1 2014-05-05 Marek Polacek <polacek@redhat.com>
4 * opts.c (common_handle_option): Call error_at instead of warning_at.
6 2014-05-05 Richard Biener <rguenther@suse.de>
8 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
9 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
10 under the TODO_verify_il umbrella.
12 2014-05-05 Richard Biener <rguenther@suse.de>
14 * passes.c (execute_function_todo): Move TODO_verify_flow under
15 the TODO_verify_ul umbrella.
17 2014-05-05 Richard Biener <rguenther@suse.de>
20 * fold-const.c (fold_binary_loc): Consistently avoid
21 canonicalizing X & CST away from a CST that is the mask
24 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
26 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
27 int argument to enum machine_mode.
28 (picochip_class_max_nregs): Ditto.
29 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
30 (picochip_class_max_nregs): Ditto.
32 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
34 * target.def: Add new target hook.
35 * doc/tm.texi: Regenerate.
36 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
37 * targhooks.c (default_keep_leaf_when_profiled): New function.
39 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
40 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
42 2014-05-05 Bin Cheng <bin.cheng@arm.com>
44 PR tree-optimization/60363
45 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
46 (copy_phi_args): New parameters. Call get_value_locus_in_path.
47 (update_destination_phis): New parameter.
48 (create_edge_and_update_destination_phis): Ditto.
49 (ssa_fix_duplicate_block_edges): Pass new arguments.
50 (thread_single_edge): Ditto.
52 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
54 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
55 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
56 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
57 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
58 Use RS6000_BTM_HARD_FLOAT.
59 (BU_MISC_2): Likewise.
60 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
61 RS6000_BTM_HARD_FLOAT.
62 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
64 (rs6000_invalid_builtin): Add hard floating builtin support.
65 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
67 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
69 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
71 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
72 Add missing function* argument.
74 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
76 * lra-constraints.c (valid_address_p): Move earlier in file.
77 Add a constraint argument to the address_info version.
78 (satisfies_memory_constraint_p): New function.
79 (satisfies_address_constraint_p): Likewise.
80 (process_alt_operands, curr_insn_transform): Use them.
81 (process_address): Pass the constraint to valid_address_p when
82 checking address operands.
84 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
86 * config/mips/mips.c (mips_isa_rev): New variable.
87 (mips_set_architecture): Set it.
88 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
90 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
91 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
92 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
93 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
94 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
95 conditions in terms of mips_isa_rev.
96 (mips_isa_rev): Declare.
98 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
100 * config/sh/sh-mem.cc: Use tabs instead of spaces.
101 (prob_unlikely, prob_likely): Make variables const.
103 2014-05-03 Denis Chertykov <chertykov@gmail.com>
105 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
107 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
109 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
111 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
113 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
114 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
115 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
117 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
118 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
120 Replace usage of ROUND_REG with sh_round_reg.
121 Use CEIL instead of ROUND_ADVANCE.
123 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
126 * config/sh/sh.c: Include stdlib headers before everything else.
128 2014-05-02 Jakub Jelinek <jakub@redhat.com>
130 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
131 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
132 (gimplify_adjust_omp_clauses): Simd region is never
133 directly nested in combined parallel. Instead, for linear
134 with copyin/copyout, if in combined for simd loop, make decl
135 firstprivate/lastprivate on OMP_FOR.
136 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
137 expand_omp_for_static_chunk): When setting endvar, also set
138 fd->loop.v to the same value.
140 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
142 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
144 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
146 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
149 2014-05-02 Marek Polacek <polacek@redhat.com>
151 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
153 2014-02-26 Kito Cheng <kito@0xlab.org>
155 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
156 to a C expression marco.
157 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
158 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
159 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
160 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
161 HONOR_REG_ALLOC_ORDER.
162 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
164 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
166 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
168 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
170 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
172 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
174 * tree-if-conv.c (is_cond_scalar_reduction): New function.
175 (convert_scalar_cond_reduction): Likewise.
176 (predicate_scalar_phi): Add recognition and transformation
177 of simple conditioanl reduction to be vectorizable.
179 2014-05-01 Marek Polacek <polacek@redhat.com>
182 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
184 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
186 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
187 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
188 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
189 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
190 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
191 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
192 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
193 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
195 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
197 * config/arc/arc.opt (mlra): Move comment above option name
198 to avoid mis-parsing as language options.
200 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
202 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
203 * config/sol2.h: ... here.
204 * config/sol2-10.h: Remove.
206 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
207 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
208 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
209 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
210 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
211 * config/sol2.h: ... here.
212 (SECTION_NAME_FORMAT): Don't redefine.
213 (STARTFILE_ARCH32_SPEC): Rename to ...
214 (STARTFILE_ARCH_SPEC): ... this.
215 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
216 * config/sparc/sol2.h: ... here.
217 (SECTION_NAME_FORMAT): Don't undef.
218 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
219 (SUBTARGET_EXTRA_SPECS): Remove.
220 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
222 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
223 (MD_STARTFILE_PREFIX): Remove.
224 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
225 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
226 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
227 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
228 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
229 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
230 * config/i386/sol2.h: ... here.
231 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
232 * config/i386/sol2-bi.h: Remove.
233 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
234 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
236 * config/i386/t-sol2-64: Rename to ...
237 * config/i386/t-sol2: ... this.
238 * config/sparc/t-sol2-64: Rename to ...
239 * config/sparc/t-sol2: ... this.
241 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
242 sol2_tm_file_head, sol2_tm_file_tail.
243 Include ${cpu_type}/sol2.h before sol2.h.
245 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
246 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
247 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
248 Reflect i386/t-sol2-64 renaming.
249 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
250 Reflect sparc/t-sol2-64 renaming.
252 2014-04-30 Richard Biener <rguenther@suse.de>
254 * passes.c (execute_function_todo): Move TODO_verify_stmts
255 and TODO_verify_ssa under the TODO_verify_il umbrella.
256 * tree-ssa.h (verify_ssa): Adjust prototype.
257 * tree-ssa.c (verify_ssa): Add parameter to tell whether
258 we should verify SSA operands.
259 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
260 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
261 whether we should verify whether not throwing stmts have EH info.
262 * graphite-scop-detection.c (create_sese_edges): Adjust.
263 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
264 * tree-eh.c (lower_try_finally_switch): Do not add the
265 default case label twice.
267 2014-04-30 Marek Polacek <polacek@redhat.com>
269 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
270 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
271 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
272 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
274 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
276 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
277 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
278 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
279 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
280 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
281 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
282 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
283 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
285 2014-04-29 David Malcolm <dmalcolm@redhat.com>
287 * tree-cfg.c (dump_function_to_file): Dump the return type of
288 functions, in a line to itself before the function body, mimicking
289 the layout of a C function.
291 2014-04-29 Jakub Jelinek <jakub@redhat.com>
293 PR tree-optimization/60971
294 * tree-tailcall.c (process_assignment): Reject conversions which
297 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
299 * calls.c (initialize_argument_information): Always treat
300 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
301 (expand_call): Likewise.
302 (emit_library_call_calue_1): Likewise.
303 * expr.c (PUSH_ARGS_REVERSED): Do not define.
304 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
307 2014-04-29 Nick Clifton <nickc@redhat.com>
309 * config/msp430/msp430.md (umulsidi): Fix typo.
310 (mulhisi3): Enable even inside interrupt handlers.
311 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
312 bigger return address pushed in large mode.
314 2014-04-29 Nick Clifton <nickc@redhat.com>
316 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
317 (arc_init_reg_tables): Use a machine_mode enum to iterate over
319 * config/m32r/m32r.c (init_reg_tables): Likewise.
320 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
321 enum to hold the modes.
323 2014-04-29 Richard Biener <rguenther@suse.de>
325 * dominance.c (free_dominance_info): Add overload with
327 (dom_info_state): Likewise.
328 (dom_info_available_p): Likewise.
329 * basic-block.h (free_dominance_info, dom_info_state,
330 dom_info_available_p): Declare overloads.
331 * passes.c (execute_function_todo): Verify that verifiers
332 don't change dominator info state. Drop dominator info
333 for IPA pass invocations.
334 * cgraph.c (release_function_body): Restore asserts that
335 dominator information is released.
337 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
339 * doc/invoke.texi: Fix typo.
340 * tree-vrp.c: Fix typos.
341 * gimple.c (infer_nonnull_range): Reorder operands of an &&
344 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
346 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
348 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
350 * config/aarch64/aarch64-builtins.c
351 (aarch64_types_storestruct_lane_qualifiers): New.
352 (TYPES_STORESTRUCT_LANE): Likewise.
353 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
354 (st3_lane): Likewise.
355 (st4_lane): Likewise.
356 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
357 (vec_store_lanesci_lane<mode>): Likewise.
358 (vec_store_lanesxi_lane<mode>): Likewise.
359 (aarch64_st2_lane<VQ:mode>): Likewise.
360 (aarch64_st3_lane<VQ:mode>): Likewise.
361 (aarch64_st4_lane<VQ:mode>): Likewise.
362 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
363 * config/aarch64/arm_neon.h
364 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
365 use new macro arguments.
366 (__ST3_LANE_FUNC): Likewise.
367 (__ST4_LANE_FUNC): Likewise.
368 * config/aarch64/iterators.md (V_TWO_ELEM): New.
369 (V_THREE_ELEM): Likewise.
370 (V_FOUR_ELEM): Likewise.
372 2014-04-28 David Malcolm <dmalcolm@redhat.com>
374 * doc/gimple.texi: Replace the description of the now-defunct
375 union gimple_statement_d with a diagram showing the
376 gimple_statement_base class hierarchy and its relationships to
377 the GSS_ and GIMPLE_ enums.
379 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
381 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
382 * config/aarch64/aarch64.c
383 (aarch64_cannot_change_mode_class): Weaken conditions.
384 (aarch64_modes_tieable_p): New.
385 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
387 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
389 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
390 (loadsync_<mode>): Change mode.
391 (load_quadpti, store_quadpti): New.
392 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
393 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
395 2014-04-28 Martin Jambor <mjambor@suse.cz>
397 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
398 same alias type as the original statement.
399 (subreplacement_assignment_data): New type.
400 (handle_unscalarized_data_in_subtree): New type of parameter,
401 generate new memory accesses with same alias type as the original
403 (load_assign_lhs_subreplacements): Likewise.
404 (sra_modify_constructor_assign): Generate new memory accesses with
405 same alias type as the original statement.
407 2014-04-28 Richard Biener <rguenther@suse.de>
409 * tree-pass.h (TODO_verify_il): Define.
410 (TODO_verify_all): Complete properly.
411 * passes.c (execute_function_todo): Move existing loop-closed
412 SSA verification under TODO_verify_il.
413 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
414 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
415 Fix tree sharing issue.
417 2014-04-28 Richard Biener <rguenther@suse.de>
420 * builtins.def (DEF_C11_BUILTIN): Add.
421 (BUILT_IN_ALIGNED_ALLOC): Likewise.
422 * coretypes.h (enum function_class): Add function_c11_misc.
423 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
424 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
425 (call_may_clobber_ref_p_1): Likewise.
426 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
427 (mark_all_reaching_defs_necessary_1): Likewise.
