocompression):
20302 Document new function attributes.
20303 * doc/invoke.texi (minterlink-compressed, mmicromips,
20304 m14k, m14ke, m14kec): Document new options.
20305 (minterlink-mips16): Update documentation.
20306 * doc/md.texi (ZC, ZD): Document new constraints.
20307 * configure.ac (gcc_cv_as_micromips): Check if linker
20308 supports the .set micromips directive.
20309 * configure: Regenerate.
20310 * config.in: Regenerate.
20311 * config/mips/mips-tables.opt: Regenerate.
20312 * config/mips/micromips.md: New file.
20313 * constraints.md (ZC, ZD): New constraints.
20314 * config/mips/predicates.md (movep_src_register): New predicate.
20315 (movep_src_operand): New predicate.
20316 (non_volatile_mem_operand): New predicate.
20317 * config/mips/mips.md (multimem): New type.
20318 (length): Differentiate between 17-bit and 18-bit branch offsets.
20319 (MOVEP1, MOVEP2): New mode iterator.
20320 (mov_<load>l): Use ZC constraint.
20321 (mov_<load>r): Likewise.
20322 (mov_<store>l): Likewise.
20323 (mov_<store>r): Likewise.
20324 (*branch_equality<mode>_inverted): Add microMIPS support.
20325 (*branch_equality<mode>): Likewise.
20326 (*jump_absolute): Likewise.
20327 (indirect_jump_<mode>): Likewise.
20328 (tablejump_<mode>): Likewise.
20329 (<optab>_internal): Likewise.
20330 (sibcall_internal): Likewise.
20331 (sibcall_value_internal): Likewise.
20332 (prefetch): Use constraint ZD.
20333 * config/mips/mips.opt (minterlink-compressed): New option.
20334 (minterlink-mips16): Now an alias for minterlink-compressed.
20335 (mmicromips): New option.
20336 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
20337 (compare_and_swap_12): Likewise.
20338 (sync_add<mode>): Likewise.
20339 (sync_<optab>_12): Likewise.
20340 (sync_old_<optab>_12): Likewise.
20341 (sync_new_<optab>_12): Likewise.
20342 (sync_nand_12): Likewise.
20343 (sync_old_nand_12): Likewise.
20344 (sync_new_nand_12): Likewise.
20345 (sync_sub<mode>): Likewise.
20346 (sync_old_add<mode>): Likewise.
20347 (sync_old_sub<mode>): Likewise.
20348 (sync_new_add<mode>): Likewise.
20349 (sync_new_sub<mode>): Likewise.
20350 (sync_<optab><mode>): Likewise.
20351 (sync_old_<optab><mode>): Likewise.
20352 (sync_new_<optab><mode>): Likewise.
20353 (sync_nand<mode>): Likewise.
20354 (sync_old_nand<mode>): Likewise.
20355 (sync_new_nand<mode>): Likewise.
20356 (sync_lock_test_and_set<mode>): Likewise.
20357 (test_and_set_12): Likewise.
20358 (atomic_compare_and_swap<mode>): Likewise.
20359 (atomic_exchange<mode>_llsc): Likewise.
20360 (atomic_fetch_add<mode>_llsc): Likewise.
20361 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
20362 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
20363 (umips_save_restore_pattern_p): Likewise.
20364 (umips_load_store_pair_p): Likewise.
20365 (umips_output_load_store_pair): Likewise.
20366 (umips_movep_target_p): Likewise.
20367 (umips_12bit_offset_address_p): Likewise.
20368 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
20369 (mips_base_mips16): Rename this...
20370 (mips_base_compression_flags): ...to this. Update all uses.
20371 (mips_attribute_table): Add micromips, nomicromips and nocompression.
20372 (mips_mips16_decl_p): Delete.
20373 (mips_nomips16_decl_p): Delete.
20374 (mips_get_compress_on_flags): New function.
20375 (mips_get_compress_off_flags): New function.
20376 (mips_get_compress_mode): New function.
20377 (mips_get_compress_on_name): New function.
20378 (mips_get_compress_off_name): New function.
20379 (mips_insert_attributes): Support multiple compression types.
20380 (mips_merge_decl_attributes): Likewise.
20381 (umips_12bit_offset_address_p): New function.
20382 (mips_start_function_definition): Emit .set micromips directive.
20383 (mips_call_may_need_jalx_p): New function.
20384 (mips_function_ok_for_sibcall): Add microMIPS support.
20385 (mips_print_operand_punctuation): Support short delay slots and
20387 (umips_swm_mask, umips_swm_encoding): New.
20388 (umips_build_save_restore): New function.
20389 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
20390 (was_mips16_p): Remove.
20391 (old_compression_mode): New.
20392 (mips_set_compression_mode): New function.
20393 (mips_set_current_function): Add microMIPS support.
20394 (mips_option_override): Likewise.
20395 (umips_save_restore_pattern_p): New function.
20396 (umips_output_save_restore): New function.
20397 (umips_load_store_pair_p_1): New function.
20398 (umips_load_store_pair_p): New function.
20399 (umips_output_load_store_pair_1): New function.
20400 (umips_output_load_store_pair): New function.
20401 (umips_movep_target_p) New function.
20402 (mips_prepare_pch_save): Add microMIPS support.
20403 * config/mips/mips.h (TARGET_COMPRESSION): New.
20404 (TARGET_CPU_CPP_BUILTINS): Update macro
20405 to use new compression flags and to support microMIPS.
20406 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
20407 (MIPS_ARCH_FLOAT_SPEC): Likewise.
20408 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
20409 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
20410 (ASM_SPEC): Support mmicromips and mno-micromips.
20411 (M16STORE_REG_P): New macro.
20412 (MIPS_CALL): Support TARGET_MICROMIPS.
20413 (MICROMIPS_J): New macro.
20414 (mips_base_mips16): Rename this...
20415 (mips_base_compression_flags): ...to this.
20416 (UMIPS_12BIT_OFFSET_P): New macro.
20417 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
20418 (MULTILIB_DIRNAMES): Likewise.
20419 2013-03-20 Richard Biener <rguenther@suse.de>
20421 PR tree-optimization/56661
20422 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
20423 the result does not have to be distinct.
20425 2013-03-20 Richard Biener <rguenther@suse.de>
20427 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
20430 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20431 Steven Bosscher <steven@gcc.gnu.org>
20433 PR rtl-optimization/56605
20434 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
20436 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
20439 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
20440 that require movd instead of movq.
20442 2013-03-20 Richard Biener <rguenther@suse.de>
20444 * tree-ssa-structalias.c (struct variable_info): Add pointer
20445 to the first field of an aggregate with sub-vars. Make
20446 this and the pointer to the next subfield its ID.
20447 (vi_next): New function.
20448 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
20449 storedanything_id, integer_id): Increment by one.
20450 (new_var_info, get_call_vi, lookup_call_clobber_vi,
20451 get_call_clobber_vi): Adjust.
20452 (solution_set_expand): Simplify and speedup.
20453 (solution_set_add): Inline into ...
20454 (set_union_with_increment): ... this. Adjust accordingly.
20455 (do_sd_constraint): Likewise.
20456 (do_ds_constraint): Likewise.
20457 (do_complex_constraint): Simplify.
20458 (build_pred_graph): Adjust.
20459 (solve_graph): Likewise. Simplify and speedup.
20460 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
20461 get_constraint_for_component_ref, get_constraint_for_1,
20462 first_vi_for_offset, first_or_preceding_vi_for_offset,
20463 create_function_info_for, create_variable_info_for_1,
20464 create_variable_info_for, intra_create_variable_infos): Adjust.
20465 (init_base_vars): Push NULL for ID zero.
20466 (compute_points_to_sets): Adjust.
20468 2013-03-20 Richard Biener <rguenther@suse.de>
20470 * cfgloop.c (verify_loop_structure): Streamline and avoid
20471 ICEing on corrupt loop tree.
20472 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
20475 2013-03-20 Richard Biener <rguenther@suse.de>
20477 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
20478 check whether an SSA update is needed.
20480 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
20482 * config/mips/constraints.md (T): Rename to...
20486 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
20487 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
20489 2013-03-19 Ian Bolton <ian.bolton@arm.com>
20491 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
20492 (*subsi3_carryin_uxtw): Likewise.
20494 2013-03-19 Ian Bolton <ian.bolton@arm.com>
20496 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
20497 (*rorsi3_insn_uxtw): Likewise.
20499 2013-03-19 Ian Bolton <ian.bolton@arm.com>
20501 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
20502 (*extrsi5_insn_uxtw): Likewise.
20504 2013-03-19 Richard Biener <rguenther@suse.de>
20506 PR tree-optimization/56273
20507 * passes.c (init_optimization_passes): Move second VRP after DOM.
20509 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
20511 * config/i386/i386.md (*movti_internal): Merge from
20512 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
20513 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
20514 nox64 isa attributes.
20516 2013-03-18 Richard Biener <rguenther@suse.de>
20518 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
20520 (merge_node_constraints): Likewise.
20521 (build_succ_graph): Likewise.
20522 (valid_graph_edge): Inline into single caller.
20523 (unify_nodes): Likewise. Use bitmap_set_bit return value
20525 (scc_visit): Fix formatting and variable use.
20526 (do_sd_constraint): Use gcc_checking_assert.
20527 (do_ds_constraint): Likewise.
20528 (do_complex_constraint): Likewise.
20529 (condense_visit): Likewise. Cleanup.
20530 (dump_pred_graph): New function.
20531 (perform_var_substitution): Dump the pred-graph before
20532 variable substitution.
20533 (find_equivalent_node): Use gcc_checking_assert.
20534 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
20536 2013-03-18 Richard Biener <rguenther@suse.de>
20538 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
20539 Remove cond_expr_stmt_list argument and do not gimplify the
20541 (vect_loop_versioning): Adjust.
20542 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
20543 Cleanup to use less temporaries.
20544 (vect_create_data_ref_ptr): Cleanup.
20546 2013-03-18 Jakub Jelinek <jakub@redhat.com>
20548 PR tree-optimization/56635
20549 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
20550 require types_compatible_p types.
20552 2013-03-18 Nick Clifton <nickc@redhat.com>
20554 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
20555 spurious backslash.
20557 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
20558 Add missing line to comment describing function.
20560 2013-03-18 Richard Biener <rguenther@suse.de>
20562 PR tree-optimization/56210
20563 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
20564 Handle string / character search functions.
20565 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
20567 2013-03-18 Richard Biener <rguenther@suse.de>
20569 PR middle-end/56483
20570 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
20571 and implement properly.
20572 * gimple.h (gimple_cond_single_var_p): Remove.
20574 2013-03-18 Richard Biener <rguenther@suse.de>
20576 * tree-data-ref.h (find_data_references_in_loop): Declare.
20577 * tree-data-ref.c (get_references_in_stmt): Use a stack
20578 vector pre-allocated in the callers.
20579 (find_data_references_in_stmt): Adjust.
20580 (graphite_find_data_references_in_stmt): Likewise.
20581 (create_rdg_vertices): Likewise.
20582 (find_data_references_in_loop): Export.
20583 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
20584 Compute dependences here...
20585 (vect_analyze_data_refs): ...not here. When we encounter
20586 a non-vectorizable data reference in basic-block vectorization
20587 truncate the data reference vector. Do not bother to
20588 fixup data-dependence information for gather loads.
20589 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
20590 of data references, as reported.
20592 2013-03-18 Richard Biener <rguenther@suse.de>
20594 PR tree-optimization/3713
20595 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
20596 has_constants and expr.
20597 (stmt_has_constants): Properly valueize SSA names when deciding
20598 whether the stmt has constants.
20600 2013-03-18 Richard Biener <rguenther@suse.de>
20602 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
20603 whole function when there is nothing to do.
20604 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
20605 * tree-vectorizer.c (vectorize_loops): Update virtual and
20606 loop-closed SSA once.
20607 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
20609 2013-03-18 Richard Biener <rguenther@suse.de>
20611 PR middle-end/56113
20612 * domwalk.c (bb_postorder): New global static.
20613 (cmp_bb_postorder): New function.
20614 (walk_dominator_tree): Replace scheme imposing an order for
20615 visiting dominator sons by one sorting them at the time they
20616 are pushed on the stack.
20618 2013-03-18 Richard Biener <rguenther@suse.de>
20620 PR tree-optimization/39326
20621 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
20622 (struct mem_ref): Replace mem member with ao_ref typed member.
20623 (MEM_ANALYZABLE): Adjust.
20624 (memref_eq): Likewise.
20625 (mem_ref_alloc): Likewise.
20626 (gather_mem_refs_stmt): Likewise.
20627 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
20628 (execute_sm_if_changed_flag_set): Adjust.
20629 (execute_sm): Likewise.
20630 (ref_always_accessed_p): Likewise.
20631 (refs_independent_p): Likewise.
20632 (can_sm_ref_p): Likewise.
20634 2013-03-18 Jakub Jelinek <jakub@redhat.com>
20637 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
20638 return 1 even for !unsignedp.
20640 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
20642 * config/i386/i386.md (isa): Add x64 and nox64.
20643 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
20644 (*pushtf): Enable *roF alternative for x64 isa only.
20645 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
20646 mode attribute of integer alternatives to DImode for TARGET_64BIT.
20647 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
20648 (*movtf_internal): Merge from *movtf_internal_rex64 and
20649 *movtf_internal_sse. Use x64 and nox64 isa attributes.
20650 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
20651 nox64 isa attributes.
20652 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
20653 nox64 isa attributes.
20654 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
20656 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
20658 * config/alpha/alpha.c (TARGET_LRA_P): New define.
20660 2013-03-17 Jakub Jelinek <jakub@redhat.com>
20663 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
20664 class names. Remove trailing comma after "ALL_REGS".
20666 2013-03-16 Jan Hubicka <jh@suse.cz>
20668 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
20669 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
20670 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
20671 of cgraph_get_create_node.
20672 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
20674 2013-03-16 Jason Merrill <jason@redhat.com>
20677 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
20678 with DW_AT_default_value.
20680 2013-03-16 Jakub Jelinek <jakub@redhat.com>
20682 * BASE-VER: Set to 4.9.0.
20684 2013-03-14 Andi Kleen <ak@linux.intel.com>
20687 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
20688 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
20689 Document _x* TSX intrinsics.
20691 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
20692 David Holsgrove <david.holsgrove@xilinx.com>
20694 * configure.ac: Add MicroBlaze TLS support detection.
20695 * configure: Regenerate.
20696 * config/microblaze/microblaze-protos.h
20697 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
20698 symbol_mentioned_p, label_mentioned_p): Add prototypes.
20699 * config/microblaze/microblaze.c (microblaze_address_type): Add
20700 ADDRESS_TLS and tls_reloc address types.
20701 (microblaze_address_info): Add tls_reloc.
20702 (TARGET_HAVE_TLS): Define.
20703 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
20704 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
20705 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
20706 load_tls_operand, microblaze_call_tls_get_addr,
20707 microblaze_legitimize_tls_address): New functions.
20708 (microblaze_classify_unspec): Handle UNSPEC_TLS.
20709 (get_base_reg): Use microblaze_tls_symbol_p.
20710 (microblaze_classify_address): Handle TLS.
20711 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
20712 label_mentioned_p and microblaze_tls_referenced_p.
20713 (microblaze_legitimize_address): Handle TLS.
20714 (microblaze_address_insns): Handle ADDRESS_TLS.
20715 (pic_address_needs_scratch): Handle TLS.
20716 (print_operand_address): Handle TLS.
20717 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
20718 (microblaze_expand_move): Handle TLS.
20719 (microblaze_legitimate_constant_p): Check
20720 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
20721 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
20722 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
20723 (PIC_OFFSET_TABLE_REGNUM): Set.
20724 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
20725 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
20726 (addsi3, movsi_internal2, movdf_internal): Update constraints
20727 * config/microblaze/predicates.md (arith_plus_operand): Define
20728 (move_operand): Redefine as move_src_operand,
20729 check microblaze_tls_referenced_p.
20731 2013-03-14 Ian Bolton <ian.bolton@arm.com>
20733 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
20734 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
20736 2013-03-14 Ian Bolton <ian.bolton@arm.com>
20738 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
20741 2013-03-14 Jakub Jelinek <jakub@redhat.com>
20743 PR tree-optimization/53265
20744 * common.opt (Waggressive-loop-optimizations): New option.
20745 * tree-ssa-loop-niter.c: Include tree-pass.h.
20746 (do_warn_aggressive_loop_optimizations): New function.
20747 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
20748 if number_of_latch_executions returned constant.
20749 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
20750 early. If number_of_latch_executions returned constant, set
20751 nb_iterations_upper_bound back to it.
20752 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
20754 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
20755 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
20757 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
20758 (MULTILIB_OSDIRNAMES): Set.
20759 * genmultilib: If defaultosdirname doesn't start with :: , set
20760 defaultosdirname2 instead, clear it and emit two . multilib_raw
20761 entries instead of just one.