428 (propagate_necessity): Likewise.
429 (eliminate_unnecessary_stmts): Likewise.
430 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
432 2014-04-28 Richard Biener <rguenther@suse.de>
434 * tree-vrp.c (vrp_var_may_overflow): Remove.
435 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
436 with overflow immediately bump to one before that value and
437 let iteration figure out overflow status.
439 2014-04-28 Richard Biener <rguenther@suse.de>
441 * configure.ac: Do valgrind header checks unconditionally.
442 Add --enable-valgrind-annotations.
443 * system.h: Guard valgrind header inclusion with
444 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
445 * alloc-pool.c (pool_alloc, pool_free): Use
446 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
447 to guard possibly dead code.
448 * config.in: Regenerated.
449 * configure: Likewise.
451 2014-04-28 Jeff Law <law@redhat.com>
453 PR tree-optimization/60902
454 * tree-ssa-threadedge.c
455 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
456 over real defs when invalidating outputs from statements that do not
457 produce useful outputs for threading.
459 2014-04-28 Richard Biener <rguenther@suse.de>
461 PR tree-optimization/60979
462 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
463 SCOPs that end in a block with a successor with abnormal
466 2014-04-28 Richard Biener <rguenther@suse.de>
468 * tree-pass.h (execute_pass_list): Adjust prototype.
469 * passes.c (pass_manager::execute_early_local_passes):
471 (do_per_function): Change callback signature, push all actual
472 work to the callbals.
473 (do_per_function_toporder): Likewise.
474 (execute_function_dump): Adjust.
475 (execute_function_todo): Likewise.
476 (clear_last_verified): Likewise.
477 (verify_curr_properties): Likewise.
478 (update_properties_after_pass): Likewise.
479 (execute_pass_list_1): Split out from ...
480 (execute_pass_list): ... here. Adjust.
481 (execute_ipa_pass_list): Likewise.
482 * cgraphunit.c (cgraph_add_new_function): Adjust.
483 (analyze_function): Likewise.
484 (expand_function): Likewise.
485 * cgraph.c (release_function_body): Free dominance info
486 here instead of asserting it was magically freed elsewhere.
488 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
490 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
491 * configure: Regenerate.
492 * config/sparc/sparc.opt (muser-mode): New option.
493 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
495 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
496 * doc/invoke.texi (SPARC options): Document -muser-mode.
498 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
500 * cselib.c (find_slot_memmode): Delete.
501 (cselib_hasher): Change compare_type to a struct.
502 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
504 (preserve_constants_and_equivs): Adjust for new compare_type.
505 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
506 (wrap_constant): Delete.
507 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
509 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
511 * doc/install.texi (Building with profile feedback): Remove
514 2014-04-26 Tom de Vries <tom@codesourcery.com>
516 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
519 2014-04-25 Cary Coutant <ccoutant@google.com>
522 * dwarf2out.c (should_move_die_to_comdat): A type definition
523 can contain a subprogram definition, but don't move it to a
525 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
526 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
528 (clone_tree_hash): Rename to...
529 (clone_tree_partial): ...this; change callers. Copy
530 DW_TAG_subprogram DIEs as declarations.
531 (copy_decls_walk): Don't copy children of a declaration into a
534 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
537 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
540 2014-04-25 Jiong Wang <jiong.wang@arm.com>
542 * config/arm/predicates.md (call_insn_operand): Add long_call check.
543 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
545 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
548 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
550 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
553 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
555 PR tree-optimization/60930
556 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
557 creating a multiply candidate by folding two constant
558 multiplicands when the result overflows.
560 2014-04-25 Jakub Jelinek <jakub@redhat.com>
562 PR tree-optimization/60960
563 * tree-vect-generic.c (expand_vector_operation): Only call
564 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
566 2014-04-25 Tom de Vries <tom@codesourcery.com>
568 * expr.c (clobber_reg_mode): New function.
569 * expr.h (clobber_reg): New function.
571 2014-04-25 Tom de Vries <tom@codesourcery.com>
573 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
576 2014-04-25 Radovan Obradovic <robradovic@mips.com>
577 Tom de Vries <tom@codesourcery.com>
579 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
581 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
582 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
583 new argument to find_all_hard_reg_sets call.
585 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
587 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
588 Use HOST_WIDE_INT_C for mask literal.
589 (aarch_rev16_shleft_mask_imm_p): Likewise.
591 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
594 * config/sparc/sparc.md (ashlsi3_extend): Delete.
596 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
598 PR preprocessor/56540
599 * config/i386/i386-c.c (ix86_target_macros): Define
600 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
602 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
604 * configure.ac (tga_func): Remove.
605 (LIB_TLS_SPEC): Remove.
606 * configure: Regenerate.
607 * config.in: Regenerate.
608 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
610 2014-04-25 Richard Biener <rguenther@suse.de>
613 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
614 call stmt use/clobber sets during stmt walk instead of
615 walking the possibly incomplete set of caller edges.
617 2014-04-25 Richard Biener <rguenther@suse.de>
620 * passes.c (apply_ipa_transforms): Inline into only caller ...
621 (execute_one_pass): ... here. Properly bring in function
622 bodies for nodes we want to apply IPA transforms to.
624 2014-04-24 Cong Hou <congh@google.com>
626 PR tree-optimization/60896
627 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
628 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
629 (vect_mark_pattern_stmts): Set the def type of all statements in
630 PATTERN_DEF_SEQ as vect_internal_def.
632 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
634 * doc/extend.texi (PowerPC Built-in Functions): Document new
635 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
637 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
639 * config/rs6000/predicates.md (const_0_to_3_operand): New
640 predicate to match 0..3 integer constants.
642 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
643 to support adding miscellaneous builtin functions.
644 (BU_DFP_MISC_2): Likewise.
645 (BU_P7_MISC_1): Likewise.
646 (BU_P7_MISC_2): Likewise.
647 (BU_P8V_MISC_3): Likewise.
648 (BU_MISC_1): Likewise.
649 (BU_MISC_2): Likewise.
650 (DIVWE): Add extended divide builtin functions.
658 (DXEX): Add decimal floating-point builtin functions.
670 (CDTBCD): Add new BCD builtin functions.
674 (BCDADD_LT): Likewise.
675 (BCDADD_EQ): Likewise.
676 (BCDADD_GT): Likewise.
677 (BCDADD_OV): Likewise.
679 (BCDSUB_LT): Likewise.
680 (BCDSUB_EQ): Likewise.
681 (BCDSUB_GT): Likewise.
682 (BCDSUB_OV): Likewise.
683 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
684 (UNPACK_TD): Likewise.
686 (UNPACK_TF): Likewise.
687 (UNPACK_TF_0): Likewise.
688 (UNPACK_TF_1): Likewise.
689 (PACK_V1TI): Likewise.
690 (UNPACK_V1TI): Likewise.
692 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
693 support for decimal floating point builtin functions.
694 (rs6000_expand_ternop_builtin): Add checks for the new builtin
695 functions that take constant arguments.
696 (rs6000_invalid_builtin): Add decimal floating point builtin
698 (rs6000_init_builtins): Setup long double, _Decimal64, and
699 _Decimal128 types for new builtin functions.
700 (builtin_function_type): Set the unsigned flags appropriately for
701 the new builtin functions.
702 (rs6000_opt_masks): Add support for decimal floating point builtin
705 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
706 floating point builtin functions.
707 (RS6000_BTM_COMMON): Likewise.
708 (RS6000_BTI_long_double): Likewise.
709 (RS6000_BTI_dfloat64): Likewise.
710 (RS6000_BTI_dfloat128): Likewise.
711 (long_double_type_internal_node): Likewise.
712 (dfloat64_type_internal_node): Likewise.
713 (dfloat128_type_internal_node): Likewise.
715 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
716 2.07 bcd arithmetic instructions.
717 (UNSPEC_BCDSUB): Likewise.
718 (UNSPEC_BCD_OVERFLOW): Likewise.
719 (UNSPEC_BCD_ADD_SUB): Likewise.
720 (bcd_add_sub): Likewise.
721 (BCD_TEST): Likewise.
722 (bcd<bcd_add_sub>): Likewise.
723 (bcd<bcd_add_sub>_test): Likewise.
724 (bcd<bcd_add_sub>_test2): Likewise.
725 (bcd<bcd_add_sub>_<code>): Likewise.
726 (peephole2 for combined bcd ops): Likewise.
728 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
729 decimal floating point builtin functions.
730 (UNSPEC_DENBCD): Likewise.
731 (UNSPEC_DXEX): Likewise.
732 (UNSPEC_DIEX): Likewise.
733 (UNSPEC_DSCLI): Likewise.
734 (UNSPEC_DSCRI): Likewise.
735 (D64_D128): Likewise.
736 (dfp_suffix): Likewise.
737 (dfp_ddedpd_<mode>): Likewise.
738 (dfp_denbcd_<mode>): Likewise.
739 (dfp_dxex_<mode>): Likewise.
740 (dfp_diex_<mode>): Likewise.
741 (dfp_dscli_<mode>): Likewise.
742 (dfp_dscri_<mode>): Likewise.
744 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
746 (UNSPEC_CDTBCD): Likewise.
747 (UNSPEC_CBCDTD): Likewise.
748 (UNSPEC_DIVE): Add support for new extended divide builtin
750 (UNSPEC_DIVEO): Likewise.
751 (UNSPEC_DIVEU): Likewise.
752 (UNSPEC_DIVEUO): Likewise.
753 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
754 pack/unpack 128-bit types.
755 (UNSPEC_PACK_128BIT): Likewise.
756 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
757 (udiv<mode>3): Use idiv_ldiv mode attribute.
758 (div<mode>3): Likewise.
759 (addg6s): Add new BCD builtin functions.
762 (UNSPEC_DIV_EXTEND): Add support for new extended divide
764 (div_extend): Likewise.
765 (div<div_extend>_<mode>"): Likewise.
766 (FP128_64): Add support for new builtin functions to pack/unpack
768 (unpack<mode>): Likewise.
769 (unpacktf_0): Likewise.
770 (unpacktf_1): Likewise.
771 (unpack<mode>_dm): Likewise.
772 (unpack<mode>_nodm): Likewise.
773 (pack<mode>): Likewise.
774 (unpackv1ti): Likewise.
775 (packv1ti): Likewise.
777 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
779 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
782 2014-04-24 Jakub Jelinek <jakub@redhat.com>
784 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
785 * gimplify.c (omp_is_private): Change last argument's type to int.
786 Only diagnose lastprivate if the simd argument is 1, only diagnose
787 linear if the simd argument is 2.
788 (gimplify_omp_for): Adjust omp_is_private callers. When adding
789 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
790 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
791 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
792 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
793 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
794 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
795 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
796 * tree-nested.c (convert_nonlocal_omp_clauses,
797 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
799 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
802 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
805 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
807 * flag-types.h (enum ivar_visibility): Add.
809 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
811 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
814 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
816 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
818 2014-04-24 Radovan Obradovic <robradovic@mips.com>
819 Tom de Vries <tom@codesourcery.com>
821 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
822 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
824 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
825 * emit-rtl.c (try_split): Same.
827 2014-04-24 Radovan Obradovic <robradovic@mips.com>
828 Tom de Vries <tom@codesourcery.com>
830 * common.opt (fuse-caller-save): New option.
832 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
834 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
837 2014-04-24 Richard Biener <rguenther@suse.de>
839 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
840 during TER and instead use the sepops interface for expanding
841 non-GIMPLE_SINGLE_RHS.
843 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
845 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
846 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
848 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
850 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
851 assembler 64-bit option.
852 * configure: Regenerate.
854 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
856 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
857 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
858 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
859 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
860 (TARGET_CRYPTO): Take TARGET_SIMD into account.
862 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
864 * config/aarch64/aarch64-builtins.c
865 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
866 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
867 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
868 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
870 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
871 (Vrevsuff): New mode attribute.
873 2014-04-24 Terry Guo <terry.guo@arm.com>
875 * config/arm/arm.h (machine_function): Define variable
876 after_arm_reorg here.
877 * config/arm/arm.c (after_arm_reorg): Remove the definition.
878 (arm_split_constant): Update the way to access variable
881 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
883 2014-04-23 Tom de Vries <tom@codesourcery.com>
885 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
887 2014-04-23 David Malcolm <dmalcolm@redhat.com>
889 * is-a.h: Update comments to reflect the following changes to the
890 "pointerness" of the API, making the template parameter match the
891 return type, allowing use of is-a.h with typedefs of pointers.
892 (is_a_helper::cast): Return a T rather then a pointer to a T, so
893 that the return type matches the parameter to the is_a_helper.
895 (dyn_cast): Likewise.
897 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
898 pointer from the is-a.h API.
900 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
901 (is_a_helper <cgraph_node *>::test): ...this, matching change to
903 (is_a_helper <varpool_node>::test): Likewise, convert to...
904 (is_a_helper <varpool_node *>::test): ...this.
906 (varpool_first_variable): Update for removal of implicit pointer
908 (varpool_next_variable): Likewise.
909 (varpool_first_static_initializer): Likewise.
910 (varpool_next_static_initializer): Likewise.
911 (varpool_first_defined_variable): Likewise.
912 (varpool_next_defined_variable): Likewise.
913 (cgraph_first_defined_function): Likewise.
914 (cgraph_next_defined_function): Likewise.
915 (cgraph_first_function): Likewise.
916 (cgraph_next_function): Likewise.
917 (cgraph_first_function_with_gimple_body): Likewise.
918 (cgraph_next_function_with_gimple_body): Likewise.
919 (cgraph_alias_target): Likewise.
920 (varpool_alias_target): Likewise.
921 (cgraph_function_or_thunk_node): Likewise.
922 (varpool_variable_node): Likewise.
923 (symtab_real_symbol_p): Likewise.
924 * cgraphunit.c (referred_to_p): Likewise.
925 (analyze_functions): Likewise.
926 (handle_alias_pairs): Likewise.
927 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
928 * gimple-ssa.h (gimple_vuse_op): Likewise.
929 (gimple_vdef_op): Likewise.
930 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
931 * gimple.c (gimple_build_asm_1): Likewise.
932 (gimple_build_try): Likewise.
933 (gimple_build_resx): Likewise.
934 (gimple_build_eh_dispatch): Likewise.
935 (gimple_build_omp_for): Likewise.
936 (gimple_omp_for_set_clauses): Likewise.
938 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
939 (is_a_helper <gimple_statement_asm *>::test): ...this.
940 (is_a_helper <gimple_statement_bind>::test): Convert to...
941 (is_a_helper <gimple_statement_bind *>::test): ...this.
942 (is_a_helper <gimple_statement_call>::test): Convert to...
943 (is_a_helper <gimple_statement_call *>::test): ...this.
944 (is_a_helper <gimple_statement_catch>::test): Convert to...
945 (is_a_helper <gimple_statement_catch *>::test): ...this.
946 (is_a_helper <gimple_statement_resx>::test): Convert to...
947 (is_a_helper <gimple_statement_resx *>::test): ...this.
948 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
949 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
950 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
951 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
952 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
953 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
954 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
955 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
956 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
957 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
958 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
959 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
960 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
961 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
962 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
963 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
964 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
965 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
966 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
967 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
968 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
969 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
970 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
971 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
972 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
973 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
974 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
975 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
976 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
977 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
978 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
979 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
980 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
981 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
982 (is_a_helper <gimple_statement_phi>::test): Convert to...
983 (is_a_helper <gimple_statement_phi *>::test): ...this.
984 (is_a_helper <gimple_statement_transaction>::test): Convert to...
985 (is_a_helper <gimple_statement_transaction *>::test): ...this.
986 (is_a_helper <gimple_statement_try>::test): Convert to...
987 (is_a_helper <gimple_statement_try *>::test): ...this.
988 (is_a_helper <gimple_statement_wce>::test): Convert to...
989 (is_a_helper <gimple_statement_wce *>::test): ...this.
990 (is_a_helper <const gimple_statement_asm>::test): Convert to...
991 (is_a_helper <const gimple_statement_asm *>::test): ...this.
992 (is_a_helper <const gimple_statement_bind>::test): Convert to...
993 (is_a_helper <const gimple_statement_bind *>::test): ...this.
994 (is_a_helper <const gimple_statement_call>::test): Convert to...
995 (is_a_helper <const gimple_statement_call *>::test): ...this.
996 (is_a_helper <const gimple_statement_catch>::test): Convert to...
997 (is_a_helper <const gimple_statement_catch *>::test): ...this.
998 (is_a_helper <const gimple_statement_resx>::test): Convert to...
999 (is_a_helper <const gimple_statement_resx *>::test): ...this.
1000 (is_a_helper <const gimple_statement_eh_dispatch>::test):
1002 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
1004 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
1006 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
1007 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
1009 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
1011 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
1013 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
1015 (is_a_helper <const gimple_statement_omp_return>::test): Convert
1017 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
1018 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
1020 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
1021 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
1023 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
1024 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
1025 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
1026 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
1028 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
1029 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
1031 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
1032 (is_a_helper <const gimple_statement_omp_target>::test): Convert
1034 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
1035 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
1037 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
1038 (is_a_helper <const gimple_statement_omp_single>::test): Convert
1040 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
1041 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
1043 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
1044 (is_a_helper <const gimple_statement_omp_task>::test): Convert
1046 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
1047 (is_a_helper <const gimple_statement_phi>::test): Convert to...
1048 (is_a_helper <const gimple_statement_phi *>::test): ...this.
1049 (is_a_helper <const gimple_statement_transaction>::test): Convert
1051 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
1052 (is_a_helper <const gimple_statement_with_ops>::test): Convert
1054 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
1055 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
1056 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
1057 (is_a_helper <const gimple_statement_with_memory_ops>::test):
1059 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
1061 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
1063 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
1065 (gimple_use_ops): Update for removal of implicit pointer from the
1067 (gimple_set_use_ops): Likewise.
1068 (gimple_vuse): Likewise.
1069 (gimple_vdef): Likewise.
1070 (gimple_vuse_ptr): Likewise.
1071 (gimple_vdef_ptr): Likewise.
1072 (gimple_set_vuse): Likewise.
1073 (gimple_set_vdef): Likewise.
1074 (gimple_omp_return_set_lhs): Likewise.
1075 (gimple_omp_return_lhs): Likewise.
1076 (gimple_omp_return_lhs_ptr): Likewise.
1077 (gimple_call_fntype): Likewise.
1078 (gimple_call_set_fntype): Likewise.
1079 (gimple_call_set_internal_fn): Likewise.
1080 (gimple_call_use_set): Likewise.
1081 (gimple_call_clobber_set): Likewise.
1082 (gimple_bind_vars): Likewise.
1083 (gimple_bind_set_vars): Likewise.
1084 (gimple_bind_body_ptr): Likewise.
1085 (gimple_bind_set_body): Likewise.
1086 (gimple_bind_add_stmt): Likewise.
1087 (gimple_bind_block): Likewise.
1088 (gimple_bind_set_block): Likewise.
1089 (gimple_asm_ninputs): Likewise.
1090 (gimple_asm_noutputs): Likewise.
1091 (gimple_asm_nclobbers): Likewise.
1092 (gimple_asm_nlabels): Likewise.
1093 (gimple_asm_input_op): Likewise.
1094 (gimple_asm_input_op_ptr): Likewise.
1095 (gimple_asm_output_op): Likewise.
1096 (gimple_asm_output_op_ptr): Likewise.
1097 (gimple_asm_set_output_op): Likewise.
1098 (gimple_asm_clobber_op): Likewise.
1099 (gimple_asm_set_clobber_op): Likewise.
1100 (gimple_asm_label_op): Likewise.
1101 (gimple_asm_set_label_op): Likewise.
1102 (gimple_asm_string): Likewise.
1103 (gimple_catch_types): Likewise.
1104 (gimple_catch_types_ptr): Likewise.
1105 (gimple_catch_handler_ptr): Likewise.
1106 (gimple_catch_set_types): Likewise.
1107 (gimple_catch_set_handler): Likewise.
1108 (gimple_eh_filter_types): Likewise.
1109 (gimple_eh_filter_types_ptr): Likewise.
1110 (gimple_eh_filter_failure_ptr): Likewise.
1111 (gimple_eh_filter_set_types): Likewise.
1112 (gimple_eh_filter_set_failure): Likewise.
1113 (gimple_eh_must_not_throw_fndecl): Likewise.
1114 (gimple_eh_must_not_throw_set_fndecl): Likewise.
1115 (gimple_eh_else_n_body_ptr): Likewise.
1116 (gimple_eh_else_e_body_ptr): Likewise.
1117 (gimple_eh_else_set_n_body): Likewise.
1118 (gimple_eh_else_set_e_body): Likewise.
1119 (gimple_try_eval_ptr): Likewise.
1120 (gimple_try_cleanup_ptr): Likewise.
1121 (gimple_try_set_eval): Likewise.
1122 (gimple_try_set_cleanup): Likewise.
1123 (gimple_wce_cleanup_ptr): Likewise.
1124 (gimple_wce_set_cleanup): Likewise.
1125 (gimple_phi_capacity): Likewise.
1126 (gimple_phi_num_args): Likewise.
1127 (gimple_phi_result): Likewise.
1128 (gimple_phi_result_ptr): Likewise.
1129 (gimple_phi_set_result): Likewise.
1130 (gimple_phi_arg): Likewise.
1131 (gimple_phi_set_arg): Likewise.
1132 (gimple_resx_region): Likewise.
1133 (gimple_resx_set_region): Likewise.
1134 (gimple_eh_dispatch_region): Likewise.
1135 (gimple_eh_dispatch_set_region): Likewise.
1136 (gimple_omp_critical_name): Likewise.
1137 (gimple_omp_critical_name_ptr): Likewise.
1138 (gimple_omp_critical_set_name): Likewise.
1139 (gimple_omp_for_clauses): Likewise.
1140 (gimple_omp_for_clauses_ptr): Likewise.
1141 (gimple_omp_for_set_clauses): Likewise.
1142 (gimple_omp_for_collapse): Likewise.
1143 (gimple_omp_for_index): Likewise.
1144 (gimple_omp_for_index_ptr): Likewise.