20763 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
20765 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
20766 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
20767 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
20768 (SUBTARGET_OVERRIDE_OPTIONS): New.
20770 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
20773 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
20774 (musermode): Convert to Var(TARGET_USERMODE).
20775 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
20776 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
20777 * config/sh/sh.c (sh_option_override): Use
20778 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
20779 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
20781 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
20783 (udivsi3_i4_single, divsi3_i4_single): Use
20784 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
20786 2013-03-13 Dave Korn <dave.korn.cygwin@....>
20788 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
20791 2013-03-13 Richard Biener <rguenther@suse.de>
20793 PR tree-optimization/56608
20794 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
20795 calls when vectorizing basic-blocks.
20797 2013-03-13 Jakub Jelinek <jakub@redhat.com>
20800 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
20803 2013-03-12 Jakub Jelinek <jakub@redhat.com>
20805 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
20807 2013-03-11 Jan Hubicka <jh@suse.cz>
20810 * lto-streamer-out.c (output_symbol_p): Skip references from
20811 constructors of external variables.
20813 2013-03-11 Jan Hubicka <jh@suse.cz>
20815 PR middle-end/56571
20816 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
20818 * emit-rtl.c (verify_rtx_sharing): Likewise.
20819 (copy_insn_1): Likewise.
20820 * rtl.c (copy_rtx): Likewise.
20822 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
20825 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
20826 output_operand_lossage message.
20828 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
20831 * arm.c (shift_op): Validate RTL pattern on the fly.
20832 (arm_print_operand, case 'S'): Don't use shift_operator to validate
20835 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20838 * config/pa/pa.md (call_value): Check for calls to powf and direct to
20839 new call patterns that clobber %fr12.
20840 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
20841 split and postreload patterns.
20842 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
20843 registers %fr12 and %fr12R as call used.
20845 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
20847 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
20848 (canon_address, record_store, replace_read, check_mem_read_rtx,
20849 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
20850 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
20851 rest_of_handle_dse): Likewise.
20853 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
20855 PR middle-end/56524
20856 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
20858 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
20859 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
20860 (save_optabs_if_changed): Replace with...
20861 (init_tree_optimization_optabs): ...this.
20862 * optabs.c (save_optabs_if_changed): Rename to...
20863 (init_tree_optimization_optabs): ...this. Take the optimization node
20864 as argument. Do nothing if the base optabs are already correct.
20865 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
20866 to recompute optabs.
20867 * function.h (function): Remove optabs field.
20868 * function.c (invoke_set_current_function_hook): Call
20869 init_tree_optimization_optabs. Use the result to initialize
20872 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
20874 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
20875 if GTMA_HAS_NO_INSTRUMENTATION.
20876 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
20877 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
20878 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
20879 * gimple-pretty-print.c (dump_gimple_transaction): Handle
20880 GTMA_HAS_NO_INSTRUMENTATION.
20882 2013-03-08 Jakub Jelinek <jakub@redhat.com>
20884 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
20887 2013-03-08 Marek Polacek <polacek@redhat.com>
20888 Jakub Jelinek <jakub@redhat.com>
20890 PR tree-optimization/56478
20891 * predict.c (is_comparison_with_loop_invariant_p): Change the
20892 type of loop_step to tree.
20893 (predict_loops): Adjust.
20894 (predict_iv_comparison): Perform the computations on double_ints.
20896 2013-03-08 Richard Biener <rguenther@suse.de>
20898 PR tree-optimization/56570
20899 * tree-cfg.c (verify_expr_location_1): Verify locations for
20901 * tree-sra.c (create_access_replacement): Strip locations
20902 from DECL_DEBUG_EXPRs.
20904 2013-03-08 Richard Biener <rguenther@suse.de>
20906 * tree-inline.c (expand_call_inline): Do not associate
20907 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
20908 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
20910 2013-03-08 Richard Biener <rguenther@suse.de>
20912 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
20913 or block changes with -Og. Fix for location / block encoding
20914 changes and PHI arguments with locations.
20916 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
20918 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
20920 (struct output_info): Likewise.
20921 (register_overhead): Remove bad gcc_assert.
20922 (bitmap_find_bit): If there is only a single bitmap element, do not
20923 count a miss as a search.
20924 (print_statistics): Update for counter type changes.
20925 (dump_bitmap_statistics): Likewise. Print headers such that they
20926 are properly lined up with the printed counters.
20928 2013-03-07 Jakub Jelinek <jakub@redhat.com>
20930 PR tree-optimization/56559
20931 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
20932 check that it has only a single use.
20934 2013-03-07 Richard Biener <rguenther@suse.de>
20936 * doc/invoke.texi (fwhole-program): Discourage use in combination
20939 2013-03-06 Jakub Jelinek <jakub@redhat.com>
20941 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
20943 PR tree-optimization/56539
20944 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
20945 instead of GSI_CONTINUE_LINKING as last argument to
20946 force_gimple_operand_gsi. Adjust function comment.
20948 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
20951 PR middle-end/56548
20952 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
20953 promoted mode, convert the result back to the original mode.
20955 2013-03-06 Richard Biener <rguenther@suse.de>
20957 PR middle-end/56294
20958 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
20959 (insert_updated_phi_nodes_compare_uids): New function.
20960 (update_ssa): Sort symbols_to_rename after UID before
20961 traversing it to insert PHI nodes.
20963 2013-03-06 Richard Biener <rguenther@suse.de>
20965 PR middle-end/50494
20966 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
20967 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
20970 2013-02-13 Richard Biener <rguenther@suse.de>
20973 * varasm.c (output_constant_def_1): Get the decl representing
20974 the constant as argument.
20975 (output_constant_def): Wrap output_constant_def_1.
20976 (make_decl_rtl): Use output_constant_def_1 with the decl
20977 representing the constant.
20978 (build_constant_desc): Optionally re-use a decl already
20979 representing the constant.
20980 (tree_output_constant_def): Adjust.
20982 2013-03-06 Joey Ye <joey.ye@arm.com>
20985 * gcc.c (convert_white_space): New function.
20986 (main): Handles white space in function name.
20988 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
20991 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
20992 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
20993 to SH_DIV_CALL_TABLE for TARGET_SH2.
20994 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
20996 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
20997 call-table options.
20999 2013-03-05 Sterling Augustine <saugustine@google.com>
21000 Cary Coutant <ccoutant@google.com>
21003 * dwarf2out.c (resolve_addr): Don't call
21004 remove_loc_list_addr_table_entries a second time for the same
21007 2013-03-05 Jakub Jelinek <jakub@redhat.com>
21010 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
21011 (avoid_complex_debug_insns): New function.
21012 (expand_debug_locations): Call it.
21014 PR rtl-optimization/56484
21015 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
21016 lifetimes of hard registers on small register class machines.
21018 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
21020 * config/microblaze/microblaze-protos.h: Rename
21021 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
21022 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
21024 (microblaze_fast_interrupt_function_p): New function.
21025 (microblaze_is_interrupt_handler): Rename to
21026 microblaze_is_interrupt_variant and add fast_interrupt check.
21027 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
21028 (save_restore_insns): Likewise.
21029 (compute_frame_size): Likewise.
21030 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
21031 (microblaze_globalize_label): Likewise.
21032 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
21033 * config/microblaze/microblaze.md: Use wrapper
21034 microblaze_is_interrupt_variant.
21036 2013-03-05 Kai Tietz <ktietz@redhat.com>
21038 * sdbout.c (sdbout_one_type): Switch to current function's section
21039 supporting cold/hot.
21041 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
21043 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
21046 2013-03-05 Jakub Jelinek <jakub@redhat.com>
21048 PR middle-end/56461
21049 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
21050 if VALGRIND_GET_VBITS is defined, temporarily make object
21051 memory all defined, and restore previous valgrind addressability
21052 and definability afterwards. Free this_object at the end.
21054 PR middle-end/56461
21055 * lra.c (lra): Call lra_clear_live_ranges if live_p,
21056 right before calling lra_create_live_ranges, also call it
21057 when clearing live_p. Only call lra_clear_live_ranges
21058 at the end if live_p.
21060 PR middle-end/56461
21061 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
21063 2013-03-05 Richard Biener <rguenther@suse.de>
21065 PR tree-optimization/56521
21066 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
21069 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
21072 * except.h (remove_unreachable_eh_regions): New prototype.
21073 * except.c (remove_eh_handler_splicer): New function, split out
21074 of remove_eh_handler.
21075 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
21076 warning about running it on many EH regions one at a time.
21077 (remove_unreachable_eh_regions_worker): New function, walk the
21078 EH tree in depth-first order and remove non-marked regions.
21079 (remove_unreachable_eh_regions): New function.
21080 * tree-eh.c (mark_reachable_handlers): New function, split out
21081 from remove_unreachable_handlers.
21082 (remove_unreachable_handlers): Use mark_reachable_handlers and
21083 remove_unreachable_eh_regions.
21084 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
21085 and remove_unreachable_eh_regions.
21087 2013-03-05 Richard Biener <rguenther@suse.de>
21089 PR middle-end/56525
21090 * loop-init.c (fix_loop_structure): Remove loops in two stages,
21091 not freeing them until the end.
21093 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21095 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
21097 2013-03-05 Richard Biener <rguenther@suse.de>
21099 PR tree-optimization/56270
21100 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
21101 of loads after scheduling an SLP instance.
21103 2013-03-05 Jakub Jelinek <jakub@redhat.com>
21105 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
21107 (check_gcc_parallelize): Run guality.exp as a separate job from
21108 vect.exp with unsorted.exp and $(dg_target_exps) separately from
21109 struct-layout-1.exp with stackalign.exp.
21111 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
21113 PR middle-end/56461
21114 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
21115 load_index sbitmap even if some bit in it isn't set.
21117 PR middle-end/56461
21118 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
21119 (discover_iteration_bound_by_body_walk): Change queues to
21120 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
21121 spelling in comment. Call safe_push on queues[bound_index] directly.
21122 Release queues[queue_index] in every iteration unconditionally.
21123 Release bounds vector.
21125 PR middle-end/56461
21126 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
21127 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
21128 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
21131 2013-03-05 Richard Biener <rguenther@suse.de>
21134 * tree-inline.c (remap_blocks_to_null): New function.
21135 (expand_call_inline): When expanding a call stmt without
21136 an associated block inline remap all callee blocks to NULL.
21138 2013-03-05 Jakub Jelinek <jakub@redhat.com>
21140 PR rtl-optimization/56494
21141 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
21142 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
21143 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
21145 PR middle-end/56461
21146 * sel-sched-ir.c (free_sched_pools): Release
21147 succs_info_pool.stack[succs_info_pool.max_top] vectors too
21148 if succs_info_pool.max_top isn't -1.
21151 * opts.c (opts_obstack, opts_concat): Moved to...
21152 * opts-common.c (opts_obstack, opts_concat): ... here.
21154 2013-03-04 Jakub Jelinek <jakub@redhat.com>
21156 PR middle-end/56461
21157 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
21159 2013-03-04 Martin Jambor <mjambor@suse.cz>
21161 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
21162 all appropriate places.
21164 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
21166 PR tree-optimization/56424
21167 * ipa-split.c (split_function): Do not set the RSO flag if result is
21168 not by reference and its type is a register type.
21170 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
21172 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
21173 (microblaze_legitimate_pic_operand): Likewise
21174 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
21175 new function microblaze_legitimate_pic_operand
21176 * config/microblaze/microblaze-protos.h
21177 (microblaze_legitimate_pic_operand): Declare.
21179 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21181 * config/microblaze/predicates.md (call_insn_simple_operand):
21182 New predicate for supported rtx code types.
21183 * config/microblaze/microblaze.md (call_internal1): Use
21184 call_insn_simple_operand predicate.
21186 2013-03-04 Jakub Jelinek <jakub@redhat.com>
21188 PR middle-end/56461
21189 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
21190 partitions.ordered_remove.
21192 PR middle-end/56461
21193 * tree-vect-stmts.c (vectorizable_conversion): Don't call
21194 vec_oprnds0.create (1) for modifier == NONE.
21196 PR middle-end/56461
21197 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
21198 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
21199 vec_oprnds1 right before pushing anything to it for
21202 PR middle-end/56461
21203 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
21204 set nbbs to 0 instead of having separate code path.
21205 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
21206 instead of false as last argument if returning NULL.
21208 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
21210 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
21211 the attribute is now called "target" instead of "option".
21212 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
21213 * doc/tm.texi.in (Target Attributes): Likewise document the correct
21214 attribute/pragma name for TARGET_OPTION_VALID_P and
21215 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
21216 * doc/tm.texi: Regenerated.
21218 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
21220 * config/microblaze/microblaze.c:
21221 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
21222 * config/microblaze/microblaze.h: Add -mxl-reorder to
21224 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
21225 instructions emitted if TARGET_REORDER.
21226 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
21227 or 0 for -m/-mno case, but initialises as 2 to detect default use case
21230 2013-03-01 Xinliang David Li <davidxl@google.com>
21232 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
21235 2013-03-01 Jakub Jelinek <jakub@redhat.com>
21237 PR middle-end/56461
21238 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
21239 vector even when returning true. Fix up function comment formatting.
21241 PR middle-end/56461
21242 * ira-build.c (ira_loop_nodes_count): New variable.
21243 (create_loop_tree_nodes): Initialize it.
21244 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
21246 PR middle-end/56461
21247 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
21248 method on dr_chain and result_chain.
21249 * tree-vect-stmts.c (vectorizable_store): Only call
21250 result_chain.create if j == 0.
21252 PR middle-end/56461
21253 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
21254 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
21255 before overwriting it.
21257 2013-03-01 Tobias Burnus <burnus@net-b.de>
21259 * doc/extended.texi (C Extensions): Change order in @menu
21261 (Other MIPS Built-in Functions): Move last MIPS entry before
21262 "picoChip Built-in Functions".
21263 (SH Built-in Functions): Move after RX Built-in Functions.
21264 * doc/gcc.texi (Introduction): Change order in @menu
21266 * doc/md.texi (Constraints): Ditto.
21267 * gty.texi (Type Information): Ditto.
21268 (User-provided marking routines for template types): Make
21270 * doc/invoke.texi (AArch64 Options): Move before
21271 "Adapteva Epiphany Options".
21273 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
21274 Jakub Jelinek <jakub@redhat.com>
21277 * asan.c (gate_asan): Lookup no_sanitize_address instead of
21278 no_address_safety_analysis attribute.
21279 * doc/extend.texi (no_address_safety_attribute): Rename to
21280 no_sanitize_address attribute, mention no_address_safety_analysis
21281 attribute as deprecated alias.
21283 2013-02-28 Jakub Jelinek <jakub@redhat.com>
21285 PR middle-end/56461
21286 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
21287 type to vec<vec<tree> > *.
21288 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
21289 to be vec<tree> instead of vec<tree> *, set vec_defs
21290 to vNULL and call vec_defs.create (number_of_vects), adjust other
21292 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
21293 vectorizable_condition): Adjust vect_get_slp_defs callers.
21295 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
21297 * config/aarch64/aarch64.c
21298 (aarch64_float_const_representable): Remove unused variable.
21300 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
21302 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
21304 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
21306 * config/aarch64/aarch64-builtins.c
21307 (aarch64_init_simd_builtins): Make static.
21309 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
21311 * config/aarch64/aarch64.c
21312 (aarch64_simd_make_constant): Make static.
21314 2013-02-28 Martin Jambor <mjambor@suse.cz>
21316 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
21317 with no initialization to the RHS of debug statements.
21319 2013-02-28 Martin Jambor <mjambor@suse.cz>
21321 PR tree-optimization/56294
21322 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
21324 (get_access_replacement): Do not call create_access_replacement.
21325 Assert a replacement exists.
21326 (get_repl_default_def_ssa_name): Create the replacement declaration
21329 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21331 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
21332 final_end_function.
21334 2013-02-28 Marek Polacek <polacek@redhat.com>
21336 PR rtl-optimization/56466
21337 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
21338 if we're changing a loop.
21339 (peel_loops_completely): Likewise.
21341 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
21344 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
21346 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
21349 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
21350 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
21351 INTX_FTYPE_FX, FX_FTYPE_INTX.
21352 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
21354 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
21356 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
21357 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
21358 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
21359 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
21360 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
21361 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
21362 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
21363 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
21364 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
21365 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
21366 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
21367 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
21368 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
21369 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
21370 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
21371 (avrxmega6): Increase max flash segments from 5 to 6.
21372 * config/avr/t-multilib: Regenerate.
21373 * config/avr/avr-tables.opt: Regenerate.
21374 * doc/avr-mmcu.texi: Regenerate.
21376 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
21378 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
21379 (avr_device_to_arch): Rename to avr_device_to_ld.
21380 (avr_device_to_as): New prototype.
21381 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
21382 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
21383 * config/avr/driver-avr.c (avr_device_to_as): New.
21384 (avr_device_to_arch): Rename to avr_device_to_ld.
21386 2013-02-27 Jakub Jelinek <jakub@redhat.com>
21388 PR middle-end/56461
21389 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
21390 method on dr_chain and result_chain.