1145 (gimple_omp_for_set_index): Likewise.
1146 (gimple_omp_for_initial): Likewise.
1147 (gimple_omp_for_initial_ptr): Likewise.
1148 (gimple_omp_for_set_initial): Likewise.
1149 (gimple_omp_for_final): Likewise.
1150 (gimple_omp_for_final_ptr): Likewise.
1151 (gimple_omp_for_set_final): Likewise.
1152 (gimple_omp_for_incr): Likewise.
1153 (gimple_omp_for_incr_ptr): Likewise.
1154 (gimple_omp_for_set_incr): Likewise.
1155 (gimple_omp_for_pre_body_ptr): Likewise.
1156 (gimple_omp_for_set_pre_body): Likewise.
1157 (gimple_omp_parallel_clauses): Likewise.
1158 (gimple_omp_parallel_clauses_ptr): Likewise.
1159 (gimple_omp_parallel_set_clauses): Likewise.
1160 (gimple_omp_parallel_child_fn): Likewise.
1161 (gimple_omp_parallel_child_fn_ptr): Likewise.
1162 (gimple_omp_parallel_set_child_fn): Likewise.
1163 (gimple_omp_parallel_data_arg): Likewise.
1164 (gimple_omp_parallel_data_arg_ptr): Likewise.
1165 (gimple_omp_parallel_set_data_arg): Likewise.
1166 (gimple_omp_task_clauses): Likewise.
1167 (gimple_omp_task_clauses_ptr): Likewise.
1168 (gimple_omp_task_set_clauses): Likewise.
1169 (gimple_omp_task_child_fn): Likewise.
1170 (gimple_omp_task_child_fn_ptr): Likewise.
1171 (gimple_omp_task_set_child_fn): Likewise.
1172 (gimple_omp_task_data_arg): Likewise.
1173 (gimple_omp_task_data_arg_ptr): Likewise.
1174 (gimple_omp_task_set_data_arg): Likewise.
1175 (gimple_omp_taskreg_clauses): Likewise.
1176 (gimple_omp_taskreg_clauses_ptr): Likewise.
1177 (gimple_omp_taskreg_set_clauses): Likewise.
1178 (gimple_omp_taskreg_child_fn): Likewise.
1179 (gimple_omp_taskreg_child_fn_ptr): Likewise.
1180 (gimple_omp_taskreg_set_child_fn): Likewise.
1181 (gimple_omp_taskreg_data_arg): Likewise.
1182 (gimple_omp_taskreg_data_arg_ptr): Likewise.
1183 (gimple_omp_taskreg_set_data_arg): Likewise.
1184 (gimple_omp_task_copy_fn): Likewise.
1185 (gimple_omp_task_copy_fn_ptr): Likewise.
1186 (gimple_omp_task_set_copy_fn): Likewise.
1187 (gimple_omp_task_arg_size): Likewise.
1188 (gimple_omp_task_arg_size_ptr): Likewise.
1189 (gimple_omp_task_set_arg_size): Likewise.
1190 (gimple_omp_task_arg_align): Likewise.
1191 (gimple_omp_task_arg_align_ptr): Likewise.
1192 (gimple_omp_task_set_arg_align): Likewise.
1193 (gimple_omp_single_clauses): Likewise.
1194 (gimple_omp_single_clauses_ptr): Likewise.
1195 (gimple_omp_single_set_clauses): Likewise.
1196 (gimple_omp_target_clauses): Likewise.
1197 (gimple_omp_target_clauses_ptr): Likewise.
1198 (gimple_omp_target_set_clauses): Likewise.
1199 (gimple_omp_target_child_fn): Likewise.
1200 (gimple_omp_target_child_fn_ptr): Likewise.
1201 (gimple_omp_target_set_child_fn): Likewise.
1202 (gimple_omp_target_data_arg): Likewise.
1203 (gimple_omp_target_data_arg_ptr): Likewise.
1204 (gimple_omp_target_set_data_arg): Likewise.
1205 (gimple_omp_teams_clauses): Likewise.
1206 (gimple_omp_teams_clauses_ptr): Likewise.
1207 (gimple_omp_teams_set_clauses): Likewise.
1208 (gimple_omp_sections_clauses): Likewise.
1209 (gimple_omp_sections_clauses_ptr): Likewise.
1210 (gimple_omp_sections_set_clauses): Likewise.
1211 (gimple_omp_sections_control): Likewise.
1212 (gimple_omp_sections_control_ptr): Likewise.
1213 (gimple_omp_sections_set_control): Likewise.
1214 (gimple_omp_for_set_cond): Likewise.
1215 (gimple_omp_for_cond): Likewise.
1216 (gimple_omp_atomic_store_set_val): Likewise.
1217 (gimple_omp_atomic_store_val): Likewise.
1218 (gimple_omp_atomic_store_val_ptr): Likewise.
1219 (gimple_omp_atomic_load_set_lhs): Likewise.
1220 (gimple_omp_atomic_load_lhs): Likewise.
1221 (gimple_omp_atomic_load_lhs_ptr): Likewise.
1222 (gimple_omp_atomic_load_set_rhs): Likewise.
1223 (gimple_omp_atomic_load_rhs): Likewise.
1224 (gimple_omp_atomic_load_rhs_ptr): Likewise.
1225 (gimple_omp_continue_control_def): Likewise.
1226 (gimple_omp_continue_control_def_ptr): Likewise.
1227 (gimple_omp_continue_set_control_def): Likewise.
1228 (gimple_omp_continue_control_use): Likewise.
1229 (gimple_omp_continue_control_use_ptr): Likewise.
1230 (gimple_omp_continue_set_control_use): Likewise.
1231 (gimple_transaction_body_ptr): Likewise.
1232 (gimple_transaction_label): Likewise.
1233 (gimple_transaction_label_ptr): Likewise.
1234 (gimple_transaction_set_body): Likewise.
1235 (gimple_transaction_set_label): Likewise.
1237 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
1238 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1239 * ipa-ref.c (ipa_record_reference): Likewise.
1240 * ipa-reference.c (analyze_function): Likewise.
1241 (ipa_reference_write_optimization_summary): Likewise.
1242 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
1243 (address_taken_from_non_vtable_p): Likewise.
1244 (comdat_can_be_unshared_p_1): Likewise.
1245 * lto-cgraph.c (lto_output_ref): Likewise.
1246 (add_references): Likewise.
1247 (compute_ltrans_boundary): Likewise.
1248 (output_symtab): Likewise.
1249 (input_ref): Likewise.
1250 (input_cgraph_1): Likewise.
1251 (output_cgraph_opt_summary): Likewise.
1252 * lto-streamer-out.c (lto_output): Likewise.
1253 (output_symbol_p): Likewise.
1254 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
1255 (lsei_start_function_in_partition): Likewise.
1256 (lsei_next_variable_in_partition): Likewise.
1257 (lsei_start_variable_in_partition): Likewise.
1258 * symtab.c (insert_to_assembler_name_hash): Likewise.
1259 (unlink_from_assembler_name_hash): Likewise.
1260 (symtab_unregister_node): Likewise.
1261 (symtab_remove_node): Likewise.
1262 (dump_symtab_node): Likewise.
1263 (verify_symtab_base): Likewise.
1264 (verify_symtab_node): Likewise.
1265 (symtab_make_decl_local): Likewise.
1266 (symtab_alias_ultimate_target): Likewise.
1267 (symtab_resolve_alias): Likewise.
1268 (symtab_get_symbol_partitioning_class): Likewise.
1269 * tree-phinodes.c (allocate_phi_node): Likewise.
1270 (reserve_phi_args_for_new_edge): Likewise.
1271 (remove_phi_args): Likewise.
1272 * varpool.c (varpool_node_for_asm): Likewise.
1273 (varpool_remove_unreferenced_decls): Likewise.
1275 2014-04-23 Jeff Law <law@redhat.com>
1277 PR tree-optimization/60902
1278 * tree-ssa-threadedge.c
1279 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
1280 invalidate outputs from statements that do not produce useful
1281 outputs for threading.
1283 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
1285 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
1286 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
1287 machine descriptions for Stack Smashing Protector.
1289 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
1291 * aarch64.md (<optab>_rol<mode>3): New pattern.
1292 (<optab>_rolsi3_uxtw): Likewise.
1293 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
1295 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
1297 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
1298 (arm_cortex_a12_tune): Likewise.
1300 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1302 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
1304 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1306 * config/arm/arm.md (arm_rev16si2): New pattern.
1307 (arm_rev16si2_alt): Likewise.
1308 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
1310 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1312 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
1313 (rev16<mode>2_alt): Likewise.
1314 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
1315 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
1316 (aarch_rev16_shleft_mask_imm_p): Likewise.
1317 (aarch_rev16_p_1): Likewise.
1318 (aarch_rev16_p): Likewise.
1319 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
1320 (aarch_rev16_shright_mask_imm_p): Likewise.
1321 (aarch_rev16_shleft_mask_imm_p): Likewise.
1323 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1325 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
1326 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
1328 (cortex_a53_extra_costs): Likewise.
1329 (cortex_a57_extra_costs): Likewise.
1330 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
1331 (cortexa7_extra_costs): Likewise.
1332 (cortexa8_extra_costs): Likewise.
1333 (cortexa12_extra_costs): Likewise.
1334 (cortexa15_extra_costs): Likewise.
1335 (v7m_extra_costs): Likewise.
1336 (arm_new_rtx_costs): Handle BSWAP.
1338 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1340 * config/arm/arm.c (cortexa8_extra_costs): New table.
1341 (arm_cortex_a8_tune): New tuning struct.
1342 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
1344 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1346 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
1348 2014-04-23 Richard Biener <rguenther@suse.de>
1350 * Makefile.in (OBJS): Remove loop-unswitch.o.
1351 * loop-unswitch.c: Delete.
1352 * tree-pass.h (make_pass_rtl_unswitch): Remove.
1353 * passes.def (pass_rtl_unswitch): Likewise.
1354 * loop-init.c (gate_rtl_unswitch): Likewise.
1355 (rtl_unswitch): Likewise.
1356 (pass_data_rtl_unswitch): Likewise.
1357 (pass_rtl_unswitch): Likewise.
1358 (make_pass_rtl_unswitch): Likewise.
1359 * rtl.h (reversed_condition): Likewise.
1360 (compare_and_jump_seq): Likewise.
1361 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
1363 * loop-unroll.c (compare_and_jump_seq): Likewise.
1365 2014-04-23 Richard Biener <rguenther@suse.de>
1367 PR tree-optimization/60903
1368 * tree-ssa-loop-im.c (analyze_memory_references): Remove
1369 commented code block.
1370 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
1371 loop flags to newly created BBs and edges.
1373 2014-04-23 Nick Clifton <nickc@redhat.com>
1375 * config/msp430/msp430.c (msp430_handle_option): Move function
1377 (msp430_option_override): Simplify mcu and mcpu option handling.
1378 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
1379 support for -mhwmult command line option.
1380 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
1381 -mhwmult command line option.
1382 (msp430_hwmult_enabled): Delete.
1383 (msp43o_output_labelref): Add support for -mhwmult command line
1385 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
1386 (umulsidi3): Likewise.
1387 * config/msp430/msp430.opt (mmcu): Add Report attribute.
1388 (mcpu, mlarge, msmall): Likewise.
1389 (mhwmult): New option.
1390 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
1392 (msp430_is_f5_mcu): Remove prototype.
1393 (msp430_use_f5_series_hwmult): Add prototype.
1394 * config/msp430/msp430-opts.h: New file.
1395 * common/config/msp430: New directory.
1396 * common/config/msp430/msp430-common.c: New file.
1397 * config.gcc (msp430): Remove target_has_targetm_common.
1398 * doc/invoke.texi: Document -mhwmult command line option.
1400 2014-04-23 Nick Clifton <nickc@redhat.com>
1402 * config/i386/cygwin.h (ENDFILE_SPEC): Include
1403 default-manifest.o if it can be found in the search path.
1404 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
1406 2014-04-23 Terry Guo <terry.guo@arm.com>
1408 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
1410 2014-04-23 Richard Biener <rguenther@suse.de>
1413 * tree-inline.c (declare_return_variable): Use mark_addressable.
1415 2014-04-23 Richard Biener <rguenther@suse.de>
1418 * loop-init.c (loop_optimizer_init): Make sure to apply
1419 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
1421 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1424 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
1426 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
1427 if flag_sanitize_undefined_trap_on_error.
1428 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
1429 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
1430 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
1431 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
1432 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
1433 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
1434 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
1435 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
1436 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
1437 * ubsan.c (ubsan_instrument_unreachable): Return
1438 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
1439 (ubsan_expand_null_ifn): Emit __builtin_trap ()
1440 if flag_sanitize_undefined_trap_on_error and
1441 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
1442 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
1443 instrument_bool_enum_load): Emit __builtin_trap () if
1444 flag_sanitize_undefined_trap_on_error and
1445 __builtin_handle_*_abort () if !flag_sanitize_recover.
1446 * doc/invoke.texi (-fsanitize-recover,
1447 -fsanitize-undefined-trap-on-error): Document.
1449 2014-04-22 Christian Bruel <christian.bruel@st.com>
1451 * config/sh/sh.md (mov<mode>): Replace movQIHI.
1452 Force immediates to SImode.
1454 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
1456 * config/nios2/nios2.md (UNSPEC_ROUND): New.
1458 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
1459 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
1460 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
1461 (nios2_fpu_insn): Add entry for round.
1462 (N2FPU_NO_ERRNO_P): Define.
1463 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
1465 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
1467 2014-04-22 Richard Henderson <rth@redhat.com>
1469 * config/aarch64/aarch64 (addti3, subti3): New expanders.
1470 (add<GPI>3_compare0): Remove leading * from name.
1471 (add<GPI>3_carryin): Likewise.
1472 (sub<GPI>3_compare0): Likewise.
1473 (sub<GPI>3_carryin): Likewise.
1474 (<su_optab>mulditi3): New expander.
1475 (multi3): New expander.
1476 (madd<GPI>): Remove leading * from name.
1478 2014-04-22 Martin Jambor <mjambor@suse.cz>
1480 * cgraphclones.c (cgraph_function_versioning): Copy
1481 ipa_transforms_to_apply instead of asserting it is empty.
1483 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
1486 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
1487 on count_exp to get mode.
1489 2014-04-22 Andrew Pinski <apinski@cavium.com>
1491 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
1492 Handle TLS for ILP32.
1493 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
1494 (tlsie_small_<mode>): this and handle PTR.
1495 (tlsie_small_sidi): New pattern.
1496 (tlsle_small): Change to an expand to handle ILP32.
1497 (tlsle_small_<mode>): New pattern.
1498 (tlsdesc_small): Rename to ...
1499 (tlsdesc_small_<mode>): this and handle PTR.
1501 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1503 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
1505 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1507 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1508 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
1509 (aarch64_types_signed_poly_qualifiers): Likewise.
1510 (aarch64_types_unsigned_signed_qualifiers): Likewise.
1511 (aarch64_types_poly_signed_qualifiers): Likewise.
1512 (TYPES_REINTERP_SS): Type macro added.
1513 (TYPES_REINTERP_SU): Likewise.
1514 (TYPES_REINTERP_SP): Likewise.
1515 (TYPES_REINTERP_US): Likewise.
1516 (TYPES_REINTERP_PS): Likewise.
1517 (aarch64_fold_builtin): New expression folding added.
1518 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
1519 Declarations removed.
1520 (REINTERP_SS): Declarations added.
1521 (REINTERP_US): Likewise.
1522 (REINTERP_PS): Likewise.
1523 (REINTERP_SU): Likewise.
1524 (REINTERP_SP): Likewise.
1525 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
1526 (vreinterpretq_p8_f64): Likewise.
1527 (vreinterpret_p16_f64): Likewise.
1528 (vreinterpretq_p16_f64): Likewise.
1529 (vreinterpret_f32_f64): Likewise.
1530 (vreinterpretq_f32_f64): Likewise.
1531 (vreinterpret_f64_f32): Likewise.
1532 (vreinterpret_f64_p8): Likewise.
1533 (vreinterpret_f64_p16): Likewise.
1534 (vreinterpret_f64_s8): Likewise.
1535 (vreinterpret_f64_s16): Likewise.
1536 (vreinterpret_f64_s32): Likewise.
1537 (vreinterpret_f64_s64): Likewise.
1538 (vreinterpret_f64_u8): Likewise.
1539 (vreinterpret_f64_u16): Likewise.
1540 (vreinterpret_f64_u32): Likewise.
1541 (vreinterpret_f64_u64): Likewise.
1542 (vreinterpretq_f64_f32): Likewise.
1543 (vreinterpretq_f64_p8): Likewise.
1544 (vreinterpretq_f64_p16): Likewise.
1545 (vreinterpretq_f64_s8): Likewise.
1546 (vreinterpretq_f64_s16): Likewise.
1547 (vreinterpretq_f64_s32): Likewise.
1548 (vreinterpretq_f64_s64): Likewise.
1549 (vreinterpretq_f64_u8): Likewise.
1550 (vreinterpretq_f64_u16): Likewise.
1551 (vreinterpretq_f64_u32): Likewise.
1552 (vreinterpretq_f64_u64): Likewise.
1553 (vreinterpret_s64_f64): Likewise.
1554 (vreinterpretq_s64_f64): Likewise.
1555 (vreinterpret_u64_f64): Likewise.
1556 (vreinterpretq_u64_f64): Likewise.
1557 (vreinterpret_s8_f64): Likewise.
1558 (vreinterpretq_s8_f64): Likewise.
1559 (vreinterpret_s16_f64): Likewise.
1560 (vreinterpretq_s16_f64): Likewise.
1561 (vreinterpret_s32_f64): Likewise.
1562 (vreinterpretq_s32_f64): Likewise.
1563 (vreinterpret_u8_f64): Likewise.
1564 (vreinterpretq_u8_f64): Likewise.
1565 (vreinterpret_u16_f64): Likewise.
1566 (vreinterpretq_u16_f64): Likewise.
1567 (vreinterpret_u32_f64): Likewise.
1568 (vreinterpretq_u32_f64): Likewise.
1570 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1572 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1573 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
1574 (vreinterpret_p8_s8): Likewise.
1575 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
1576 (vreinterpret_p8_s16): Likewise.
1577 (vreinterpret_p8_s32): Likewise.
1578 (vreinterpret_p8_s64): Likewise.
1579 (vreinterpret_p8_f32): Likewise.
1580 (vreinterpret_p8_u8): Likewise.
1581 (vreinterpret_p8_u16): Likewise.
1582 (vreinterpret_p8_u32): Likewise.
1583 (vreinterpret_p8_u64): Likewise.
1584 (vreinterpret_p8_p16): Likewise.
1585 (vreinterpretq_p8_s8): Likewise.
1586 (vreinterpretq_p8_s16): Likewise.
1587 (vreinterpretq_p8_s32): Likewise.
1588 (vreinterpretq_p8_s64): Likewise.
1589 (vreinterpretq_p8_f32): Likewise.
1590 (vreinterpretq_p8_u8): Likewise.
1591 (vreinterpretq_p8_u16): Likewise.
1592 (vreinterpretq_p8_u32): Likewise.
1593 (vreinterpretq_p8_u64): Likewise.
1594 (vreinterpretq_p8_p16): Likewise.
1595 (vreinterpret_p16_s8): Likewise.
1596 (vreinterpret_p16_s16): Likewise.
1597 (vreinterpret_p16_s32): Likewise.
1598 (vreinterpret_p16_s64): Likewise.
1599 (vreinterpret_p16_f32): Likewise.
1600 (vreinterpret_p16_u8): Likewise.
1601 (vreinterpret_p16_u16): Likewise.
1602 (vreinterpret_p16_u32): Likewise.
1603 (vreinterpret_p16_u64): Likewise.
1604 (vreinterpret_p16_p8): Likewise.
1605 (vreinterpretq_p16_s8): Likewise.
1606 (vreinterpretq_p16_s16): Likewise.
1607 (vreinterpretq_p16_s32): Likewise.
1608 (vreinterpretq_p16_s64): Likewise.
1609 (vreinterpretq_p16_f32): Likewise.
1610 (vreinterpretq_p16_u8): Likewise.
1611 (vreinterpretq_p16_u16): Likewise.
1612 (vreinterpretq_p16_u32): Likewise.
1613 (vreinterpretq_p16_u64): Likewise.
1614 (vreinterpretq_p16_p8): Likewise.
1615 (vreinterpret_f32_s8): Likewise.
1616 (vreinterpret_f32_s16): Likewise.
1617 (vreinterpret_f32_s32): Likewise.
1618 (vreinterpret_f32_s64): Likewise.
1619 (vreinterpret_f32_u8): Likewise.
1620 (vreinterpret_f32_u16): Likewise.
1621 (vreinterpret_f32_u32): Likewise.
1622 (vreinterpret_f32_u64): Likewise.
1623 (vreinterpret_f32_p8): Likewise.
1624 (vreinterpret_f32_p16): Likewise.
1625 (vreinterpretq_f32_s8): Likewise.
1626 (vreinterpretq_f32_s16): Likewise.
1627 (vreinterpretq_f32_s32): Likewise.
1628 (vreinterpretq_f32_s64): Likewise.
1629 (vreinterpretq_f32_u8): Likewise.
1630 (vreinterpretq_f32_u16): Likewise.
1631 (vreinterpretq_f32_u32): Likewise.
1632 (vreinterpretq_f32_u64): Likewise.
1633 (vreinterpretq_f32_p8): Likewise.
1634 (vreinterpretq_f32_p16): Likewise.
1635 (vreinterpret_s64_s8): Likewise.
1636 (vreinterpret_s64_s16): Likewise.
1637 (vreinterpret_s64_s32): Likewise.
1638 (vreinterpret_s64_f32): Likewise.
1639 (vreinterpret_s64_u8): Likewise.
1640 (vreinterpret_s64_u16): Likewise.
1641 (vreinterpret_s64_u32): Likewise.
1642 (vreinterpret_s64_u64): Likewise.
1643 (vreinterpret_s64_p8): Likewise.