21392 PR middle-end/56461
21393 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
21394 pointer_set_destroy on not_executed_last_iteration.
21396 PR middle-end/56461
21397 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
21399 PR middle-end/56461
21400 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
21401 FOR_EACH_DEFINED_FUNCTION when freeing state.
21403 PR middle-end/56461
21404 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
21406 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
21409 PR middle-end/56461
21410 * ipa-cp.c (decide_whether_version_node): Call vec_free on
21411 known_aggs[i].items and release known_aggs vector.
21413 PR middle-end/56461
21414 * ipa-reference.c (propagate): Free node_info even for alias nodes.
21416 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21418 * config/microblaze/microblaze.c (microblaze_emit_compare):
21419 Use xor for EQ/NE comparisions.
21420 * config/microblaze/microblaze.md (cstoresf4): Add constraints
21421 (cbranchsf4): Adjust operator to comparison_operator.
21423 2013-02-27 Jakub Jelinek <jakub@redhat.com>
21425 PR middle-end/56461
21426 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
21428 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
21429 vec_safe_push, always update *slot.
21430 (redirect_edge_var_map_clear): Use vec_free.
21431 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
21432 (free_var_map_entry): Use vec_free.
21433 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
21434 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
21436 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
21438 PR middle-end/45472
21439 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
21440 when the may_trap_p bit of the exprs being merged differs.
21441 Reorder tests for speculativeness in the logical and operator.
21443 2013-02-27 Jakub Jelinek <jakub@redhat.com>
21445 * incpath.c (add_standard_paths): Use reconcat instead of concat
21446 where appropriate and avoid leaking memory.
21448 * opts.h: Include obstack.h.
21449 (opts_concat): New prototype.
21450 (opts_obstack): New declaration.
21451 * opts.c (opts_concat): New function.
21452 (opts_obstack): New variable.
21453 (init_options_struct): Call gcc_init_obstack on opts_obstack.
21454 (finish_options): Use opts_concat instead of concat
21455 and XOBNEWVEC instead of XNEWVEC.
21456 * opts-common.c (generate_canonical_option, decode_cmdline_option,
21457 generate_option): Likewise.
21458 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
21459 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
21462 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
21463 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
21465 2013-02-26 Jakub Jelinek <jakub@redhat.com>
21467 PR middle-end/56461
21468 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
21470 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
21472 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
21473 (arm_block_move_unaligned_straight): Likewise.
21474 (arm_adjust_block_mem): Likewise.
21476 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
21479 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
21480 temp, cond and label.
21481 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
21484 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
21485 * config/c6x/c6x.h (dbx_register_map): Update declaration.
21488 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
21489 of prologue/epilogue functions.
21492 * config/tilegx/tilegx.c (tilegx_expand_prologue):
21493 Remove unused variable cfa_offset.
21494 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
21497 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
21498 type promotion to unsigned.
21501 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
21502 for HOST_WIDE_INT of 32 bit / same size as int.
21503 (arm_block_move_unaligned_straight): Likewise.
21504 (arm_adjust_block_mem): Likewise.
21507 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
21510 2013-02-26 Marek Polacek <polacek@redhat.com>
21512 PR tree-optimization/56426
21513 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
21515 2013-02-26 Richard Biener <rguenther@suse.de>
21518 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
21519 unused variable loops.
21521 2013-02-26 Jakub Jelinek <jakub@redhat.com>
21523 PR tree-optimization/56448
21524 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
21525 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
21526 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
21527 later operands of the references, or even first operand for
21528 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
21530 PR tree-optimization/56443
21531 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
21532 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
21533 to type_for_mode langhook.
21535 2013-02-25 Matt Turner <mattst88@gmail.com>
21537 * doc/invoke.texi: Document r4700.
21539 2013-02-25 Richard Biener <rguenther@suse.de>
21541 PR tree-optimization/56175
21542 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
21544 (simplify_bitwise_binary): ... here. Also guard the conversion
21545 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
21547 2013-02-25 Catherine Moore <clm@codesourcery.com>
21550 2013-02-24 Catherine Moore <clm@codesourcery.com>
21551 Maciej W. Rozycki <macro@codesourcery.com>
21552 Tom de Vries <tom@codesourcery.com>
21553 Nathan Sidwell <nathan@codesourcery.com>
21554 Iain Sandoe <iain@codesourcery.com>
21555 Nathan Froyd <froydnj@codesourcery.com>
21556 Chao-ying Fu <fu@mips.com>
21558 * doc/extend.texi: (micromips, nomicromips, nocompression):
21559 Document new function attributes.
21560 * doc/invoke.texi (minterlink-compressed, mmicromips,
21561 m14k, m14ke, m14kec): Document new options.
21562 (minterlink-mips16): Update documentation.
21563 * doc/md.texi (ZC, ZD): Document new constraints.
21564 * configure.ac (gcc_cv_as_micromips): Check if linker
21565 supports the .set micromips directive.
21566 * configure: Regenerate.
21567 * config.in: Regenerate.
21568 * config/mips/mips-tables.opt: Regenerate.
21569 * config/mips/micromips.md: New file.
21570 * constraints.md (ZC, AD): New constraints.
21571 * config/mips/predicates.md (movep_src_register): New predicate.
21572 (movep_src_operand): New predicate.
21573 (non_volatile_mem_operand): New predicate.
21574 * config/mips/mips.md (multimem): New type.
21575 (length): Differentiate between 17-bit and 18-bit branch offsets.
21576 (MOVEP1, MOVEP2): New mode iterator.
21577 (mov_<load>l): Use ZC constraint.
21578 (mov_<load>r): Likewise.
21579 (mov_<store>l): Likewise.
21580 (mov_<store>r): Likewise.
21581 (*branch_equality<mode>_inverted): Add microMIPS support.
21582 (*branch_equality<mode>): Likewise.
21583 (*jump_absolute): Likewise.
21584 (indirect_jump_<mode>): Likewise.
21585 (tablejump_<mode>): Likewise.
21586 (<optab>_internal): Likewise.
21587 (sibcall_internal): Likewise.
21588 (sibcall_value_internal): Likewise.
21589 (prefetch): Use constraint ZD.
21590 * config/mips/mips.opt (minterlink-compressed): New option.
21591 (minterlink-mips16): Now an alias for minterlink-compressed.
21592 (mmicromips): New option.
21593 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
21594 (compare_and_swap_12): Likewise.
21595 (sync_add<mode>): Likewise.
21596 (sync_<optab>_12): Likewise.
21597 (sync_old_<optab>_12): Likewise.
21598 (sync_new_<optab>_12): Likewise.
21599 (sync_nand_12): Likewise.
21600 (sync_old_nand_12): Likewise.
21601 (sync_new_nand_12): Likewise.
21602 (sync_sub<mode>): Likewise.
21603 (sync_old_add<mode>): Likewise.
21604 (sync_old_sub<mode>): Likewise.
21605 (sync_new_add<mode>): Likewise.
21606 (sync_new_sub<mode>): Likewise.
21607 (sync_<optab><mode>): Likewise.
21608 (sync_old_<optab><mode>): Likewise.
21609 (sync_new_<optab><mode>): Likewise.
21610 (sync_nand<mode>): Likewise.
21611 (sync_old_nand<mode>): Likewise.
21612 (sync_new_nand<mode>): Likewise.
21613 (sync_lock_test_and_set<mode>): Likewise.
21614 (test_and_set_12): Likewise.
21615 (atomic_compare_and_swap<mode>): Likewise.
21616 (atomic_exchange<mode>_llsc): Likewise.
21617 (atomic_fetch_add<mode>_llsc): Likewise.
21618 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
21619 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
21620 (umips_save_restore_pattern_p): Likewise.
21621 (umips_load_store_pair_p): Likewise.
21622 (umips_output_load_store_pair): Likewise.
21623 (umips_movep_target_p): Likewise.
21624 (umips_12bit_offset_address_p): Likewise.
21625 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
21626 (mips_base_mips16): Rename this...
21627 (mips_base_compression_flags): ...to this. Update all uses.
21628 (mips_attribute_table): Add micromips, nomicromips and nocompression.
21629 (mips_mips16_decl_p): Delete.
21630 (mips_nomips16_decl_p): Delete.
21631 (mips_get_compress_on_flags): New function.
21632 (mips_get_compress_off_flags): New function.
21633 (mips_get_compress_mode): New function.
21634 (mips_get_compress_on_name): New function.
21635 (mips_get_compress_off_name): New function.
21636 (mips_insert_attributes): Support multiple compression types.
21637 (mips_merge_decl_attributes): Likewise.
21638 (umips_12bit_offset_address_p): New function.
21639 (mips_start_function_definition): Emit .set micromips directive.
21640 (mips_call_may_need_jalx_p): New function.
21641 (mips_function_ok_for_sibcall): Add microMIPS support.
21642 (mips_print_operand_punctuation): Support short delay slots and
21644 (umips_swm_mask, umips_swm_encoding): New.
21645 (umips_build_save_restore): New function.
21646 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
21647 (was_mips16_p): Remove.
21648 (old_compression_mode): New.
21649 (mips_set_compression_mode): New function.
21650 (mips_set_current_function): Add microMIPS support.
21651 (mips_option_override): Likewise.
21652 (umips_save_restore_pattern_p): New function.
21653 (umips_output_save_restore): New function.
21654 (umips_load_store_pair_p_1): New function.
21655 (umips_load_store_pair_p): New function.
21656 (umips_output_load_store_pair_1): New function.
21657 (umips_output_load_store_pair): New function.
21658 (umips_movep_target_p) New function.
21659 (mips_prepare_pch_save): Add microMIPS support.
21660 * config/mips/mips.h (TARGET_COMPRESSION): New.
21661 (TARGET_CPU_CPP_BUILTINS): Update macro
21662 to use new compression flags and to support microMIPS.
21663 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
21664 (MIPS_ARCH_FLOAT_SPEC): Likewise.
21665 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
21666 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
21667 (ASM_SPEC): Support mmicromips and mno-micromips.
21668 (M16STORE_REG_P): New macro.
21669 (MIPS_CALL): Support TARGET_MICROMIPS.
21670 (MICROMIPS_J): New macro.
21671 (mips_base_mips16): Rename this...
21672 (mips_base_compression_flags): ...to this.
21673 (UMIPS_12BIT_OFFSET_P): New macro.
21674 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
21675 (MULTILIB_DIRNAMES): Likewise.
21677 2013-02-25 Tom de Vries <tom@codesourcery.com>
21679 PR rtl-optimization/56131
21680 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
21681 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
21682 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
21684 2013-02-25 Tobias Burnus <burnus@net-b.de>
21686 * doc/invoke.texi (-fsanitize=): Move from optimization
21687 to debugging options.
21689 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
21691 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
21693 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
21694 Alexander Monakov <amonakov@ispras.ru>
21696 PR middle-end/56077
21697 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
21698 flush pending lists also on non-jumps. Adjust comment.
21700 2013-02-24 Catherine Moore <clm@codesourcery.com>
21701 Maciej W. Rozycki <macro@codesourcery.com>
21702 Tom de Vries <tom@codesourcery.com>
21703 Nathan Sidwell <nathan@codesourcery.com>
21704 Iain Sandoe <iain@codesourcery.com>
21705 Nathan Froyd <froydnj@codesourcery.com>
21706 Chao-ying Fu <fu@mips.com>
21708 * doc/extend.texi: (micromips, nomicromips, nocompression):
21709 Document new function attributes.
21710 * doc/invoke.texi (minterlink-compressed, mmicromips,
21711 m14k, m14ke, m14kec): Document new options.
21712 (minterlink-mips16): Update documentation.
21713 * doc/md.texi (ZC, ZD): Document new constraints.
21714 * configure.ac (gcc_cv_as_micromips): Check if linker
21715 supports the .set micromips directive.
21716 * configure: Regenerate.
21717 * config.in: Regenerate.
21718 * config/mips/mips-tables.opt: Regenerate.
21719 * config/mips/micromips.md: New file.
21720 * constraints.md (ZC, AD): New constraints.
21721 * config/mips/predicates.md (movep_src_register): New predicate.
21722 (movep_src_operand): New predicate.
21723 (non_volatile_mem_operand): New predicate.
21724 * config/mips/mips.md (multimem): New type.
21725 (length): Differentiate between 17-bit and 18-bit branch offsets.
21726 (MOVEP1, MOVEP2): New mode iterator.
21727 (mov_<load>l): Use ZC constraint.
21728 (mov_<load>r): Likewise.
21729 (mov_<store>l): Likewise.
21730 (mov_<store>r): Likewise.
21731 (*branch_equality<mode>_inverted): Add microMIPS support.
21732 (*branch_equality<mode>): Likewise.
21733 (*jump_absolute): Likewise.
21734 (indirect_jump_<mode>): Likewise.
21735 (tablejump_<mode>): Likewise.
21736 (<optab>_internal): Likewise.
21737 (sibcall_internal): Likewise.
21738 (sibcall_value_internal): Likewise.
21739 (prefetch): Use constraint ZD.
21740 * config/mips/mips.opt (minterlink-compressed): New option.
21741 (minterlink-mips16): Now an alias for minterlink-compressed.
21742 (mmicromips): New option.
21743 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
21744 (compare_and_swap_12): Likewise.
21745 (sync_add<mode>): Likewise.
21746 (sync_<optab>_12): Likewise.
21747 (sync_old_<optab>_12): Likewise.
21748 (sync_new_<optab>_12): Likewise.
21749 (sync_nand_12): Likewise.
21750 (sync_old_nand_12): Likewise.
21751 (sync_new_nand_12): Likewise.
21752 (sync_sub<mode>): Likewise.
21753 (sync_old_add<mode>): Likewise.
21754 (sync_old_sub<mode>): Likewise.
21755 (sync_new_add<mode>): Likewise.
21756 (sync_new_sub<mode>): Likewise.
21757 (sync_<optab><mode>): Likewise.
21758 (sync_old_<optab><mode>): Likewise.
21759 (sync_new_<optab><mode>): Likewise.
21760 (sync_nand<mode>): Likewise.
21761 (sync_old_nand<mode>): Likewise.
21762 (sync_new_nand<mode>): Likewise.
21763 (sync_lock_test_and_set<mode>): Likewise.
21764 (test_and_set_12): Likewise.
21765 (atomic_compare_and_swap<mode>): Likewise.
21766 (atomic_exchange<mode>_llsc): Likewise.
21767 (atomic_fetch_add<mode>_llsc): Likewise.
21768 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
21769 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
21770 (umips_save_restore_pattern_p): Likewise.
21771 (umips_load_store_pair_p): Likewise.
21772 (umips_output_load_store_pair): Likewise.
21773 (umips_movep_target_p): Likewise.
21774 (umips_12bit_offset_address_p): Likewise.
21775 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
21776 (mips_base_mips16): Rename this...
21777 (mips_base_compression_flags): ...to this. Update all uses.
21778 (mips_attribute_table): Add micromips, nomicromips and nocompression.
21779 (mips_mips16_decl_p): Delete.
21780 (mips_nomips16_decl_p): Delete.
21781 (mips_get_compress_on_flags): New function.
21782 (mips_get_compress_off_flags): New function.
21783 (mips_get_compress_mode): New function.
21784 (mips_get_compress_on_name): New function.
21785 (mips_get_compress_off_name): New function.
21786 (mips_insert_attributes): Support multiple compression types.
21787 (mips_merge_decl_attributes): Likewise.
21788 (umips_12bit_offset_address_p): New function.
21789 (mips_start_function_definition): Emit .set micromips directive.
21790 (mips_call_may_need_jalx_p): New function.
21791 (mips_function_ok_for_sibcall): Add microMIPS support.
21792 (mips_print_operand_punctuation): Support short delay slots and
21794 (umips_swm_mask, umips_swm_encoding): New.
21795 (umips_build_save_restore): New function.
21796 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
21797 (was_mips16_p): Remove.
21798 (old_compression_mode): New.
21799 (mips_set_compression_mode): New function.
21800 (mips_set_current_function): Add microMIPS support.
21801 (mips_option_override): Likewise.
21802 (umips_save_restore_pattern_p): New function.
21803 (umips_output_save_restore): New function.
21804 (umips_load_store_pair_p_1): New function.
21805 (umips_load_store_pair_p): New function.
21806 (umips_output_load_store_pair_1): New function.
21807 (umips_output_load_store_pair): New function.
21808 (umips_movep_target_p) New function.
21809 (mips_prepare_pch_save): Add microMIPS support.
21810 * config/mips/mips.h (TARGET_COMPRESSION): New.
21811 (TARGET_CPU_CPP_BUILTINS): Update macro
21812 to use new compression flags and to support microMIPS.
21813 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
21814 (MIPS_ARCH_FLOAT_SPEC): Likewise.
21815 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
21816 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
21817 (ASM_SPEC): Support mmicromips and mno-micromips.
21818 (M16STORE_REG_P): New macro.
21819 (MIPS_CALL): Support TARGET_MICROMIPS.
21820 (MICROMIPS_J): New macro.
21821 (mips_base_mips16): Rename this...
21822 (mips_base_compression_flags): ...to this.
21823 (UMIPS_12BIT_OFFSET_P): New macro.
21824 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
21825 (MULTILIB_DIRNAMES): Likewise.
21827 2013-02-24 Jakub Jelinek <jakub@redhat.com>
21830 * target-globals.c (save_target_globals): For init_reg_sets and
21831 target_reinit remporarily set this_fn_optabs to this_target_optabs.
21833 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
21835 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
21836 * config/aarch64/t-aarch64
21837 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
21839 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
21841 PR inline-asm/56148
21842 * lra-constraints.c (process_alt_operands): Reload operand
21843 conflicting with earlier clobber only if no more other conflicting
21846 2013-02-22 Jakub Jelinek <jakub@redhat.com>
21849 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
21850 if not linking a shared library.
21852 2013-02-22 Seth LaForge <sethml@google.com>
21854 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
21856 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
21858 * config/arm/arm.md (split for extendsidi): Update condition.
21859 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
21860 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
21861 (qhs_zextenddi_cstr): Likewise.
21863 2013-02-21 Jakub Jelinek <jakub@redhat.com>
21865 PR middle-end/56420
21866 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
21867 avoid signed wrapping.
21868 (expand_mult): Handle properly multiplication by
21869 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
21870 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
21871 in the compiler if coeff is HOST_WIDE_INT_MIN.
21872 (expand_divmod): Don't make ext_op1 static, change it's type to
21873 uhwi. Avoid undefined behavior in -INTVAL (op1).
21875 PR rtl-optimization/50339
21876 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
21878 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
21879 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
21880 into splitting_ashiftrt field.
21881 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
21883 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
21886 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
21888 PR middle-end/56108
21889 * trans-mem.c (execute_tm_mark): Do not expand transactions that
21890 are sure to go irrevocable.
21892 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
21894 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
21895 scalars are valid operands.
21897 2013-02-21 Martin Jambor <mjambor@suse.cz>
21899 PR tree-optimization/56310
21900 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
21901 only matching indices and non-negative final offsets.
21902 (intersect_aggregates_with_edge): Pass src_idx to
21903 agg_replacements_to_vector. Pass src_idx insstead of index to
21904 intersect_with_agg_replacements.
21906 2013-02-21 Martin Jambor <mjambor@suse.cz>
21908 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
21909 instead of hard-wired defaults.
21911 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
21913 * doc/invoke.texi (MIPS Options): Update documentation of the
21914 floating-point multiply-accumulate instruction restrictions.
21916 2013-02-21 Kostya Serebryany <kcc@google.com>
21918 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
21919 asan_shadow_offset on x86_64 linux.
21921 2013-02-21 Richard Biener <rguenther@suse.de>
21923 PR tree-optimization/56415
21925 2013-02-11 Richard Biener <rguenther@suse.de>
21927 PR tree-optimization/56273
21928 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
21931 2013-02-21 Jakub Jelinek <jakub@redhat.com>
21934 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
21937 PR inline-asm/56405
21938 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
21939 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
21941 2013-02-20 Jan Hubicka <jh@suse.cz>
21943 PR tree-optimization/56265
21944 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
21945 when target is referenced for first time.
21947 2013-02-20 Richard Biener <rguenther@suse.de>
21949 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
21950 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
21951 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
21952 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
21953 not return anything.
21954 (rename_ssa_copies): Do not remove unused locals.
21955 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
21956 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
21957 * passes.c (execute_function_todo): Do not schedule unused locals
21958 removal if cleanup_tree_cfg did something.
21959 * tree-ssa-live.c (remove_unused_locals): Dump statistics
21960 about the number of removed locals.
21962 2013-02-20 Richard Biener <rguenther@suse.de>
21964 PR tree-optimization/56398
21965 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
21967 2013-02-20 Martin Jambor <mjambor@suse.cz>
21969 PR tree-optimization/55334
21970 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
21971 restricted pointers to arrays.
21973 2013-02-20 Richard Biener <rguenther@suse.de>
21974 Jakub Jelinek <jakub@redhat.com>
21976 PR tree-optimization/56396
21977 * tree-ssa-ccp.c (n_const_val): New static variable.
21978 (get_value): Return NULL for SSA names we don't have a lattice
21980 (ccp_initialize): Initialize n_const_val.
21981 * tree-ssa-copy.c (n_copy_of): New static variable.
21982 (init_copy_prop): Initialize n_copy_of.
21983 (get_value): Return NULL_TREE for SSA names we don't have a
21986 2013-02-20 Martin Jambor <mjambor@suse.cz>
21988 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
21990 2013-02-20 Richard Biener <rguenther@suse.de>
21992 * genpreds.c (write_lookup_constraint): Do not compare first
21993 letter of the constraint again.
21995 2013-02-20 Richard Biener <rguenther@suse.de>
21997 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
21999 (get_use_iv_cost): Terminate hashtable walk when coming across
22002 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
22004 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
22005 reassociation for avx2 targets.
22007 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22009 * config/microblaze/microblaze.c: microblaze_has_clz = 0
22010 Add version check for v8.10.a to enable microblaze_has_clz
22011 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
22012 version and TARGET_PATTERN_COMPARE check
22013 * config/microblaze/microblaze.md: New clzsi2 instruction
22015 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22017 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
22018 function before branching.
22020 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
22022 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
22024 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
22026 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
22028 PR middle-end/55889
22029 * sel-sched.c: Include ira.h.
22030 (implicit_clobber_conflict_p): New function.
22031 (moveup_expr): Use it.
22032 * Makefile.in (sel-sched.o): Depend on ira.h.
22034 2013-02-19 Richard Biener <rguenther@suse.de>
22036 PR tree-optimization/56384
22037 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
22038 (vn_hash_type): Split out from ...
22039 (vn_hash_constant_with_type): ... here.
22040 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
22041 (vn_phi_eq): Compare types from vn_phi_s structure.
22042 (vn_phi_lookup): Populate vn_phi_s type.
22043 (vn_phi_insert): Likewise.
22045 2013-02-19 Jakub Jelinek <jakub@redhat.com>
22047 PR tree-optimization/56350
22048 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
22049 if haven't found reduction or nested cycle operand, rather than
22050 asserting we must find it.
22052 PR tree-optimization/56381
22053 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
22056 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
22057 Jakub Jelinek <jakub@redhat.com>
22060 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
22061 (swap_optab_enable): Same.
22062 (init_all_optabs): Use argument instead of global.
22063 * tree.h (struct tree_optimization_option): New field target_optabs.
22064 * expr.h (init_all_optabs): Add argument to prototype.
22065 (TREE_OPTIMIZATION_OPTABS): New.
22066 (save_optabs_if_changed): Protoize.
22067 * optabs.h: Declare this_fn_optabs.
22068 * optabs.c (save_optabs_if_changed): New.
22069 Declare this_fn_optabs.
22070 (init_optabs): Add argument to init_all_optabs() call.
22071 * function.c (invoke_set_current_function_hook): Handle per
22073 * function.h (struct function): New field optabs.
22074 * config/mips/mips.c (mips_set_mips16_mode): Handle when
22075 optimization_current_node has changed.
22076 * target-globals.h (save_target_globals_default_opts): Protoize.
22077 * target-globals.c (save_target_globals_default_opts): New.
22079 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22082 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
22083 registers %fr12 and %fr12R as call used.
22086 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
22087 and HImode, require all displacements to be an integer multiple of
22089 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
22090 only allow QImode and HImode when reload is in progress and strict is
22091 true. Likewise for symbolic addresses. Use base14_operand to check
22092 displacements in REG+BASE addresses.
22094 2013-02-18 Richard Biener <rguenther@suse.de>
22096 PR tree-optimization/56366
22097 * tree-vect-loop.c (get_initial_def_for_induction): Properly
22098 handle sign-conversion of outer-loop initial induction value.
22100 2013-02-18 Richard Biener <rguenther@suse.de>
22102 PR middle-end/56349
22103 * cfghooks.c (merge_blocks): If we merge a latch into another
22104 block adjust references to it.
22105 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
22106 (verify_loop_structure): Verify that a recorded latch is in fact
22109 2013-02-18 Richard Biener <rguenther@suse.de>
22111 PR tree-optimization/56321
22112 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
22113 order SSA name release and virtual operand unlinking.
22115 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22117 * config/microblaze/microblaze.md (save_stack_block): Define.
22118 (restore_stack_block): Likewise.
22120 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22122 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
22123 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
22124 * config/microblaze/microblaze.c (microblaze_option_override):
22125 Bail out early for PIC modes when target does not support PIC.
22127 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22129 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
22130 Replace with a microblaze version.
22131 (microblaze_trampoline_init): Adapt for microblaze.
22132 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
22135 2013-02-16 Jakub Jelinek <jakub@redhat.com>
22136 Dodji Seketeli <dodji@redhat.com>
22139 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
22140 (instrument_mem_region_access): Do not forget to always put
22141 instrumentation of the of 'base' and 'base + len' in a "if (len !=
22142 0) statement, even for cases where either 'base' or 'base + len'
22143 are not instrumented -- because they have been previously
22144 instrumented. Simplify the logic by putting all the statements
22145 instrument 'base + len' inside a sequence, and then insert that
22146 sequence right before the current insertion point. Then, to
22147 instrument 'base + len', just get an iterator on that statement.
22148 And do not forget to update the pointer to iterator the function
22149 received as argument.
22151 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
22153 PR rtl-optimization/56348
22154 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
22156 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
22158 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
22159 (clean_graph_dump_file): Pass base to start_graph_dump.
22161 2013-02-14 Richard Henderson <rth@redhat.com>
22164 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
22166 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
22168 * collect2-aix.h: Define F_LOADONLY.
22170 2013-02-14 Richard Biener <rguenther@suse.de>
22173 * varasm.c (output_constant_def_1): Get the decl representing
22174 the constant as argument.
22175 (output_constant_def): Wrap output_constant_def_1.
22176 (make_decl_rtl): Use output_constant_def_1 with the decl
22177 representing the constant.
22178 (build_constant_desc): Optionally re-use a decl already
22179 representing the constant.
22180 (tree_output_constant_def): Adjust.
22182 2013-02-14 Dodji Seketeli <dodji@redhat.com>
22185 * asan.c (instrument_builtin_call): Really put the length of the
22186 second source argument into src1_len.
22188 2013-02-13 Jakub Jelinek <jakub@redhat.com>
22190 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
22191 argument. If it is false, don't create edge from then_bb to
22193 (insert_if_then_before_iter): Pass true to it.
22194 (build_check_stmt): Pass false to it.
22195 (transform_statements): Flush hash table only on extended basic
22196 block boundaries, rather than at the beginning of every bb.
22197 Don't flush hash table on nonfreeing_call_p calls.
22198 * tree-flow.h (nonfreeing_call_p): New prototype.
22199 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
22201 2013-02-13 David S. Miller <davem@davemloft.net>
22203 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
22205 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
22208 * ira.c (max_regno_before_ira): Move from ...
22210 (fix_reg_equiv_init): Use max_regno_before_ira instead of
22213 2013-02-13 Jakub Jelinek <jakub@redhat.com>
22215 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
22217 2013-02-13 Richard Biener <rguenther@suse.de>
22220 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
22221 globals in MEM_REFs.
22223 2013-02-13 Richard Biener <rguenther@suse.de>
22225 * loop-init.c (loop_optimizer_init): Clear loop state when
22226 re-initializing preserved loops.
22227 * loop-unswitch.c (unswitch_single_loop): Return whether
22228 we unswitched the loop. Do not verify loop state here.
22229 (unswitch_loops): When we unswitched a loop discover new loops.
22231 2013-02-13 Kostya Serebryany <kcc@google.com>
22233 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
22235 * sanitizer.def: Rename __asan_init to __asan_init_v1.
22237 2013-02-12 Dodji Seketeli <dodji@redhat.com>
22239 Avoid instrumenting duplicated memory access in the same basic block
22240 * Makefile.in (asan.o): Add new dependency on hash-table.h
22241 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
22242 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
22243 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
22244 (free_mem_ref_resources, has_mem_ref_been_instrumented)
22245 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
22246 (get_mem_ref_of_assignment): New functions.
22247 (get_mem_refs_of_builtin_call): Extract from
22248 instrument_builtin_call and tweak a little bit to make it fit with
22250 (instrument_builtin_call): Use the new
22251 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
22252 of is_gimple_builtin_call.
22253 (instrument_derefs, instrument_mem_region_access): Insert the
22254 instrumented memory reference into the hash table.
22255 (maybe_instrument_assignment): Renamed instrument_assignment into
22256 this, and change it to advance the iterator when instrumentation
22257 actually happened and return true in that case. This makes it
22258 homogeneous with maybe_instrument_assignment, and thus give a
22259 chance to callers to be more 'regular'.
22260 (transform_statements): Clear the memory reference hash table
22261 whenever we enter a new BB, when we cross a function call, or when
22262 we are done transforming statements. Use
22263 maybe_instrument_assignment instead of instrumentation. No more
22264 need to special case maybe_instrument_assignment and advance the
22265 iterator after calling it; it's now handled just like
22266 maybe_instrument_call. Update comment.
22268 2013-02-13 Richard Biener <rguenther@suse.de>
22270 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
22271 Fix loop discovery code.
22273 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
22275 PR inline-asm/56148
22276 * lra-constraints.c (process_alt_operands): Match early clobber
22277 operand with itself. Check conflicts with earlyclobber only if
22278 the operand is not reloaded. Prefer to reload conflicting operand
22279 if earlyclobber and matching operands are the same.
22281 2013-02-12 Richard Biener <rguenther@suse.de>
22284 * lto-streamer-out.c (write_symbol): Do not output symbols
22285 for hard register variables.
22287 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
22290 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
22291 (umulsidi3_insn, mulsidi3_insn): New insns.
22293 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
22295 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
22296 (struct tune_params): Add vec_costs field.
22297 * config/arm/arm.c (arm_builtin_vectorization_cost)
22298 (arm_add_stmt_cost): New functions.
22299 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
22300 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
22301 (arm_default_vec_cost): New struct of type cpu_vec_costs.
22302 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
22303 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
22304 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
22305 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
22307 2013-02-12 Richard Biener <rguenther@suse.de>
22310 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
22311 decls again if possible.
22313 2013-02-12 Richard Biener <rguenther@suse.de>
22315 PR middle-end/56288
22316 * tree-ssa.c (verify_ssa_name): Fix check, move
22317 SSA_NAME_IN_FREE_LIST check up.
22319 2013-02-12 Jakub Jelinek <jakub@redhat.com>
22320 Steven Bosscher <steven@gcc.gnu.org>
22322 PR rtl-optimization/56151
22323 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
22324 equal to op0 or op1, and last_insn pattern is CODE operation
22325 with MEM dest and one of the operands matches that MEM.
22327 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
22329 * doc/extend.texi: Document Function Multiversioning and "default"
22330 parameter string to target attribute.
22331 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
22332 target attribute parameter is "default".
22333 (ix86_compare_version_priority): Remove checks for target attribute.
22334 (ix86_mangle_function_version_assembler_name): Change error to sorry.
22335 Remove check for target attribute equal to NULL. Add assert.
22336 (ix86_generate_version_dispatcher_body): Change error to sorry.
22338 2013-02-11 Iain Sandoe <iain@codesourcery.com>
22339 Jack Howarth <howarth@bromo.med.uc.edu>
22340 Patrick Marlier <patrick.marlier@gmail.com>
22343 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
22344 define ENDFILE_SPEC as TM_DESTRUCTOR.
22345 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
22347 2013-02-11 Alexander Potapenko <glider@google.com>
22348 Jack Howarth <howarth@bromo.med.uc.edu>
22349 Jakub Jelinek <jakub@redhat.com>
22352 * config/darwin.c (cdtor_record): Rename ctor_record.
22353 (sort_cdtor_records): Rename sort_ctor_records.
22354 (finalize_dtors): New routine to sort destructors by
22355 priority before use in assemble_integer.
22356 (machopic_asm_out_destructor): Use finalize_dtors if needed.
22358 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
22360 PR rtl-optimization/56275
22361 * simplify-rtx.c (avoid_constant_pool_reference): Check that
22362 offset is non-negative and less than cmode size before
22363 calling simplify_subreg.
22365 2013-02-11 Richard Biener <rguenther@suse.de>
22367 PR tree-optimization/56264
22368 * cfgloop.h (fix_loop_structure): Adjust prototype.
22369 * loop-init.c (fix_loop_structure): Return the number of
22370 newly discovered loops.
22371 * tree-cfgcleanup.c (repair_loop_structures): When new loops
22372 are discovered, do a full loop-closed SSA rewrite.
22374 2013-02-11 Richard Biener <rguenther@suse.de>
22376 PR tree-optimization/56273
22377 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
22379 (check_array_ref): Fix missing newline in dumps.
22380 (search_for_addr_array): Likewise.