1644 (vreinterpret_s64_p16): Likewise.
1645 (vreinterpretq_s64_s8): Likewise.
1646 (vreinterpretq_s64_s16): Likewise.
1647 (vreinterpretq_s64_s32): Likewise.
1648 (vreinterpretq_s64_f32): Likewise.
1649 (vreinterpretq_s64_u8): Likewise.
1650 (vreinterpretq_s64_u16): Likewise.
1651 (vreinterpretq_s64_u32): Likewise.
1652 (vreinterpretq_s64_u64): Likewise.
1653 (vreinterpretq_s64_p8): Likewise.
1654 (vreinterpretq_s64_p16): Likewise.
1655 (vreinterpret_u64_s8): Likewise.
1656 (vreinterpret_u64_s16): Likewise.
1657 (vreinterpret_u64_s32): Likewise.
1658 (vreinterpret_u64_s64): Likewise.
1659 (vreinterpret_u64_f32): Likewise.
1660 (vreinterpret_u64_u8): Likewise.
1661 (vreinterpret_u64_u16): Likewise.
1662 (vreinterpret_u64_u32): Likewise.
1663 (vreinterpret_u64_p8): Likewise.
1664 (vreinterpret_u64_p16): Likewise.
1665 (vreinterpretq_u64_s8): Likewise.
1666 (vreinterpretq_u64_s16): Likewise.
1667 (vreinterpretq_u64_s32): Likewise.
1668 (vreinterpretq_u64_s64): Likewise.
1669 (vreinterpretq_u64_f32): Likewise.
1670 (vreinterpretq_u64_u8): Likewise.
1671 (vreinterpretq_u64_u16): Likewise.
1672 (vreinterpretq_u64_u32): Likewise.
1673 (vreinterpretq_u64_p8): Likewise.
1674 (vreinterpretq_u64_p16): Likewise.
1675 (vreinterpret_s8_s16): Likewise.
1676 (vreinterpret_s8_s32): Likewise.
1677 (vreinterpret_s8_s64): Likewise.
1678 (vreinterpret_s8_f32): Likewise.
1679 (vreinterpret_s8_u8): Likewise.
1680 (vreinterpret_s8_u16): Likewise.
1681 (vreinterpret_s8_u32): Likewise.
1682 (vreinterpret_s8_u64): Likewise.
1683 (vreinterpret_s8_p8): Likewise.
1684 (vreinterpret_s8_p16): Likewise.
1685 (vreinterpretq_s8_s16): Likewise.
1686 (vreinterpretq_s8_s32): Likewise.
1687 (vreinterpretq_s8_s64): Likewise.
1688 (vreinterpretq_s8_f32): Likewise.
1689 (vreinterpretq_s8_u8): Likewise.
1690 (vreinterpretq_s8_u16): Likewise.
1691 (vreinterpretq_s8_u32): Likewise.
1692 (vreinterpretq_s8_u64): Likewise.
1693 (vreinterpretq_s8_p8): Likewise.
1694 (vreinterpretq_s8_p16): Likewise.
1695 (vreinterpret_s16_s8): Likewise.
1696 (vreinterpret_s16_s32): Likewise.
1697 (vreinterpret_s16_s64): Likewise.
1698 (vreinterpret_s16_f32): Likewise.
1699 (vreinterpret_s16_u8): Likewise.
1700 (vreinterpret_s16_u16): Likewise.
1701 (vreinterpret_s16_u32): Likewise.
1702 (vreinterpret_s16_u64): Likewise.
1703 (vreinterpret_s16_p8): Likewise.
1704 (vreinterpret_s16_p16): Likewise.
1705 (vreinterpretq_s16_s8): Likewise.
1706 (vreinterpretq_s16_s32): Likewise.
1707 (vreinterpretq_s16_s64): Likewise.
1708 (vreinterpretq_s16_f32): Likewise.
1709 (vreinterpretq_s16_u8): Likewise.
1710 (vreinterpretq_s16_u16): Likewise.
1711 (vreinterpretq_s16_u32): Likewise.
1712 (vreinterpretq_s16_u64): Likewise.
1713 (vreinterpretq_s16_p8): Likewise.
1714 (vreinterpretq_s16_p16): Likewise.
1715 (vreinterpret_s32_s8): Likewise.
1716 (vreinterpret_s32_s16): Likewise.
1717 (vreinterpret_s32_s64): Likewise.
1718 (vreinterpret_s32_f32): Likewise.
1719 (vreinterpret_s32_u8): Likewise.
1720 (vreinterpret_s32_u16): Likewise.
1721 (vreinterpret_s32_u32): Likewise.
1722 (vreinterpret_s32_u64): Likewise.
1723 (vreinterpret_s32_p8): Likewise.
1724 (vreinterpret_s32_p16): Likewise.
1725 (vreinterpretq_s32_s8): Likewise.
1726 (vreinterpretq_s32_s16): Likewise.
1727 (vreinterpretq_s32_s64): Likewise.
1728 (vreinterpretq_s32_f32): Likewise.
1729 (vreinterpretq_s32_u8): Likewise.
1730 (vreinterpretq_s32_u16): Likewise.
1731 (vreinterpretq_s32_u32): Likewise.
1732 (vreinterpretq_s32_u64): Likewise.
1733 (vreinterpretq_s32_p8): Likewise.
1734 (vreinterpretq_s32_p16): Likewise.
1735 (vreinterpret_u8_s8): Likewise.
1736 (vreinterpret_u8_s16): Likewise.
1737 (vreinterpret_u8_s32): Likewise.
1738 (vreinterpret_u8_s64): Likewise.
1739 (vreinterpret_u8_f32): Likewise.
1740 (vreinterpret_u8_u16): Likewise.
1741 (vreinterpret_u8_u32): Likewise.
1742 (vreinterpret_u8_u64): Likewise.
1743 (vreinterpret_u8_p8): Likewise.
1744 (vreinterpret_u8_p16): Likewise.
1745 (vreinterpretq_u8_s8): Likewise.
1746 (vreinterpretq_u8_s16): Likewise.
1747 (vreinterpretq_u8_s32): Likewise.
1748 (vreinterpretq_u8_s64): Likewise.
1749 (vreinterpretq_u8_f32): Likewise.
1750 (vreinterpretq_u8_u16): Likewise.
1751 (vreinterpretq_u8_u32): Likewise.
1752 (vreinterpretq_u8_u64): Likewise.
1753 (vreinterpretq_u8_p8): Likewise.
1754 (vreinterpretq_u8_p16): Likewise.
1755 (vreinterpret_u16_s8): Likewise.
1756 (vreinterpret_u16_s16): Likewise.
1757 (vreinterpret_u16_s32): Likewise.
1758 (vreinterpret_u16_s64): Likewise.
1759 (vreinterpret_u16_f32): Likewise.
1760 (vreinterpret_u16_u8): Likewise.
1761 (vreinterpret_u16_u32): Likewise.
1762 (vreinterpret_u16_u64): Likewise.
1763 (vreinterpret_u16_p8): Likewise.
1764 (vreinterpret_u16_p16): Likewise.
1765 (vreinterpretq_u16_s8): Likewise.
1766 (vreinterpretq_u16_s16): Likewise.
1767 (vreinterpretq_u16_s32): Likewise.
1768 (vreinterpretq_u16_s64): Likewise.
1769 (vreinterpretq_u16_f32): Likewise.
1770 (vreinterpretq_u16_u8): Likewise.
1771 (vreinterpretq_u16_u32): Likewise.
1772 (vreinterpretq_u16_u64): Likewise.
1773 (vreinterpretq_u16_p8): Likewise.
1774 (vreinterpretq_u16_p16): Likewise.
1775 (vreinterpret_u32_s8): Likewise.
1776 (vreinterpret_u32_s16): Likewise.
1777 (vreinterpret_u32_s32): Likewise.
1778 (vreinterpret_u32_s64): Likewise.
1779 (vreinterpret_u32_f32): Likewise.
1780 (vreinterpret_u32_u8): Likewise.
1781 (vreinterpret_u32_u16): Likewise.
1782 (vreinterpret_u32_u64): Likewise.
1783 (vreinterpret_u32_p8): Likewise.
1784 (vreinterpret_u32_p16): Likewise.
1785 (vreinterpretq_u32_s8): Likewise.
1786 (vreinterpretq_u32_s16): Likewise.
1787 (vreinterpretq_u32_s32): Likewise.
1788 (vreinterpretq_u32_s64): Likewise.
1789 (vreinterpretq_u32_f32): Likewise.
1790 (vreinterpretq_u32_u8): Likewise.
1791 (vreinterpretq_u32_u16): Likewise.
1792 (vreinterpretq_u32_u64): Likewise.
1793 (vreinterpretq_u32_p8): Likewise.
1794 (vreinterpretq_u32_p16): Likewise.
1796 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1798 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1800 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1803 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1804 (vqnegd_s64): Likewise.
1805 (vqabs_s64): Likewise.
1806 (vqabsd_s64): Likewise.
1808 2014-04-22 Richard Henderson <rth@redhat.com>
1810 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1811 computation to the top of the loop.
1813 2014-04-22 Renlin <renlin.li@arm.com>
1814 Jiong Wang <jiong.wang@arm.com>
1816 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1817 * config/aarch64/aarch64.c (aarch64_layout_frame)
1818 (aarch64_initial_elimination_offset): Likewise.
1820 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1822 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1825 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1827 * machmode.h (bitwise_mode_for_mode): Declare.
1828 * stor-layout.h (bitwise_type_for_mode): Likewise.
1829 * stor-layout.c (bitwise_mode_for_mode): New function.
1830 (bitwise_type_for_mode): Likewise.
1831 * builtins.c (fold_builtin_memory_op): Use it instead of
1832 int_mode_for_mode and build_nonstandard_integer_type.
1834 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1836 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1837 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1838 (*-*-solaris2*): Simplify.
1839 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1840 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1841 *-*-solaris2.9* handling.
1843 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1845 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1846 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1848 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1849 * configure: Regenerate.
1851 * config/i386/sol2-9.h: Remove.
1853 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1854 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1855 Remove Solaris 9 references.
1857 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1859 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1860 (floatuns<GPI:mode><GPF:mode>2): Remove.
1861 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1862 and floatuns conversions.
1863 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1864 and floatuns conversions.
1865 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1866 (w1,w2): New mode attributes for inequal width conversions.
1868 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1870 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1871 the output asm format.
1873 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1875 * config/aarch64/aarch64-simd.md
1876 (aarch64_cm<optab>di): Always split.
1877 (*aarch64_cm<optab>di): New.
1878 (aarch64_cmtstdi): Always split.
1879 (*aarch64_cmtstdi): New.
1881 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1883 PR tree-optimization/60823
1884 * omp-low.c (ipa_simd_modify_function_body): Go through
1885 all SSA_NAMEs and for those refering to vector arguments
1886 which are going to be replaced adjust SSA_NAME_VAR and,
1887 if it is a default definition, change it into a non-default
1888 definition assigned at the beginning of function from new_decl.
1889 (ipa_simd_modify_stmt_ops): Rewritten.
1890 * tree-dfa.c (set_ssa_default_def): When removing default def,
1891 check for NULL loc instead of NULL *loc.