22382 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
22384 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
22386 2013-02-09 Jakub Jelinek <jakub@redhat.com>
22389 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
22391 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
22393 PR rtl-optimization/56246
22394 * lra-constraints.c (simplify_operand_subreg): Try to reuse
22396 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
22397 constraints are satisfied.
22399 2013-02-08 Jeff Law <law@redhat.com>
22402 * emit-rtl.c (reg_is_parm_p): New function.
22403 * regs.h (reg_is_parm_p): New prototype.
22404 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
22405 callee-clobbered registers.
22407 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
22410 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
22411 If there is no implicit builtin declaration, just return NULL.
22413 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
22415 * config/i386/sse.md (FMAMODEM): New mode iterator.
22416 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
22417 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
22419 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
22421 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
22422 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
22423 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
22425 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22427 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
22428 (microblaze*-*-elf): Likewise.
22429 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
22431 * config/microblaze/microblaze-c.c: Add builtin defines for
22432 _LITTLE_ENDIAN and _BIG_ENDIAN.
22433 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
22434 add to TARGET_DEFAULT flags.
22435 Expand ASM_SPEC and LINK_SPEC.
22436 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
22437 * config/microblaze/microblaze.md: Update extendsidi2 and
22438 movdi_internal instructions to use low-order / high-order reg
22440 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
22441 options and inversemask / mask of LITTLE_ENDIAN.
22442 * config/microblaze/t-microblaze: Expand multilib options to
22443 include mlittle-endian (le) and update exceptions patterns.
22445 2013-02-08 Jakub Jelinek <jakub@redhat.com>
22447 PR rtl-optimization/56195
22448 * lra-constraints.c (get_reload_reg): Don't reuse regs
22449 if they have smaller mode than requested, if they have
22450 wider mode than requested, try to return a SUBREG.
22452 PR tree-optimization/56250
22453 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
22454 if type is unsigned and code isn't MULT_EXPR.
22456 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
22458 PR tree-optimization/56064
22459 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
22460 bits according to mode.
22461 * fixed-value.h (fixed_from_double_int)
22462 (const_fixed_from_double_int): Adjust comments.
22464 2013-02-08 Richard Biener <rguenther@suse.de>
22467 * lto-streamer.h (struct data_in): Remove current_file, current_line
22468 and current_col members.
22469 * lto-streamer-out.c (lto_output_location): Stream changed bits
22470 en-block for efficiency.
22471 * lto-streamer-in.c (clear_line_info): Remove.
22472 (lto_input_location): Cache current file, line and column
22473 globally via local statics. Read changed bits en-block.
22474 (input_function): Do not call clear_line_info.
22475 (lto_read_body): Likewise.
22476 (lto_input_toplevel_asms): Likewise.
22478 2013-02-08 Michael Matz <matz@suse.de>
22480 PR tree-optimization/52448
22481 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
22482 (nt_call_phase): New static.
22483 (add_or_mark_expr): Only mark accesses with newer phase than any
22485 (nonfreeing_call_p): New.
22486 (nt_init_block): Update nt_call_phase, mark blocks as visited.
22487 (nt_fini_block): Keep blocks marked as visited.
22488 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
22490 2013-02-08 Richard Biener <rguenther@suse.de>
22492 * ira.c (ira): Free broken dominator information.
22494 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
22496 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
22498 2013-02-08 Marek Polacek <polacek@redhat.com>
22500 * cfgloop.c (verify_loop_structure): Add more checking of headers.
22502 2013-02-08 Richard Biener <rguenther@suse.de>
22504 PR middle-end/56181
22505 * cfgloop.h (flow_loops_find): Adjust.
22506 (bb_loop_header_p): Declare.
22507 * cfgloop.c (bb_loop_header_p): New function split out from ...
22508 (flow_loops_find): ... here. Adjust function signature,
22509 support incremental loop structure update.
22510 (verify_loop_structure): Cleanup. Verify a loop is a loop.
22511 * cfgloopmanip.c (fix_loop_structure): Move ...
22512 * loop-init.c (fix_loop_structure): ... here.
22513 (apply_loop_flags): Split out from ...
22514 (loop_optimizer_init): ... here.
22515 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
22516 in incremental mode, only remove dead loops here.
22518 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
22521 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
22522 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
22523 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
22524 (*round<mode>3.libgcc): New insns for fixed-modes.
22525 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
22526 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
22527 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
22528 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
22529 implementations. Define to __builtin_avr_absFX,
22530 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
22531 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
22532 __builtin_avr_countlsFX, respectively.
22533 * config/avr/avr-c.c (target.h): Include it.
22534 (enum avr_builtin_id): New enum.
22535 (avr_resolve_overloaded_builtin): New static function.
22536 (avr_register_target_pragmas): Use it to set
22537 targetm.resolve_overloaded_builtin.
22538 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
22539 tree nodes used by DEF_BUILTIN.
22540 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
22541 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
22542 <AVR_BUILTIN_xxBITS>: Same.
22544 2013-02-08 Richard Biener <rguenther@suse.de>
22546 * cfgloop.c (verify_loop_structure): Properly handle
22547 a loop exiting to another loop header.
22548 * ira-int.h (ira_loops): Remove.
22549 * ira.c (ira_loops): Remove.
22550 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
22551 (do_reload): Use loop_optimizer_finalize.
22552 * ira-build.c (create_loop_tree_nodes): Use get_loops and
22553 number_of_loops to access the loop tree.
22554 (more_one_region_p): Likewise.
22555 (finish_loop_tree_nodes): Likewise.
22556 (rebuild_regno_allocno_maps): Likewise.
22557 (mark_loops_for_removal): Likewise.
22558 (mark_all_loops_for_removal): Likewise.
22559 (remove_unnecessary_regions): Likewise.
22560 (ira_build): Likewise.
22561 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
22563 2013-02-08 Richard Biener <rguenther@suse.de>
22565 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
22566 * ipa-pure-const.c (analyze_function): Avoid calling
22567 mark_irreducible_loops twice.
22568 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
22570 2013-02-07 David S. Miller <davem@davemloft.net>
22572 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
22574 * var-tracking.c (vt_add_function_parameter): Test the presence of
22575 HAVE_window_save properly and do not remap argument registers when
22576 we have a leaf function.
22578 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
22581 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
22583 * config/i386/i386.c (ix86_print_operand): Ditto.
22585 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
22587 * lra-constraints.c (process_alt_operands): Fix recently added comment.
22589 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
22591 PR rtl-optimization/56225
22592 * lra-constraints.c (process_alt_operands): Check that reload hard
22593 reg can hold value for strict_low_part.
22595 2013-02-07 Jakub Jelinek <jakub@redhat.com>
22598 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
22599 dwarf2out_end_function.
22600 (in_first_function_p, maybe_at_text_label_p,
22601 first_loclabel_num_not_at_text_label): New variables.
22602 (dwarf2out_var_location): In the first function find out
22603 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
22604 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
22607 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
22609 PR rtl-optimization/56178
22610 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
22611 SUBREG of a register. Tidy up related block of code.
22612 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
22613 note if the source is a register or a SUBREG of a register.
22615 2013-02-07 Jakub Jelinek <jakub@redhat.com>
22618 * config/rs6000/rs6000.md (ptrm): New mode attr.
22619 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
22620 call_value_indirect_aix<pttrsize>,
22621 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
22624 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
22626 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
22627 if -bnortl. Convert to strcmp and strncmp.
22629 2013-02-07 Alan Modra <amodra@gmail.com>
22632 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
22633 addresses won't wrap when offsetting.
22634 (rs6000_secondary_reload): Provide secondary reloads needed for
22635 wrapping LO_SUM addresses.
22637 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
22639 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
22640 MACH, just __MACH__.
22642 2013-02-06 Richard Biener <rguenther@suse.de>
22644 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
22645 instead of calling fix_loop_structure.
22647 2013-02-06 Jakub Jelinek <jakub@redhat.com>
22649 PR middle-end/56217
22650 * omp-low.c (use_pointer_for_field): Return false if
22651 lower_send_shared_vars doesn't generate any copy-out code.
22653 2013-02-06 Tom de Vries <tom@codesourcery.com>
22655 PR rtl-optimization/56131
22656 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
22657 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
22658 of the label is NULL. Add comment.
22660 2013-02-05 Jakub Jelinek <jakub@redhat.com>
22662 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
22665 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
22666 (STATIC_LIBTSAN_LIBS): Likewise.
22667 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
22668 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
22669 is defined, don't add anything else beyond that.
22670 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
22671 (LINK_COMMAND_SPEC): Use them.
22673 PR tree-optimization/56205
22674 * tree-stdarg.c (check_all_va_list_escapes): Return true if
22675 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
22676 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
22678 2013-02-05 Richard Biener <rguenther@suse.de>
22680 PR tree-optimization/53342
22681 PR tree-optimization/53185
22682 * tree-vectorizer.h (vect_check_strided_load): Remove.
22683 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
22684 not disallow peeling for vectorized strided loads.
22685 (vect_check_strided_load): Make static and simplify.
22686 (vect_analyze_data_refs): Adjust.
22687 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
22688 correctly when vectorizing strided loads.
22690 2013-02-05 Richard Biener <rguenther@suse.de>
22692 * doc/install.texi: Refer to ISL, not PPL.
22694 2013-02-05 Jan Hubicka <jh@suse.cz>
22696 PR tree-optimization/55789
22697 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
22699 2013-02-05 Jan Hubicka <jh@suse.cz>
22701 PR tree-optimization/55789
22702 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
22703 the dead call anyway.
22705 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
22708 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
22710 2013-02-04 Alexander Potapenko <glider@google.com>
22711 Jack Howarth <howarth@bromo.med.uc.edu>
22712 Jakub Jelinek <jakub@redhat.com>
22715 * config/darwin.c (sort_ctor_records): Stabilized qsort
22716 on constructor priority by using original position.
22717 (finalize_ctors): New routine to sort constructors by
22718 priority before use in assemble_integer.
22719 (machopic_asm_out_constructor): Use finalize_ctors if needed.
22721 2013-02-04 Jakub Jelinek <jakub@redhat.com>
22724 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
22725 about visibility on artificial decls.
22726 * config/sol2.c (solaris_assemble_visibility): Likewise.
22728 2013-02-04 Kai Tietz <ktietz@redhat.com>
22731 * config/i386/i386.c (function_value_ms_64): Add additional valtype
22732 argument and improve checking of return-argument types for 16-byte
22734 (ix86_function_value_1): Add additional valtype argument on call
22735 of function_value_64.
22736 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
22737 handling infunction_value_64 function.
22739 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
22741 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
22743 2013-02-04 Richard Biener <rguenther@suse.de>
22745 PR tree-optimization/56188
22746 * tree-ssa-structalias.c (label_visit): Consider case with
22747 initially non-empty points-to set.
22748 (perform_var_substitution): Dump node mapping and clean up.
22750 2013-02-04 Richard Guenther <rguenther@suse.de>
22753 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
22754 node prevail as last resort.
22755 (lto_symtab_merge_decls): Remove guard on LTRANS here.
22756 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
22758 2013-02-04 Richard Biener <rguenther@suse.de>
22760 PR tree-optimization/56113
22761 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
22763 (equiv_class_lookup_or_add): ... this.
22764 (label_visit): Adjust and fix error in previous patch.
22765 (perform_var_substitution): Adjust.
22767 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
22769 * config/sh/divtab.c: Fix formatting and comments throughout the file.
22770 * config/sh/sh4-300.md: Likewise.
22771 * config/sh/sh4a.md: Likewise.
22772 * config/sh/constraints.md: Likewise.
22773 * config/sh/sh.md: Likewise.
22774 * config/sh/netbsd-elf.h: Likewise.
22775 * config/sh/predicates.md: Likewise.
22776 * config/sh/sh-protos.h: Likewise.
22777 * config/sh/ushmedia.h: Likewise.
22778 * config/sh/linux.h: Likewise.
22779 * config/sh/sh.c: Likewise.
22780 * config/sh/superh.h: Likewise.
22781 * config/sh/elf.h: Likewise.
22782 * config/sh/sh4.md: Likewise.
22783 * config/sh/sh.h: Likewise.
22785 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22787 * config/pa/constraints.md: Adjust unused letters. Change "T"
22788 constraint to match_test floating_point_store_memory_operand().
22789 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
22790 (base14_operand): New.
22791 (floating_point_store_memory_operand): New.
22792 (integer_store_memory_operand): Revise to use base14_operand and
22793 reg_plus_base_memory_operand.
22794 (move_dest_operand): Allow symbolic_memory_operands.
22795 (symbolic_memory_operand): Check for LO_SOM.
22796 (symbolic_operand): Change default case to break.
22797 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
22798 CONST_DOUBLE values to be reloaded by putting them into memory when
22799 the destination is a floating point register.
22800 (movdf): Remove code to handle CONST_DOUBLE.
22802 (reload_indf_r1): New.
22803 (reload_insf_r1): New.
22804 Consistently use "Q" and "T" constraints with integer and floating
22805 point move instructions, respectively.
22806 (movdi): Remove FAIL.
22807 Change predicate for source operand unamed DImode move from
22808 general_operand to move_src_operand.
22809 (umulsidi3): Change predicate for destination operand to
22811 Likewise for similar unamed patterns.
22812 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
22813 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
22814 (hppa_legitimize_address): Simplify mask calculation.
22815 (pa_emit_move_sequence): Revised handling of secondary reloads from
22816 REG+D addresses for floating point loads and stores. Directly handle
22817 loading CONST0_RTX (mode) to a floating point register.
22818 (pa_secondary_reload): Handle reloading DF and SFmode constant values
22819 to floating point registers. Don't restrict secondary reloads to
22820 floating point registers to integer modes. Revise some comments and
22822 (TARGET_LEGITIMATE_ADDRESS_P): Define.
22823 (pa_legitimate_address_p): New.
22824 (pa_legitimize_reload_address): New.
22825 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
22826 (STRICT_REG_OK_FOR_BASE_P): New.
22827 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
22828 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
22830 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
22831 Andrew Dixie <andrewd@gentrack.com>
22833 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
22836 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
22838 * expmed.c (extract_bit_field_1): Pass the full width of the
22839 structure to get_best_reg_extraction_insn.
22841 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
22844 * configure.ac (use_cxa_atexit): Add AIX.
22845 * configure: Regenerate.
22847 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
22849 2013-02-01 Jakub Jelinek <jakub@redhat.com>
22852 * final.c (need_profile_function): New variable.
22853 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
22854 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
22855 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
22856 notes, targetm.asm_out.function_prologue doesn't emit anything,
22857 HAVE_prologue and profiler should be emitted before prologue,
22858 set need_profile_function instead of emitting it.
22859 (final_scan_insn): If need_profile_function, emit
22860 profile_function on the first NOTE_INSN_BASIC_BLOCK or
22861 NOTE_INSN_FUNCTION_BEG note.
22863 2013-02-01 Richard Henderson <rth@redhat.com>
22865 * config/rs6000/rs6000.md (smulditi3): New.
22868 * config/alpha/alpha.md (umulditi3): New.
22870 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
22872 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
22873 (ASM_OUTPUT_ALIGNED_LOCAL): New.
22875 2013-02-01 Richard Biener <rguenther@suse.de>
22877 PR tree-optimization/56113
22878 * tree-ssa-structalias.c (label_visit): Reduce work for
22879 single-predecessor nodes.
22881 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
22883 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
22884 range isn't testing for zero.
22886 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
22888 PR middle-end/56113
22889 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
22891 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
22892 Nick Clifton <nickc@redhat.com>
22894 * config/v850/constraints.md (Q): Define as a memory constraint.
22895 * config/v850/predicates.md (label_ref_operand): New predicate.
22896 (e3v5_shift_operand): New predicate.
22897 (ior_operator): New predicate.
22898 * config/v850/t-v850: Add e3v5 multilib.
22899 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
22900 (v850_gen_movdi): Prototype.
22901 * config/v850/v850.c: Add support for e3v5 architecture.
22902 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
22904 (construct_save_jarl): Add e3v5 long JARL support.
22905 (v850_adjust_insn_length): New function. Adjust length of call
22906 insns when using e3v5 instructions.
22907 (v850_gen_movdi): New function: Generate instructions to move a
22909 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
22910 (CPP_SPEC): Define __v850e3v5__ as appropriate.
22911 (TARGET_USE_FPU): Enable for e3v5.
22912 (CONST_OK_FOR_W): New macro.
22913 (ADJUST_INSN_LENGTH): Define.
22914 * config/v850/v850.md (UNSPEC_LOOP): Define.
22915 (attr cpu): Add v850e3v5.
22916 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
22917 (movdi): New pattern.
22918 (movdi_internal): New pattern.
22919 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
22920 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
22921 (cstoresf4): Likewise.
22922 (cstoredf4): Likewise.
22923 (insv): New pattern.
22924 (rotlso3_a): New pattern.
22925 (rotlsi3_b): New pattern
22926 (rotlsi3_v850e3v5): New pattern.
22927 (doloop_begin): New pattern.