1893 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1895 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1896 restrictions on core registers for DImode values in Thumb2.
1898 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1900 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1901 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1903 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1905 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1906 (*iordi_notzesidi_di): Likewise.
1907 (*iordi_notsesidi_di): Likewise.
1909 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1911 * config/arm/arm-protos.h (tune_params): New struct members.
1912 * config/arm/arm.c: Initialise tune_params per processor.
1913 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1914 for speed, based on new tune_params.
1916 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1918 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1920 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1922 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1924 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1925 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1926 (vrnda_f64): Likewise.
1927 (vrndi_f64): Likewise.
1928 (vrndm_f64): Likewise.
1929 (vrndn_f64): Likewise.
1930 (vrndp_f64): Likewise.
1931 (vrndx_f64): Likewise.
1933 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1935 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1936 GET_MODE_SIZE argument is enum machine_mode.
1938 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1941 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1942 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1944 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
1947 * asan.c (asan_emit_stack_protection): Force the base to align to
1948 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1949 appropriate bits if STRICT_ALIGNMENT.
1950 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1952 (expand_used_vars): Leave a space in the stack frame for alignment
1953 if STRICT_ALIGNMENT.
1955 2014-04-21 David Malcolm <dmalcolm@redhat.com>
1957 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1959 (gimple_store_p): Likewise.
1960 (gimple_assign_load_p): Likewise.
1961 (gimple_assign_cast_p): Likewise.
1962 (gimple_clobber_p): Likewise.
1964 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1965 rather than a gimple.
1966 (gimple_assign_cast_p): Likewise.
1968 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1971 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1972 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1974 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1975 more debug information for E500 if -mdebug=reg.
1977 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
1980 * config/i386/i386.c (ix86_expand_builtin)
1981 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
1982 register for target RTX.
1983 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
1985 2014-04-18 Cong Hou <congh@google.com>
1987 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
1988 the widen-mult pattern by handling two operands with different sizes,
1989 and operands whose size is smaller than half of the result type.
1991 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1993 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
1994 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
1995 (do_estimate_edge_time): Compute it.
1996 * ipa-inline.c (want_inline_small_function_p): Bypass
1997 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
1999 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
2001 * ipa-inline.c (spec_rem): New static variable.
2002 (dump_overall_stats): New function.
2003 (dump_inline_stats): New function.
2005 2014-04-18 Richard Henderson <rth@redhat.com>
2007 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
2008 to GET_MODE_SIZE, not a reg_class_t.
2010 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2012 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
2013 (vsx_xxmrglw_<mode>): Likewise.
2015 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
2018 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
2019 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
2020 (rs6000_init_hard_regno_mode_ok): Likewise.
2022 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
2024 * ipa-inline.c (inline_small_functions): Account only non-cold
2026 * doc/invoke.texi (inline-unit-growth): Update documentation.
2028 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
2030 * config/rs6000/rs6000.md (addti3, subti3): New.
2032 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
2035 * config/i386/i386.c (ix86_expand_clear): Remove outdated
2036 comment. Check optimize_insn_for_size_p instead of
2037 optimize_insn_for_speed_p.
2039 2014-04-17 Martin Jambor <mjambor@suse.cz>
2041 * gimple-iterator.c (gsi_start_edge): New function.
2042 * gimple-iterator.h (gsi_start_edge): Declare.
2043 * tree-sra.c (single_non_eh_succ): New function.
2044 (disqualify_ops_if_throwing_stmt): Renamed to
2045 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
2046 having one non-EH successor BB.
2047 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
2048 generate loads into replacements.
2049 (sra_modify_assign): Likewise and and also use the simple path for
2051 (sra_modify_function_body): Commit statements on edges.
2053 2014-04-17 Richard Biener <rguenther@suse.de>
2056 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
2057 comparison results and add clarifying comment.
2059 2014-04-17 Jakub Jelinek <jakub@redhat.com>
2061 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
2062 (blank_mode): Initialize it.
2063 (emit_mode_size_inline, emit_mode_nunits_inline,
2064 emit_mode_inner_inline): New functions.
2065 (emit_insn_modes_h): Call them and surround their output with
2066 #if GCC_VERSION >= 4001 ... #endif.
2067 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
2068 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
2069 mode_* arrays if the argument is __builtin_constant_p.
2070 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
2071 is enum machine_mode.
2073 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2075 * passes.c (opt_pass::execute): Adjust.
2076 (pass_manager::execute_pass_mode_switching): Likewise.
2077 (early_local_passes::execute): Likewise.
2078 (execute_one_pass): Pass cfun to the pass's execute method.
2079 * tree-pass.h (opt_pass::execute): Add function * argument.
2080 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
2081 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
2082 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
2083 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
2084 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
2085 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
2086 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
2087 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
2088 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
2089 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
2090 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
2091 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
2092 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
2093 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
2094 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
2095 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
2096 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
2097 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
2098 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
2099 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2100 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2101 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2102 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2103 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2104 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2105 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2106 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
2107 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
2108 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
2111 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2113 * passes.c (opt_pass::gate): Take function * argument.
2114 (gate_all_early_local_passes): Merge into
2115 (early_local_passes::gate): this.
2116 (gate_all_early_optimizations): Merge into
2117 (all_early_optimizations::gate): this.
2118 (gate_all_optimizations): Mege into
2119 (all_optimizations::gate): this.
2120 (gate_all_optimizations_g): Merge into
2121 (all_optimizations_g::gate): this.
2122 (gate_rest_of_compilation): Mege into
2123 (rest_of_compilation::gate): this.
2124 (gate_postreload): Merge into
2125 (postreload::gate): this.
2126 (dump_one_pass): Pass cfun to the pass's gate method.
2127 (execute_ipa_summary_passes): Likewise.
2128 (execute_one_pass): Likewise.
2129 (ipa_write_summaries_2): Likewise.
2130 (ipa_write_optimization_summaries_1): Likewise.
2131 (ipa_read_summaries_1): Likewise.
2132 (ipa_read_optimization_summaries_1): Likewise.
2133 (execute_ipa_stmt_fixups): Likewise.
2134 * tree-pass.h (opt_pass::gate): Add function * argument.
2135 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
2136 combine-stack-adj.c, combine.c, compare-elim.c,
2137 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
2138 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
2139 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
2140 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
2141 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
2142 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
2143 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
2144 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
2145 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
2146 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
2147 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
2148 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
2149 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
2150 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
2151 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
2152 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2153 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2154 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2155 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2156 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2157 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2158 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2159 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
2160 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
2161 var-tracking.c, vtable-verify.c, web.c: Adjust.
2163 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2165 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
2166 * configure: Regenerate.
2168 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2170 * passes.c (dump_one_pass): don't check pass->has_gate.
2171 (execute_ipa_summary_passes): Likewise.
2172 (execute_one_pass): Likewise.
2173 (ipa_write_summaries_2): Likewise.
2174 (ipa_write_optimization_summaries_1): Likewise.
2175 (ipa_read_optimization_summaries_1): Likewise.
2176 (execute_ipa_stmt_fixups): Likewise.
2177 * tree-pass.h (pass_data::has_gate): Remove.
2178 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
2179 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
2180 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
2181 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
2182 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
2183 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
2184 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
2185 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
2186 gimple-low.c, gimple-ssa-isolate-paths.c,
2187 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
2188 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
2189 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
2190 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
2191 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
2192 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
2193 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
2194 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
2195 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
2196 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
2197 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
2198 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
2199 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2200 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2201 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2202 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2203 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2204 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2205 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2206 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
2207 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
2208 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
2211 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2213 * pass_manager.h (pass_manager::register_dump_files_1): Remove
2215 * passes.c (pass_manager::register_dump_files_1): Merge into
2216 (pass_manager::register_dump_files): this, and remove its handling of
2217 properties since the pass always has the properties anyway.
2218 (pass_manager::pass_manager): Adjust.
2220 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2222 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
2223 * passes.c (pass_manager::register_dump_files_1): Remove dead code
2224 dealing with properties.
2225 (pass_manager::register_dump_files): Adjust.
2227 2014-03-20 Mark Wielaard <mjw@redhat.com>
2229 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
2230 then represent the bound as normal constant value.
2232 2014-04-17 Jakub Jelinek <jakub@redhat.com>
2235 Forward port from 4.8 branch
2236 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
2238 * config/i386/bmiintrin.h (_blsi_u32): New.
2242 (_blsmsk_u32): Ditto.
2243 (_blsmsk_u64): Ditto.
2244 (_tzcnt_u32): Ditto.
2245 (_tzcnt_u64): Ditto.
2247 2014-04-17 Kito Cheng <kito@0xlab.org>
2249 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
2251 2014-04-17 Richard Biener <rguenther@suse.de>
2254 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
2255 boolean results for comparisons.
2257 2014-04-17 Richard Biener <rguenther@suse.de>
2259 PR tree-optimization/60836
2260 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
2261 initial PHI args to be gimple values.
2263 2014-04-17 Richard Biener <rguenther@suse.de>
2265 PR tree-optimization/60841
2266 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
2267 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
2268 of stmts to SLP build.
2269 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
2270 (vect_analyze_slp): Likewise.
2271 (vect_analyze_slp_instance): Likewise.
2272 (vect_build_slp_tree): Limit overall SLP tree growth.
2273 * tree-vectorizer.h (vect_analyze_data_refs,
2274 vect_analyze_slp): Adjust prototypes.
2276 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2278 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
2281 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2283 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
2284 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
2285 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
2286 for TARGET_SLOW_PSHUFB
2288 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2290 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
2291 * config/i386/i386.c (intel_cost): Ditto.
2293 2014-04-17 Joey Ye <joey.ye@arm.com>
2295 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
2297 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2299 * opts.c (common_handle_option): Disable -fipa-reference coorectly
2302 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2304 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
2305 (type_all_derivations_known_p): New predicate.
2306 (type_all_ctors_visible_p): New predicate.
2307 (type_possibly_instantiated_p): New predicate.
2308 (get_odr_type): Compute all_derivations_known.
2309 (dump_odr_type): Dump the flag.
2310 (maybe_record_type): Cleanup.
2311 (record_target_from_binfo): Add bases_to_consider array;
2312 record bases for types w/o instances and skip CXX destructor.
2313 (possible_polymorphic_call_targets_1): Add bases_to_consider
2314 and consider_construction parameters; check if type may have instance.
2315 (get_polymorphic_call_info): Set maybe_in_construction to true
2316 when we know nothing.
2317 (record_targets_from_bases): Skip CXX destructors; they are
2318 never called for types in construction.
2319 (possible_polymorphic_call_targets): Do not record target when
2320 type may not have instance.
2322 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2325 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
2326 external aliases alive, too.
2328 2014-04-16 Andrew Pinski <apinski@cavium.com>
2330 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
2333 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2335 * final.c (compute_alignments): Do not apply loop alignment to a block
2336 falling through to the exit.
2338 2014-04-16 Catherine Moore <clm@codesourcery.com>
2340 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
2341 Adjust constraints for microMIPS store patterns.
2343 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2345 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
2347 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2349 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
2350 (append_use): Run at -O0.
2351 (append_vdef): Likewise.