22928 (fix_loop_counter): New pattern.
22929 (doloop_end): New pattern.
22930 (branch_normal): Add e3v5 long branch support.
22931 (branch_invert): Likewise.
22932 (branch_z_normal): Likewise.
22933 (branch_z_invert): Likewise.
22934 (branch_nz_normal): Likewise.
22935 (branch_nz_invert): Likewise.
22936 (call_internal_short): Add e3v5 register-indirect JARL support.
22937 (call_internal_long): Likewise.
22938 (call_value_internal_short): Likewise.
22939 (call_value_internal_long): Likewise.
22940 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
22941 (mloop): New option.
22942 * config.gcc: Add support for configuring v840e3v5 target.
22943 * doc/invoke.texi: Document new v850 specific command line options.
22945 2013-01-31 Paul Koning <ni1d@arrl.net>
22949 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
22950 children if parent is a class.
22951 (prune_unused_types_prune): Don't add DW_AT_declaration.
22953 2013-01-31 Richard Biener <rguenther@suse.de>
22955 PR tree-optimization/56157
22956 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
22957 match up operand with SLP child.
22959 2013-01-31 Jason Merrill <jason@redhat.com>
22962 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
22963 parameters the first time.
22964 (gen_scheduled_generic_parms_dies): Check completeness here.
22966 2013-01-31 Richard Biener <rguenther@suse.de>
22968 PR middle-end/53073
22969 * common.opt (faggressive-loop-optimizations): New flag,
22970 enabled by default.
22971 * doc/invoke.texi (faggressive-loop-optimizations): Document.
22972 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
22973 infer_loop_bounds_from_undefined by it.
22975 2013-01-31 Richard Biener <rguenther@suse.de>
22977 PR tree-optimization/56150
22978 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
22979 visit virtual operands.
22980 (find_uses_to_rename_bb): Likewise.
22982 2013-01-31 Richard Biener <rguenther@suse.de>
22984 PR tree-optimization/56150
22985 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
22986 mixed store non-store stmts.
22988 2013-01-30 Jakub Jelinek <jakub@redhat.com>
22991 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
22992 LIBASAN_EARLY_SPEC is defined.
22993 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
22994 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
22996 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
22999 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
23000 invalid args instead of ICEing on it.
23001 (ix86_valid_target_attribute_tree): Return error_mark_node if
23002 ix86_valid_target_attribute_inner_p failed.
23003 (ix86_valid_target_attribute_p): Return false only if
23004 ix86_valid_target_attribute_tree returned error_mark_node. Allow
23005 target("default") attribute.
23006 (sorted_attr_string): Change argument from const char * to tree,
23007 merge in all target attribute arguments rather than just one.
23008 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
23009 instead of free. Avoid using strcat.
23010 (ix86_mangle_function_version_assembler_name): Mangle
23011 target("default") as if no target attribute is present. Adjust
23012 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
23013 instead of xmalloc and XDELETEVEC instead of free.
23014 (ix86_function_versions): Don't return true if one of the decls
23015 doesn't have target attribute. If they don't and one of the decls
23016 is DECL_FUNCTION_VERSIONED, report an error. Adjust
23017 sorted_attr_string caller. Use XDELETEVEC instead of free.
23018 (ix86_supports_function_versions): Remove.
23019 (make_name): Fix up formatting.
23020 (make_dispatcher_decl): Remove resolver_name and its initialization.
23021 Avoid leaking memory.
23022 (is_function_default_version): Return true if there is
23023 target("default") attribute rather than no target attribute at all.
23024 (make_resolver_func): Avoid leaking memory.
23025 (ix86_generate_version_dispatcher_body): Likewise.
23026 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
23027 * target.def (supports_function_versions): Remove.
23028 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
23029 * doc/tm.texi: Regenerated.
23031 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
23033 PR rtl-optimization/56144
23034 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
23035 for values with side effects.
23037 2013-01-30 Richard Biener <rguenther@suse.de>
23039 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
23040 (sparseset_pop): Likewise.
23041 * cfganal.c (compute_idf): Likewise. Increase work-stack size
23042 to be able to use quick_push in the worker loop.
23044 2013-01-30 Marek Polacek <polacek@redhat.com>
23046 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
23048 2013-01-30 Richard Biener <rguenther@suse.de>
23051 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
23053 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
23055 PR tree-optimization/56064
23056 * fixed-value.c (fixed_from_double_int): New function.
23057 * fixed-value.h (fixed_from_double_int): New prototype.
23058 (const_fixed_from_double_int): New static inline function.
23059 * fold-const.c (native_interpret_fixed): New static function.
23060 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
23061 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
23062 (native_encode_fixed): New static function.
23063 (native_encode_expr) <FIXED_CST>: Use it.
23064 (native_interpret_int): Move double_int worker code to...
23065 * double-int.c (double_int::from_buffer): ...this new static method.
23066 * double-int.h (double_int::from_buffer): Prototype it.
23068 2013-01-30 Richard Biener <rguenther@suse.de>
23070 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
23071 New pointer-map and obstack.
23072 (init_alias_vars): Allocate pointer-map and obstack.
23073 (delete_points_to_sets): Free them.
23074 (find_what_var_points_to): Cache result.
23075 (find_what_p_points_to): Adjust for changed interface of
23076 find_what_var_points_to.
23077 (compute_points_to_sets): Likewise.
23078 (ipa_pta_execute): Likewise.
23080 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
23082 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
23083 * configure: Regenerate.
23084 * config.in: Regenerate.
23085 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
23086 #nobits/#progbits if supported.
23088 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
23091 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
23092 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
23093 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
23095 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
23097 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
23098 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
23100 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
23102 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
23103 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
23105 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
23107 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
23109 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
23110 * config/arm/cortex-a7.md: New bypasses using
23111 arm_mac_accumulator_is_result.
23113 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
23115 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
23116 (cortex_a7_neon_mla): Likewise.
23117 (cortex_a7_fpfmad): New reservation.
23118 (cortex_a7_fpmacs): Use ffmas and update required units.
23119 (cortex_a7_fpmuld): Update required units and latency.
23120 (cortex_a7_fpmacd): Likewise.
23121 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
23122 (cortex_a7_neon). Likewise.
23123 (bypass) Update participating units.
23125 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
23127 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
23128 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
23130 * config/arm/vfp11.md (vfp_farith): Use ffmas.
23131 (vfp_fmul): Use ffmad.
23132 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
23133 (cortex_r4_fmacd): Use ffmad.
23134 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
23135 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
23136 (cortex_a9_fmacd): Use ffmad.
23137 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
23138 (cortex_a8_vfp_macd): Use ffmad.
23139 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
23140 (cortex_a5_fpmacd): Use ffmad.
23141 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
23142 (cortex_a15_vfp_macd): Use ffmad.
23143 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
23145 2013-01-29 Jason Merrill <jason@redhat.com>
23148 * varasm.c (default_assemble_visibility): Don't warn about
23149 visibility on artificial decls.
23151 2013-01-29 Richard Biener <rguenther@suse.de>
23153 PR tree-optimization/56113
23154 * tree-ssa-structalias.c (equiv_class_lookup): Also return
23156 (label_visit): Free duplicate bitmaps and record the leader instead.
23157 (perform_var_substitution): Adjust.
23159 2013-01-29 Richard Biener <rguenther@suse.de>
23161 PR tree-optimization/55270
23162 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
23163 the CFG, schedule loops for fixup.
23165 2013-01-29 Nick Clifton <nickc@redhat.com>
23167 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
23170 2013-01-28 Leif Ekblad <leif@rdos.net>
23172 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
23173 * config/i386/i386.h (TARGET_RDOS): New macro.
23174 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
23175 * config/i386/i386.c (ix86_option_override_internal): For 64bit
23176 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
23177 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
23178 DEFAULT_LARGE_SECTION_THRESHOLD.
23179 * config/i386/i386.md (R14_REG, R15_REG): New constants.
23180 * config/i386/rdos.h: New file.
23181 * config/i386/rdos64.h: New file.
23183 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
23186 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
23189 2013-01-28 Jakub Jelinek <jakub@redhat.com>
23191 PR rtl-optimization/56117
23192 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
23193 call cselib_lookup_from_insn on the MEM before calling
23194 add_insn_mem_dependence.
23196 2013-01-28 Richard Biener <rguenther@suse.de>
23198 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
23199 to a stmt that didn't have one.
23200 (copy_phis_for_bb): Likewise for PHI arguments.
23201 (copy_debug_stmt): Likewise for debug stmts.
23203 2013-01-28 Richard Biener <rguenther@suse.de>
23205 PR tree-optimization/56034
23206 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
23207 (partition_builtin_p): Adjust.
23208 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
23209 it is the last partition.
23210 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
23211 up the vertex for the definition.
23212 (classify_partition): Classify whether a partition is a
23213 PKIND_REDUCTION, thus has uses outside of the loop.
23214 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
23215 Merge all PKIND_REDUCTION partitions into the last partition.
23216 (tree_loop_distribution): Seed partitions from reductions as well.
23218 2013-01-28 Jakub Jelinek <jakub@redhat.com>
23220 PR tree-optimization/56125
23221 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
23222 pow(x,c) into sqrt(x) * powi(x, n/2) or
23223 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
23224 optimizing for size.
23225 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
23226 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
23229 PR tree-optimization/56094
23230 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
23231 to UNKNOWN_LOCATION while gimplifying expr.
23233 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
23236 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
23237 operand 0 in movabs insn template for -masm=intel asm alternative.
23238 (*movabs<mode>_2): Ditto for operand 1.
23240 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
23243 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
23246 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23248 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
23251 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23253 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
23254 Undef to avoid warning.
23256 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
23258 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
23259 * configure: Regenerate.
23261 2013-01-25 Jakub Jelinek <jakub@redhat.com>
23263 PR tree-optimization/56098
23264 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
23265 for stmts with volatile ops.
23266 (cond_store_replacement): Don't optimize if assign has volatile ops.
23267 (cond_if_else_store_replacement_1): Don't optimize if either
23268 then_assign or else_assign have volatile ops.
23269 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
23272 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
23274 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
23276 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
23278 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
23279 missing ':' in asm example.
23281 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
23283 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
23284 entries into lane and laneq entries.
23285 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
23286 Remove AdvSIMD scalar modes.
23287 (aarch64_sq<r>dmulh_laneq<mode>): New.
23288 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
23290 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
23291 builtin implementations to relfect changes in RTL in aarch64-simd.md.
23292 * config/aarch64/iterators.md (VCOND): New.
23295 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
23298 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
23299 Add NULL LIBNAME argument to existing definitions.
23300 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
23301 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
23302 * config/avr/avr.c (DEF_BUILTIN): Same.
23303 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
23304 (avr_expand_builtin): Expand to a vanilla call if a libgcc
23305 implementation is available (DECL_ASSEMBLER_NAME is set).
23306 (avr_fold_absfx): New static function.
23307 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
23308 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
23309 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
23310 AVR_BUILTIN_ABSLLK.
23311 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
23312 (abshk, absk, abslk, absllk): Provide as static inline functions.
23314 2013-01-25 Marek Polacek <polacek@redhat.com>
23316 PR tree-optimization/56035
23317 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
23319 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
23321 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
23322 (*movtf_internal_rex64): Add (!o,C) alternative
23323 (*movxf_internal_rex64): Ditto.
23324 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
23326 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
23328 * doc/invoke.texi: fix typo.
23329 * doc/objc.texi: fix typo.
23331 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
23333 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
23334 for the first two alternatives.
23336 2013-01-24 Diego Novillo <dnovillo@google.com>
23338 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
23339 (ggc-zone.o): Remove.
23340 * configure.ac: Remove option --with-gc.
23341 * configure: Re-generate.
23342 * doc/install.texi: Remove documentation for --with-gc.
23343 * gengtype.c (write_enum_defn): Remove. Update all users.
23344 (write_Types_process_field): Remove generation of gt_e_* argument.
23345 (output_type_enum): Remove. Update all users.
23346 (write_enum_defn): Remove. Update all users.
23347 (enum alloc_zone): Remove. Update all users.
23348 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
23349 * ggc-common.c (ggc_splay_alloc): Remove first argument.
23350 Update all callers.
23351 (struct ptr_data): Remove field TYPE. Update all users.
23352 (gt_pch_note_object): Remove argument TYPE. Update all users.
23353 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
23355 * ggc-none.c (ggc_alloc_typed_stat): Remove.
23356 (struct alloc_zone): Remove.
23357 (ggc_internal_alloc_zone_stat): Remove.
23358 (ggc_internal_cleared_alloc_zone_stat): Remove.
23359 * ggc-page.c (ggc_alloc_typed_stat): Remove.
23360 (ggc_pch_count_object): Remove last argument. Update all users.
23361 (ggc_pch_alloc_object): Remove last argument. Update all users.
23362 (struct alloc_zone): Remove.
23363 * ggc-zone.c: Remove.
23364 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
23365 (struct alloc_zone): Remove.
23366 (ggc_alloc_typed_stat): Remove.
23367 (ggc_alloc_typed): Remove.
23368 (ggc_splay_alloc): Remove first argument.
23369 (rtl_zone): Remove. Update all users.
23370 (tree_zone): Remove. Update all users.
23371 (tree_id_zone): Remove. Update all users.
23372 (ggc_internal_zone_alloc_stat): Remove. Update all users.
23373 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
23374 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
23375 * tree-ssanames.c: Remove references to zone allocator in comments.
23377 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
23379 * config/avr/avr.c (avr_out_fract): Make register numbers that
23380 might be outside of source operand signed.
23382 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
23384 * config/i386/constraints.md (Yf): New constraint.
23385 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
23386 of f constraint to conditionaly disable x87 register preferences.
23387 (*movdf_internal): Ditto.
23388 (*movsf_internal): Ditto.
23390 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
23392 PR inline-asm/55934
23393 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
23394 that have operands with impossible constraints.
23395 Add a FIXME for a speed-up opportunity.
23396 * lra-constraints.c (process_alt_operands): Verify that a class
23397 selected from constraints on asms is valid for the operand mode.
23398 (curr_insn_transform): Remove incorrect comment.
23400 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
23402 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
23403 TOC operand is a valid symbol ref in the constant pool.
23405 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23407 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
23409 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
23412 * config/avr/stdfix.h: New file.
23413 * t-avr (stdfix-gcc.h): New rule to build it.
23414 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
23416 2013-01-23 Kostya Serebryany <kcc@google.com>
23418 * config/darwin.h: remove dependency on
23419 CoreFoundation (asan on Mac OS).
23421 2013-01-23 Jakub Jelinek <jakub@redhat.com>
23424 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
23425 instead of cmpdi_operand for first comparison operand.
23426 Don't assert that comparison operands aren't both constants.
23428 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
23430 * doc/install.texi (Downloading the Source): Update references to
23431 downloading separate components.
23433 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
23435 * doc/extend.texi (__int128): Improve grammar.
23437 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
23440 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
23441 alternative to (o,r).
23442 (*movdi_internal_rex64): Remove (!o,n) alternative.
23443 (DImode immediate->memory splitter): Remove.
23444 (DImode immediate->memory peephole2): Remove.
23445 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
23446 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
23447 alternative to (!o,*r).
23448 (*movtf_internal_sse): New pattern.
23449 (*movxf_internal_rex64): New pattern.
23450 (*movxf_internal): Disable for TARGET_64BIT.
23451 (*movdf_internal_rex64): Remove (!o,F) alternative.
23453 2013-01-22 Jakub Jelinek <jakub@redhat.com>
23455 PR middle-end/56074
23456 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
23457 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
23458 * tree-vect-loop-manip.c (find_loop_location): Also ignore
23459 stmt locations where LOCATION_LOCUS of the stmt location is
23460 UNKNOWN_LOCATION or BUILTINS_LOCATION.
23463 * config/i386/i386.md (UNSPEC_STOS): New.
23464 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
23465 *strsetqi_1): Add UNSPEC_STOS.
23467 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
23470 * doc/invoke.texi: Remove left over -Wsynth example.
23472 2013-01-21 Jakub Jelinek <jakub@redhat.com>
23474 PR tree-optimization/56051
23475 * fold-const.c (fold_binary_loc): Don't fold
23476 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
23477 a narrowing conversion, or widening conversion from signed
23480 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
23482 PR rtl-optimization/56023
23483 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
23484 dependent on debug instruction.
23486 2013-01-21 Martin Jambor <mjambor@suse.cz>
23488 PR middle-end/56022
23489 * function.c (allocate_struct_function): Call
23490 invoke_set_current_function_hook earlier.
23492 2013-01-21 Jakub Jelinek <jakub@redhat.com>
23494 * reload1.c (init_reload): Only initialize reload_obstack
23495 during the first call.
23497 2013-01-21 Marek Polacek <polacek@redhat.com>
23499 * cfgloop.c (verify_loop_structure): Fix up grammar.
23501 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
23503 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
23504 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
23506 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23509 * config/arm/marvell-pj4.md: Update copyright year.
23510 Fix up use of alu to alu_reg and simple_alu_imm.
23512 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
23514 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
23516 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
23519 * lra-constraints.c (curr_insn_transform): Don't reuse original
23520 insn for secondary memory move when memory mode should be different.
23522 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23524 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
23525 atomic_storedi_1): New patterns.
23527 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
23529 btver2 pipeline descriptions.
23530 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
23532 * config/i386/i386.md (btver2_decode): New type attributes.
23533 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
23535 * config/i386/btver2.md: New file describing btver2 pipelines.
23537 2013-01-19 Andrew Pinski <apinski@cavium.com>
23539 PR tree-optimization/52631
23540 * tree-ssa-sccvn (visit_use): Before looking up the original
23541 statement, try looking up the simplified expression.
23543 2013-01-19 Anthony Green <green@moxielogic.com>
23545 * config/moxie/moxie.c (moxie_expand_prologue): Set
23546 current_function_static_stack_size.
23548 2013-01-18 Jakub Jelinek <jakub@redhat.com>
23550 PR tree-optimization/56029
23551 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
23552 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
23554 2013-01-18 Sharad Singhai <singhai@google.com>
23556 PR tree-optimization/55995
23557 * dumpfile.c (dump_loc): Print location only if available.
23558 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
23560 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
23563 * lra-constraints.c (curr_insn_transform): Reuse original insn for
23564 secondary memory move.
23565 (inherit_reload_reg): Use rclass instead of cl for
23566 check_secondary_memory_needed_p.
23568 2013-01-18 Jakub Jelinek <jakub@redhat.com>
23570 PR middle-end/56015
23571 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
23572 the case where writing real complex part of target modifies op1.
23574 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
23576 * config/aarch64/aarch64-simd.md
23577 (aarch64_vcond_internal<mode>): Handle unordered cases.
23578 * config/aarch64/iterators.md (v_cmp_result): New.
23580 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
23581 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23583 * config/arm/marvell-pj4.md: New file.
23584 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
23585 * config/arm/arm.md (generic_sched): Add marvell_pj4.
23586 (generic_vfp): Likewise.
23587 * config/arm/arm-cores.def: Add marvell-pj4.
23588 * config/arm/arm-tune.md: Regenerate.
23589 * config/arm/arm-tables.opt: Regenerate.
23590 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
23591 * doc/invoke.texi: Document marvell-pj4.
23593 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
23595 * config/aarch64/arm_neon.h: Map scalar types to standard types.
23597 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
23602 * var-tracking.c (negative_power_of_two_p): New.
23603 (global_get_addr_cache, local_get_addr_cache): New.
23604 (get_addr_from_global_cache, get_addr_from_local_cache): New.
23605 (vt_canonicalize_addr): Rewrite using the above. Adjust the
23607 (vt_stack_offset_p): Remove.
23608 (vt_canon_true_dep): Always canonicalize loc's address.
23609 (clobber_overlapping_mems): Make sure we have a MEM.
23610 (local_get_addr_clear_given_value): New.
23611 (val_reset): Clear local cached entries.
23612 (compute_bb_dataflow): Create and release the local cache.
23613 Disable duplicate MEMs clobbering.
23614 (emit_notes_in_bb): Clobber MEMs likewise.
23615 (vt_emit_notes): Create and release the local cache.
23616 (vt_initialize, vt_finalize): Create and release the global
23617 cache, respectively.
23618 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
23620 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
23622 PR libmudflap/53359
23623 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
23624 not found in the symtab.
23626 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
23629 PR rtl-optimization/55547
23630 PR rtl-optimization/53827
23633 * alias.c (offset_overlap_p): New, factored out of...
23634 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
23635 the conservative special case for symbolic constants. Don't
23636 adjust zero sizes on alignment.
23638 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
23640 PR rtl-optimization/52573
23641 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
23642 REG_UNUSED for the same register.
23644 2013-01-17 Richard Biener <rguenther@suse.de>
23645 Marek Polacek <polacek@redhat.com>
23647 PR rtl-optimization/55833
23648 * loop-unswitch.c (unswitch_loops): Move loop verification...
23649 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
23650 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
23651 Set it to true when we're removing a loop from hierarchy tree in
23652 an irreducible region.
23653 (fix_bb_placements): Adjust caller.
23654 (fix_loop_placements): Likewise.
23656 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
23658 * config/avr/builtins.def (DEF_BUILTIN): Factor out
23659 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
23660 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
23661 Remove ID. Adjust comments.
23662 * config/avr/avr-c.c (avr_builtin_name): Remove.
23663 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
23664 * config/avr/avr.c (avr_tolower): New static function.
23665 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
23666 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
23667 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
23670 2013-01-17 Jan Hubicka <jh@suse.cz>
23672 PR tree-optimization/55273
23673 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
23675 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
23678 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
23679 store through atomic_store<mode>_1.
23680 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
23682 2013-01-17 Martin Jambor <mjambor@suse.cz>
23684 PR tree-optimizations/55264
23685 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
23686 for virtual methods.
23687 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
23688 virtual methods before inlining is over.
23689 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
23691 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
23694 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
23696 PR rtl-optimization/56005
23697 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
23698 pending reads for prefetch.
23700 2013-01-16 Ian Bolton <ian.bolton@arm.com>
23702 * config/aarch64/aarch64.md
23703 (*cstoresi_neg_uxtw): New pattern.
23704 (*cmovsi_insn_uxtw): New pattern.
23705 (*<optab>si3_uxtw): New pattern.
23706 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
23707 (*<optab>si3_insn_uxtw): New pattern.
23708 (*bswapsi2_uxtw): New pattern.
23710 2013-01-16 Richard Biener <rguenther@suse.de>
23712 * tree-inline.c (tree_function_versioning): Remove set but
23713 never used variable.
23715 2013-01-16 Richard Biener <rguenther@suse.de>
23717 PR tree-optimization/55964
23718 * tree-flow.h (rename_variables_in_loop): Remove.
23719 (rename_variables_in_bb): Likewise.
23720 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
23721 (copy_loop_before): Adjust and delete update-ssa status.
23722 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
23723 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
23724 (rename_variables_in_loop): Remove.
23725 (slpeel_update_phis_for_duplicate_loop): Likewise.
23726 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
23727 use available cfg machinery instead of duplicating it.
23728 Update PHI nodes and perform poor-mans SSA update here.
23729 (slpeel_tree_peel_loop_to_edge): Adjust.
23731 2013-01-16 Richard Biener <rguenther@suse.de>
23733 PR tree-optimization/54767
23734 PR tree-optimization/53465
23735 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
23736 (vrp_visit_phi_node): For PHI arguments coming via backedges
23737 drop all symbolical range information.
23738 (execute_vrp): Compute backedges.
23740 2013-01-16 Richard Biener <rguenther@suse.de>
23742 * doc/install.texi: Update CLooG and ISL requirements to
23745 2013-01-16 Christian Bruel <christian.bruel@st.com>
23748 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
23749 (broken_move): Handle UNSPECV_SP_SWITCH_B.
23750 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
23752 2013-01-16 DJ Delorie <dj@redhat.com>
23754 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
23755 (UNSPECV_SP_SWITCH_E): New.
23756 (sp_switch_1): Change to an unspec.
23757 (sp_switch_2): Change to an unspec. Don't use post-inc when we
23760 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
23762 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
23763 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
23764 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
23765 (expand_mem_thread_fence): Ditto.
23766 (expand_mem_signal_fence): Ditto.
23767 (expand_atomic_load): Ditto.
23768 (expand_atomic_store): Ditto.
23770 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
23772 PR rtl-optimization/55547
23773 PR rtl-optimization/53827
23776 * alias.c (memrefs_conflict_p): Set sizes to negative after
23779 2013-01-15 Jakub Jelinek <jakub@redhat.com>
23782 * function.c (thread_prologue_and_epilogue_insns): Always
23783 add crtl->drap_reg to set_up_by_prologue.set, even if
23784 stack_realign_drap is false.
23786 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
23788 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
23789 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
23790 *call): Fix indention.
23792 2013-01-15 Tom de Vries <tom@codesourcery.com>
23795 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
23798 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
23800 PR rtl-optimization/55153
23801 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
23803 2013-01-15 Martin Jambor <mjambor@suse.cz>
23805 PR tree-optimization/55920
23806 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
23807 accesses as grp_to_be_debug_replaced.
23809 2013-01-15 Jakub Jelinek <jakub@redhat.com>
23811 PR tree-optimization/55920
23812 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
23813 there is non-useless type conversion needed from debug rhs to lhs,
23814 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
23816 2013-01-15 Joseph Myers <joseph@codesourcery.com>
23817 Mikael Pettersson <mikpe@it.uu.se>
23820 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
23822 (ASM_OUTPUT_CASE_LABEL): Remove.
23823 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
23824 * final.c (shorten_branches): Update alignment of labels before
23825 jump tables if CASE_VECTOR_SHORTEN_MODE.
23827 2013-01-15 Richard Biener <rguenther@suse.de>
23830 * system.h: Do not include gmp.h for building host tools.
23832 2013-01-15 Richard Biener <rguenther@suse.de>
23834 PR middle-end/55882
23835 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
23836 account for bitpos when computing alignment.
23838 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
23840 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
23841 (ix86_target_macros_internal): Likewise.
23843 * config/i386/i386.c (m_CORE2I7): Removed.
23844 (m_CORE_HASWELL): New macro.
23845 (m_CORE_ALL): Likewise.
23846 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
23847 (initial_ix86_arch_features): Likewise.
23848 (processor_target_table): Initializations for Core avx2.
23849 (cpu_names): New names "core-avx2".
23850 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
23851 PROCESSOR_CORE_HASWELL.
23852 (ix86_issue_rate): New case.
23853 (ia32_multipass_dfa_lookahead): Likewise.
23854 (ix86_sched_init_global): Likewise.
23856 * config/i386/i386.h (TARGET_HASWELL): New macro.
23857 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
23858 (processor_type): New PROCESSOR_HASWELL.
23860 2013-01-15 Jakub Jelinek <jakub@redhat.com>
23862 PR tree-optimization/55955
23863 * tree-vect-loop.c (vectorizable_reduction): Give up early on
23864 *SHIFT_EXPR and *ROTATE_EXPR codes.
23866 PR tree-optimization/48766
23867 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
23868 -ftrapv disable -fwrapv.
23870 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
23873 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
23874 etc. to 1 and not to __flash.
23875 Use LL suffix for __INT24_MAX__ with -mint8.
23876 Use ULL suffix for __UINT24_MAX__ with -mint8.
23878 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
23880 * config/avr/avr-arch.h
23881 (struct base_arch_s): Use typedef avr_arch_t instead.
23882 (struct arch_info_s): Use typedef avr_arch_info_t instead.
23883 (struct mcu_type_s): Use typedef avr_mcu_t instead.
23884 * config/avr/avr.c: Same.
23885 * config/avr/avr-devices.c: Same.
23886 * config/avr/driver-avr.c: Same.
23887 * config/avr/gen-avr-mmcu-texi.c: Same.
23888 * config/avr/avr-mcus.def: Adjust comment.
23890 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
23892 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
23893 * config/aarch64/iterators.md (VALLDI): New.
23895 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
23896 Andi Kleen <ak@linux.intel.com>
23899 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
23900 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
23903 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
23905 * config/avr/avr-stdint.h: Remove trailing blanks.
23906 * config/avr/avr-log.h: Same.
23907 * config/avr/avr-arch.h: Same.
23908 * config/avr/avr-devices.c: Same.
23909 * config/avr/avr-dimode.md: Same.
23910 * config/avr/predicates.md: Same.
23911 * config/avr/avr-c.c: Same. And fix typo.
23913 * config/avr/avr-protos.h: Same. And:
23914 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
23915 (init_cumulative_args): Rename to avr_init_cumulative_args.
23916 (expand_prologue): Rename to avr_expand_prologue.
23917 (expand_epilogue): Rename to avr_expand_epilogue.
23918 (adjust_insn_length): Rename to avr_adjust_insn_length.
23919 (notice_update_cc): Rename to avr_notice_update_cc.
23920 (final_prescan_insn): Rename to avr_final_prescan_insn.
23921 * config/avr/avr.c: Same.
23922 * config/avr/avr.h: Same.
23923 * config/avr/avr.md: Remove trailing blanks.
23924 (prologue): Use avr_expand_prologue.
23925 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
23927 2013-01-14 Richard Biener <rguenther@suse.de>
23929 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
23930 verify_location, collect_subblocks): New functions.
23931 (verify_gimple_in_cfg): Verify that locations only reference
23932 BLOCKs in the functions BLOCK tree.
23934 2013-01-14 Richard Biener <rguenther@suse.de>
23936 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
23938 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
23940 (insert_out_of_ssa_copy_on_edge): Likewise.
23941 (rewrite_close_phi_out_of_ssa): Likewise.
23942 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
23944 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
23945 propagated constants.
23946 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
23949 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
23951 * config/avr/avr-modes.def: Add GPL copyright notice.
23953 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
23955 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
23956 MEMMODEL_MASK to determine memory model.
23957 (atomic_store<mode>): Ditto from operands[2].
23958 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
23960 2013-01-13 Jakub Jelinek <jakub@redhat.com>
23963 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
23964 (fold_gimple_assign): Don't call unshare_expr here.
23965 (fold_ctor_reference): Call unshare_expr.
23967 2013-01-13 Terry Guo <terry.guo@arm.com>
23969 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
23970 * doc/fragments.texi: Document MULTILIB_REUSE.
23971 * gcc.c (multilib_reuse): New internal spec.
23972 (set_multilib_dir): Also search multilib from multilib_reuse.
23973 * genmultilib (tmpmultilib3): Refactor code.
23974 (tmpmultilib4): Ditto.
23975 (multilib_reuse): New multilib argument.
23977 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
23979 * Makefile.in: Update copyright.
23981 2013-01-12 Tom de Vries <tom@codesourcery.com>
23983 PR middle-end/55890
23984 * calls.c (expand_call): Check if arg_nr is valid.
23986 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
23988 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
23989 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
23990 documentation. Add missing '__' in front of
23991 __builtin_ia32_packssdw256.
23993 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23996 * config/s390/s390.c (s390_preferred_reload_class): Do not return
23997 NO_REGS for larl operands.
23998 (s390_reload_larl_operand): Use s390_load_address instead of
24001 2013-01-11 Richard Biener <rguenther@suse.de>
24003 * tree-cfg.c (verify_node_sharing_1): Split out from ...
24004 (verify_node_sharing): ... here.
24005 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
24007 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
24009 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
24010 Substitute TREECHECKING.
24011 * configure: Regenerate.
24012 * Makefile.in (TREECHECKING): New.
24014 2013-01-11 Richard Guenther <rguenther@suse.de>
24016 PR tree-optimization/44061
24017 * tree-vrp.c (extract_range_basic): Compute zero as
24018 value-range for __builtin_constant_p of function parameters.
24020 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
24022 Update copyright years.
24024 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
24026 PR rtl-optimization/55672
24027 * lra-eliminations.c (mark_not_eliminable): Permit addition with
24028 const to be eliminable.
24030 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
24032 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
24033 * configure: Regenerate.
24035 2013-01-10 Richard Biener <rguenther@suse.de>
24037 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
24039 2013-01-10 Richard Biener <rguenther@suse.de>
24042 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
24043 locations for virtual PHI arguments.
24044 (rewrite_update_phi_arguments): Likewise.
24046 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
24048 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
24051 2013-01-10 Jakub Jelinek <jakub@redhat.com>
24053 PR tree-optimization/55921
24054 * tree-complex.c (expand_complex_asm): New function.
24055 (expand_complex_operations_1): Call it for GIMPLE_ASM.
24057 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24060 * config/s390/s390.c (s390_symref_operand_p)
24061 (s390_loadrelative_operand_p): Merge the two functions.
24062 (s390_check_qrst_address, print_operand_address): Add parameters
24063 to s390_loadrelative_operand_p invokation.
24064 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
24065 (s390_reload_larl_operand, s390_secondary_reload): Use
24066 s390_loadrelative_operand_p instead of s390_symref_operand_p.
24067 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
24069 2013-01-09 Mike Stump <mikestump@comcast.net>
24071 * dse.c (record_store): Remove unnecessary assert.
24073 2013-01-09 Jan Hubicka <jh@suse.cz>
24075 PR tree-optimization/55569
24076 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
24077 * cfgloop.h (scale_loop_profile): Likewise.
24079 2013-01-09 Jan Hubicka <jh@suse.cz>
24082 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
24084 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
24086 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
24088 PR middle-end/55114
24089 * expr.h (maybe_emit_group_store): Declare.
24090 * expr.c (maybe_emit_group_store): New function.
24091 * builtins.c (expand_builtin_int_roundingfn): Call it.
24092 (expand_builtin_int_roundingfn_2): Likewise.
24094 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
24096 PR rtl-optimization/55829
24097 * lra-constraints.c (match_reload): Add code for absent output.
24098 (curr_insn_transform): Add code for reloads of matched inputs
24101 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
24103 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
24104 attribute of movddup insn to DF.
24105 (*vec_interleave_lowv2df): Ditto.
24106 (vec_dupv2df): Ditto.
24108 2013-01-09 Jan Hubicka <jh@suse.cz>
24110 PR tree-optimiation/55875
24111 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
24112 EVERY_ITERATION parameter.
24113 (number_of_iterations_exit): Check if exit is executed every iteration.
24114 (idx_infer_loop_bounds): Similarly here.
24115 (n_of_executions_at_most): Simplify
24116 to only test for cases where statement is dominated by the
24117 particular bound; handle correctly the "postdominance" test.
24118 (scev_probably_wraps_p): Use max loop iterations info
24119 as a global bound first.
24121 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
24122 Nick Clifton <nickc@redhat.com>
24124 * config/v850/v850.md (cbranchsf4): New pattern.
24125 (cstoresf4): New pattern.
24126 (cbranchdf4): New pattern.
24127 (cstoredf4): New pattern.
24128 (movsicc): Disallow floating point comparisons.
24129 (cmpsf_le_insn): Fix order of operators.
24130 (cmpsf_lt_insn): Likewise.
24131 (cmpsf_eq_insn): Likewise.
24132 (cmpdf_le_insn): Likewise.
24133 (cmpdf_lt_insn): Likewise.
24134 (cmpdf_eq_insn): Likewise.
24135 (cmpsf_ge_insn): Use LE comparison.
24136 (cmpdf_ge_insn): Likewise.
24137 (cmpsf_gt_insn): Use LT comparison.
24138 (cmpdf_gt_insn): Likewise.
24139 (cmpsf_ne_insn): Delete pattern.
24140 (cmpdf_ne_insn): Delete pattern.
24141 * config/v850/v850.c (v850_gen_float_compare): Use
24142 gen_cmpdf_eq_insn for NE comparison.
24143 (v850_float_z_comparison_operator)
24144 (v850_float_nz_comparison_operator): Move from here ...
24145 * config/v850/predicates.md: ... to here. Move GT and GE
24146 comparisons into v850_float_z_comparison_operator.
24147 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
24149 (v850_float_nz_comparison_operator): Likewise.
24151 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24153 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
24154 with calls to gen_insvsi/gen_insvdi.
24156 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
24158 * config/i386/i386.c (initial_ix86_tune_features): Set up
24159 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
24161 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
24162 Jakub Jelinek <jakub@redhat.com>
24164 PR tree-optimization/48189
24165 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
24166 If nitercst is 0, don't predict the exit edge.
24168 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24170 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
24171 in asm_fprintf with reg_names.
24172 (aarch64_print_operand_address): Likewise.
24173 (aarch64_return_addr): Likewise.
24174 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
24176 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24178 * config/pa/pa.h (VAL_U6_BITS_P): Define.
24179 (INT_U6_BITS): Likewise.
24180 * config/pa/predicates.md (uint6_operand): New predicate.
24181 (shift5_operand, shift6_operand): Likewise.
24182 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
24184 (lshrdi3): Use shift6_operand.
24185 (shrpsi4, shrpdi4): New insn patterns.
24186 (extzv): Delete expander.
24187 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
24188 predicates in unamed zero extract patterns. Tighten common constraint.
24189 (extv): Delete expander.
24190 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
24191 predicates in unamed sign extract patterns. Tighten common constraint.
24192 (insv): Delete expander.
24193 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
24194 predicates in unamed insert patterns. Tighten common constraint.
24195 Change uint32_operand predicate to uint6_operand predicate in unamed
24196 DImode pattern to insert constant values of type 1...1xxxx.
24198 2013-01-04 Jan Hubicka <jh@suse.cz>
24200 PR tree-optimization/55823
24201 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
24204 2013-01-08 Jakub Jelinek <jakub@redhat.com>
24205 Uros Bizjak <ubizjak@gmail.com>
24207 PR rtl-optimization/55845
24208 * df-problems.c (can_move_insns_across): Stop scanning at
24209 volatile_insn_p source instruction or give up if
24210 across_from .. across_to range contains any volatile_insn_p
24213 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
24215 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
24216 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
24218 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
24219 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
24221 2013-01-08 Jakub Jelinek <jakub@redhat.com>
24224 * asan.c (asan_clear_shadow): New function.
24225 (asan_emit_stack_protection): Use it.
24227 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
24229 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
24230 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
24231 with tab instead of space.
24233 2013-01-08 Nick Clifton <nickc@redhat.com>
24235 * config/rl78/rl78.c (rl78_expand_prologue): Always select
24236 register bank 0 at the start of an interrupt handler.
24237 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
24240 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
24242 * config/aarch64/aarch64-simd.md
24243 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
24244 (aarch64_simd_bsl): Likewise.
24245 (aarch64_vcond_internal<mode>): Likewise.
24246 (vcond<mode><mode>): Likewise.
24247 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
24248 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
24250 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
24252 * config/aarch64/aarch64-builtins.c
24253 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
24255 2013-01-08 Martin Jambor <mjambor@suse.cz>
24258 * tree-sra.c (analyze_access_subtree): Return true also after
24259 potentially creating a debug-only replacement.
24261 2013-01-08 Jakub Jelinek <jakub@redhat.com>
24263 PR middle-end/55890
24264 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
24266 PR tree-optimization/54120
24267 * tree-vrp.c (range_fits_type_p): Don't allow
24268 src_precision < precision from signed vr to unsigned_p
24269 if vr->min or vr->max is negative.
24270 (simplify_float_conversion_using_ranges): Test can_float_p
24271 against CODE_FOR_nothing.
24273 2013-01-08 Jakub Jelinek <jakub@redhat.com>
24274 Richard Biener <rguenther@suse.de>
24276 PR middle-end/55851
24277 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
24278 types instead of just INTEGER_TYPE types.
24280 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
24282 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
24285 2013-01-07 Steve Ellcey <sellcey@mips.com>
24288 * config/mips/mips.opt: Change mad to mmad to match documentation.
24290 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
24293 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
24294 .progmemx.data now.
24296 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
24299 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
24300 (avr_addrspace_t): Add .section_name field.
24301 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
24303 (avr_addrspace): Same. Initialize .section_name. Remove last
24304 NULL entry. Put __memx into .progmemx.data.
24305 (progmem_section_prefix): Remove.
24306 (avr_asm_init_sections): No need to initialize progmem_section.
24307 (avr_asm_named_section): Use avr_addrspace[].section_name to get
24308 section name prefix.
24309 (avr_asm_select_section): Ditto. And use get_unnamed_section to
24310 retrieve the progmem section.
24311 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
24312 boundary to run over avr_addrspace[].
24313 (avr_register_target_pragmas): Ditto.
24315 2013-01-06 Jakub Jelinek <jakub@redhat.com>
24317 * varasm.c (output_constant_def_contents): For asan_protect_global
24318 protected strings, adjust DECL_ALIGN if needed, before testing for
24320 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
24321 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
24323 (output_object_block): For asan protected decls, emit asan padding
24324 after their contents.
24325 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
24326 (asan_finish_file): Test it here instead.
24328 2013-01-07 Nick Clifton <nickc@redhat.com>
24329 Matthias Klose <doko@debian.org>
24330 Doug Kwan <dougkwan@google.com>
24331 H.J. Lu <hongjiu.lu@intel.com>
24334 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
24336 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
24338 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
24340 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
24342 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
24344 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
24347 * doc/install.texi (Cross-Compiler-Specific Options): Document
24350 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
24352 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
24353 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
24354 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
24355 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
24356 vqmovun_high_s64): Fix source operand number and update copyright.
24358 2013-01-07 Richard Biener <rguenther@suse.de>
24360 PR middle-end/55890
24361 * gimple.h (gimple_call_builtin_p): New overload.
24362 * gimple.c (validate_call): New function.
24363 (gimple_call_builtin_p): Likewise.
24364 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
24365 Use gimple_call_builtin_p.
24366 (find_func_clobbers): Likewise.
24367 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
24368 (strlen_optimize_stmt): Likewise.
24370 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
24372 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
24373 (vld1q_dup_*): Likewise.
24374 (vld1_*): Likewise.
24375 (vld1q_*): Likewise.
24376 (vld1_lane_*): Likewise.
24377 (vld1q_lane_*): Likewise.
24379 2013-01-07 Richard Biener <rguenther@suse.de>
24381 * lto-streamer.h (LTO_minor_version): Bump to 2.
24383 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
24385 * config/aarch64/aarch64-protos.h
24386 (aarch64_const_double_zero_rtx_p): Rename to...
24387 (aarch64_float_const_zero_rtx_p): ...this.
24388 (aarch64_float_const_representable_p): New.
24389 (aarch64_output_simd_mov_immediate): Likewise.
24390 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
24391 move immediate case.
24392 * config/aarch64/aarch64.c
24393 (aarch64_const_double_zero_rtx_p): Rename to...
24394 (aarch64_float_const_zero_rtx_p): ...this.
24395 (aarch64_print_operand): Allow printing of new constants.
24396 (aarch64_valid_floating_const): New.
24397 (aarch64_legitimate_constant_p): Check for valid floating-point
24399 (aarch64_simd_valid_immediate): Likewise.
24400 (aarch64_vect_float_const_representable_p): New.
24401 (aarch64_float_const_representable_p): Likewise.
24402 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
24403 (aarch64_output_simd_mov_immediate): New.
24404 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
24405 (*movdf_aarch64): Likewise.
24406 * config/aarch64/constraints.md (Ufc): New.
24407 (Y): call aarch64_float_const_zero_rtx.
24408 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
24410 2013-01-07 Richard Biener <rguenther@suse.de>
24412 PR tree-optimization/55888
24413 PR tree-optimization/55862
24414 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
24415 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
24416 not if it is contained therein.
24418 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
24420 * config/avr/t-avr: Typo.
24422 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
24425 * config/avr/t-avr: Don't automatically rebuild
24426 $(srcdir)/config/avr/t-multilib
24427 $(srcdir)/config/avr/avr-tables.opt
24428 $(srcdir)/doc/avr-mmcu.texi
24429 (avr-mcus): New phony target to build them on request.
24430 (s-avr-mlib, s-avr-mmcu-texi): Remove.
24431 * avr/avr-mcus.def: Adjust comments.
24433 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
24435 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
24437 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
24439 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
24441 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
24443 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
24445 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
24447 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
24448 to generate profiling.
24449 * config/rs6000/aix64.h (LIB_SPEC): Same.
24451 2013-01-04 Andrew Pinski <apinski@cavium.com>
24453 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
24455 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
24457 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
24459 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
24461 (ix86_expand_move): Ditto.
24462 (ix86_zero_extend_to_Pmode): Ditto.
24463 (ix86_expand_call): Ditto.
24464 (ix86_expand_special_args_builtin): Ditto.
24465 (ix86_expand_builtin): Ditto.
24467 2013-01-04 Richard Biener <rguenther@suse.de>
24469 PR tree-optimization/55862
24470 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
24471 translating them through PHI nodes.
24473 2013-01-04 Martin Jambor <mjambor@suse.cz>
24475 PR tree-optimization/55755
24476 * tree-sra.c (sra_modify_assign): Do not check that an access has no
24477 children when trying to avoid producing a VIEW_CONVERT_EXPR.
24479 2013-01-04 Marek Polacek <polacek@redhat.com>
24481 PR middle-end/55859
24482 * opts.c (default_options_optimization): Clarify error message.
24484 2013-01-04 Richard Biener <rguenther@suse.de>
24486 PR middle-end/55863
24487 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
24490 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24493 * config/pa/pa.md (movsi): Revert previous change.
24494 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
24497 2013-01-03 Richard Henderson <rth@redhat.com>
24499 * config/i386/i386.c (ix86_expand_move): Always assign to op1
24500 after eliminating TLS symbols.
24502 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
24505 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
24506 * graphite-poly.c (debug_gmp_value): Likewise.
24508 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
24511 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
24512 selected code model, define __code_mode_small__, __code_model_medium__,
24513 __code_model_large__, __code_model_32__ or __code_model_kernel__.
24514 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
24515 xchg temporary register with %k. Declare temporary register as
24517 [__x86_64__]: For medium and large code models, preserve %rbx register.
24519 2013-01-03 Richard Biener <rguenther@suse.de>
24521 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
24522 (dump_subscript): Adjust.
24523 (finalize_ddr_dependent): Do not dump redundant info.
24524 (analyze_siv_subscript): Adjust.
24525 (subscript_dependence_tester): Likewise.
24526 (compute_affine_dependence): Likewise.
24528 2013-01-03 Richard Biener <rguenther@suse.de>
24531 2013-01-03 Richard Biener <rguenther@suse.de>
24533 PR tree-optimization/55857
24534 * tree-vect-stmts.c (vectorizable_load): Do not setup
24535 re-alignment for invariant loads.
24537 2013-01-02 Richard Biener <rguenther@suse.de>
24539 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
24540 invariant load do not generate a vector load from the scalar location.
24542 2013-01-03 Richard Biener <rguenther@suse.de>
24544 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
24545 for not vectorizing.
24546 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
24547 not build INDIRECT_REFs, call get_name once only.
24548 (vect_create_data_ref_ptr): Likewise. Dump base object kind
24549 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
24551 2013-01-03 Richard Biener <rguenther@suse.de>
24553 PR tree-optimization/55857
24554 * tree-vect-stmts.c (vectorizable_load): Do not setup
24555 re-alignment for invariant loads.
24557 2013-01-03 Richard Biener <rguenther@suse.de>
24560 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
24561 prefer a built-in decl.
24563 2013-01-03 Jakub Jelinek <jakub@redhat.com>
24565 * gcc.c (process_command): Update copyright notice dates.
24566 * gcov.c (print_version): Likewise.
24567 * gcov-dump.c (print_version): Likewise.
24569 PR rtl-optimization/55838
24570 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
24571 iv0.step, iv1.step and step.
24573 2013-01-03 Jakub Jelinek <jakub@redhat.com>
24574 Marc Glisse <marc.glisse@inria.fr>
24576 PR tree-optimization/55832
24577 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
24578 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
24579 integer_{one,zero}_node.
24581 2013-01-03 Jakub Jelinek <jakub@redhat.com>
24584 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
24585 * var-tracking.c (reverse_op): Don't add reverse ops to
24586 VALUEs that have already
24587 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
24589 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
24591 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
24593 2013-01-02 Teresa Johnson <tejohnson@google.com>
24595 * dumpfile.c (dump_loc): Print filename with location.
24596 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
24597 new location_t parameter to emit complete unroll message with
24598 new dump framework.
24599 (canonicalize_loop_induction_variables): Compute loops location
24600 and pass to try_unroll_loop_completely.
24601 * loop-unroll.c (report_unroll_peel): New function.
24602 (peel_loops_completely): Use new dump format with location
24603 for main dumpfile message, and invoke report_unroll_peel on success.
24604 (decide_unrolling_and_peeling): Ditto.
24605 (decide_peel_once_rolling): Remove old dumpfile message subsumed
24606 by report_unroll_peel.
24607 (decide_peel_completely): Ditto.
24608 (decide_unroll_constant_iterations): Ditto.
24609 (decide_unroll_runtime_iterations): Ditto.
24610 (decide_peel_simple): Ditto.
24611 (decide_unroll_stupid): Ditto.
24612 * cfgloop.c (get_loop_location): New function.
24613 * cfgloop.h (get_loop_location): Declare.
24615 2013-01-02 Sriraman Tallam <tmsriram@google.com>
24617 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
24620 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24622 PR middle-end/55198
24623 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
24624 BLKmode objects when EXPAND_MEMORY is specified.
24626 2013-01-02 Sriraman Tallam <tmsriram@google.com>
24628 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
24630 (fold_builtin_cpu): Do not share cpu model decls across statements.
24632 2013-01-02 Jason Merrill <jason@redhat.com>
24635 * tree.c (build_array_type_1): Revert earlier change.
24637 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
24639 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
24641 * config/aarch64/aarch64-tune.md: Re-generate.
24643 2013-01-02 Richard Biener <rguenther@suse.de>
24645 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
24646 invariant load do not generate a vector load from the scalar location.
24648 2013-01-02 Richard Biener <rguenther@suse.de>
24651 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
24652 * configure: Regenerate.
24654 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
24656 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
24657 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
24658 (expand_builtin_int_roundingfn_2): Keep the original target around
24659 for the fallback case.
24661 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
24663 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
24664 to be clear for sign changes.
24666 2013-01-01 Jan Hubicka <jh@suse.cz>
24668 * ipa-inline-analysis.c: Fix formatting.
24670 2013-01-01 Jakub Jelinek <jakub@redhat.com>
24672 PR tree-optimization/55831
24673 * tree-vect-loop.c (get_initial_def_for_induction): Use
24674 gsi_after_labels instead of gsi_start_bb.
24676 Copyright (C) 2013 Free Software Foundation, Inc.
24678 Copying and distribution of this file, with or without modification,
24679 are permitted in any medium without royalty provided the copyright
24680 notice and this notice are preserved.