2352 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
2353 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
2355 2014-04-16 Jakub Jelinek <jakub@redhat.com>
2357 PR tree-optimization/60844
2358 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
2359 (propagate_op_to_single_use, remove_visited_stmt_chain,
2360 linearize_expr, repropagate_negates, reassociate_bb): Use it
2361 instead of gsi_remove.
2363 2014-04-16 Martin Jambor <mjambor@suse.cz>
2365 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
2366 ipa_transforms_to_apply.
2367 (cgraph_function_versioning): Assert that old_node has empty
2368 ipa_transforms_to_apply.
2369 * trans-mem.c (ipa_tm_create_version): Likewise.
2370 * tree-inline.c (tree_function_versioning): Do not duplicate
2371 ipa_transforms_to_apply.
2373 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2376 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
2378 Pass necessary as flags on 64-bit Solaris/x86.
2379 Use lowercase relocs for x86_64-*-*.
2380 * configure: Regenerate.
2382 2014-04-15 Jan Hubicka <jh@suse.cz>
2384 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
2385 (maybe_record_node, likely_target_p): Use it.
2387 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2390 Revert following patch
2392 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2395 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2396 software floating point or no floating point registers, do not
2397 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2398 in GPRs that occurs after we tested for GPRs that would never be
2401 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2402 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2403 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2404 specifically allow DDmode, since that does not use the SPE SIMD
2407 2014-03-21 Mark Wielaard <mjw@redhat.com>
2409 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
2410 as unsigned or int depending on type and value used.
2412 2014-04-15 Richard Biener <rguenther@suse.de>
2414 PR rtl-optimization/56965
2415 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
2416 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
2418 * alias.c (true_dependence_1): Do not call
2419 nonoverlapping_component_refs_p.
2420 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
2421 nonoverlapping_component_refs_p.
2422 (indirect_refs_may_alias_p): Likewise.
2424 2014-04-15 Teresa Johnson <tejohnson@google.com>
2426 * cfg.c (dump_bb_info): Fix flags check.
2427 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
2429 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2431 PR rtl-optimization/60663
2432 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
2435 2014-04-15 Richard Biener <rguenther@suse.de>
2437 * lto-streamer.h (LTO_major_version): Bump to 4.
2439 2014-04-15 Richard Biener <rguenther@suse.de>
2441 * common.opt (lto_partition_model): New enum.
2442 (flto-partition=): Merge separate options with a single with argument,
2443 add -flto-partition=one support.
2444 * flag-types.h (enum lto_partition_model): Declare.
2445 * opts.c (finish_options): Remove duplicate -flto-partition=
2447 * lto-wrapper.c (run_gcc): Adjust.
2449 2014-04-15 Richard Biener <rguenther@suse.de>
2451 * alias.c (ncr_compar): New function.
2452 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
2454 2014-04-15 Richard Biener <rguenther@suse.de>
2456 * alias.c (record_component_aliases): Do not walk BINFOs.
2458 2014-04-15 Richard Biener <rguenther@suse.de>
2460 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
2461 Add struct function argument and adjust.
2462 (find_func_aliases_for_call): Likewise.
2463 (find_func_aliases): Likewise.
2464 (find_func_clobbers): Likewise.
2465 (intra_create_variable_infos): Likewise.
2466 (compute_points_to_sets): Likewise.
2467 (ipa_pta_execute): Adjust. Do not push/pop cfun.
2469 2014-04-15 Richard Biener <rguenther@suse.de>
2471 * tree.c (iterative_hash_expr): Use enum tree_code_class
2472 to store TREE_CODE_CLASS.
2473 (tree_block): Likewise.
2474 (tree_set_block): Likewise.
2475 * tree.h (fold_build_pointer_plus_loc): Use
2476 convert_to_ptrofftype_loc.
2478 2014-04-15 Jakub Jelinek <jakub@redhat.com>
2481 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
2484 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
2486 * cfgloop.h (struct loop): Move force_vectorize down.
2487 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
2488 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
2489 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
2490 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
2491 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
2492 * tree-core.h (enum annot_expr_kind): Add new kind values.
2493 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
2494 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
2496 * tree.def (ANNOTATE_EXPR): Tweak comment.
2498 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2500 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
2503 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
2505 * tree.h (TYPE_IDENTIFIER): Declare.
2506 * tree.c (subrange_type_for_debug_p): Use it.
2507 * godump.c (go_format_type): Likewise.
2508 * dwarf2out.c (is_cxx_auto, modified_type_die,
2509 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
2510 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
2512 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2515 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
2517 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
2519 * config/i386/i386.c (examine_argument): Return bool. Return true if
2520 parameter should be passed in memory.
2521 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
2522 (construct_container): Update calls to examine_argument.
2523 (function_arg_advance_64): Ditto.
2524 (return_in_memory_32): Merge with ix86_return_in_memory.
2525 (return_in_memory_64): Ditto.
2526 (return_in_memory_ms_64): Ditto.
2528 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2530 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
2531 * coverage.c (coverage_compute_profile_id): Handle externally visible
2534 2014-04-14 Martin Jambor <mjambor@suse.cz>
2536 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
2537 DECL_DISREGARD_INLINE_LIMITS functions.
2539 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2542 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
2544 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2547 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
2548 optimize_insn_for_speed_p instead of
2549 optimize_function_for_speed_p.
2551 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
2553 * doc/invoke.texi (free): Document AArch64.
2555 2014-04-14 Richard Biener <rguenther@suse.de>
2557 PR tree-optimization/60042
2558 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
2559 (insert_into_preds_of_block): Do not prevent PHI insertion
2560 for REFERENCE exprs here ...
2561 (eliminate_dom_walker::before_dom_children): ... but prevent
2562 their use here under similar conditions when applied to the
2563 IL after PRE optimizations.
2565 2014-04-14 Richard Biener <rguenther@suse.de>
2567 * passes.def: Move early points-to after early SRA.
2569 2014-04-14 Richard Biener <rguenther@suse.de>
2571 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
2572 check for which sign-changes we allow when forwarding
2573 a converted value into a switch.
2575 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2577 * stor-layout.c (place_field): Finalize non-constant offset for the
2580 2014-04-14 Richard Biener <rguenther@suse.de>
2582 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
2584 (expand_switch_using_bit_tests_p): Likewise.
2585 (process_switch): Compute and pass on speed_p based on the
2587 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
2588 optimize_bb_for_speed_p.
2590 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2592 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
2593 * function.h (struct function): Rename has_force_vect_loops into
2594 has_force_vectorize_loops.
2595 * lto-streamer-in.c (input_cfg): Adjust for renaming.
2596 (input_struct_function_base): Likewise.
2597 * lto-streamer-out.c (output_cfg): Likewise.
2598 (output_struct_function_base): Likewise.
2599 * omp-low.c (expand_omp_simd): Likewise.
2600 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2601 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2602 (version_loop_for_if_conversion): Likewise.
2603 (tree_if_conversion): Likewise.
2604 (main_tree_if_conversion): Likewise.
2605 (gate_tree_if_conversion): Likewise.
2606 * tree-inline.c (copy_loops): Likewise.
2607 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2608 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2609 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2610 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2611 * tree-vectorizer.c (vectorize_loops): Likewise.
2612 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2614 2014-04-14 Richard Biener <rguenther@suse.de>
2617 * lto-streamer-out.c (wrap_refs): New function.
2618 (lto_output): Wrap symbol references in global initializes in
2619 type-preserving MEM_REFs.
2621 2014-04-14 Christian Bruel <christian.bruel@st.com>
2623 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
2625 2014-04-14 Christian Bruel <christian.bruel@st.com>
2627 * config/sh/sh.md (setmemqi): New expand pattern.
2628 * config/sh/sh.h (CLEAR_RATIO): Define.
2629 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2630 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2632 2014-04-14 Richard Biener <rguenther@suse.de>
2635 * fold-const.c (negate_expr_p): Don't negate directional rounding
2637 (fold_negate_expr): Likewise.
2639 2014-04-14 Richard Biener <rguenther@suse.de>
2641 PR tree-optimization/59817
2642 PR tree-optimization/60453
2643 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2644 recursion to catch all CHRECs in the scalar evolution and restrict
2645 the predicate for the remains appropriately.
2647 2014-04-12 Catherine Moore <clm@codesourcery.com>
2649 * config/mips/constraints.md: Add new register constraint "kb".
2650 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2651 (*movhi_internal): Likewise.
2652 (*movqi_internal): Likewise.
2653 * config/mips/mips.h (M16_STORE_REGS): New register class.
2654 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2655 (REG_CLASS_CONTENTS): Likewise.
2656 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2658 2014-04-11 Tobias Burnus <burnus@net-b.de>
2661 * doc/invoke.texi (-Wformat-signedness): Document it.
2662 (Wformat=2): Mention that this enables -Wformat-signedness.
2664 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2666 * common/config/epiphany/epiphany-common.c
2667 (epiphany_option_optimization_table): Enable section anchors by
2668 default at -O1 or higher.
2669 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2670 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2671 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2672 carries no extra cost.
2673 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2674 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2675 * config/epiphany/predicates.md (memclob_operand): New predicate.
2676 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2677 Use memclob_operand predicate and X constraint for operand 3.
2679 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2681 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2682 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2685 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2687 PR rtl-optimization/60651
2688 * mode-switching.c (optimize_mode_switching): Make sure to emit
2689 sets of a lower numbered entity before sets of a higher numbered
2690 entity to a mode of the same or lower priority.
2691 When creating a seginfo for a basic block that starts with a code
2692 label, move the insertion point past the code label.
2693 (new_seginfo): Document and enforce requirement that
2694 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2695 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2696 * doc/tm.texi: Regenerate.
2698 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2701 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2703 2013-04-11 Jakub Jelinek <jakub@redhat.com>
2705 * BASE-VER: Set to 4.10.0.
2707 2014-04-11 Tobias Burnus <burnus@net-b.de>
2710 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
2711 * doc/gcc.texi (Service): Update description in the @menu
2712 * doc/invoke.texi (Option Summary): Remove misplaced and
2715 2014-04-11 Steve Ellcey <sellcey@mips.com>
2716 Jakub Jelinek <jakub@redhat.com>
2719 * expr.c (convert_move): Use emit_store_flag_force instead of
2720 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2723 2014-04-11 Richard Biener <rguenther@suse.de>
2726 * varasm.c (assemble_alias): Avoid endless error reporting
2727 recursion by setting TREE_ASM_WRITTEN.
2729 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2731 * config/s390/s390.md: Add a splitter for NOT rtx.
2733 2014-04-11 Jakub Jelinek <jakub@redhat.com>
2735 PR rtl-optimization/60663
2736 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
2738 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
2739 Jakub Jelinek <jakub@redhat.com>
2742 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2743 flag from decl_node to node.
2745 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2748 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2749 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2750 ameliorating the cases where it can be.
2752 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
2755 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2757 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2758 (loadsync_<mode>): Change mode.
2759 (load_quadpti, store_quadpti): New.
2760 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2761 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2762 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2764 2014-04-09 Cong Hou <congh@google.com>
2767 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2770 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2772 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2773 instead of vnor to exploit possible fusion opportunity in the
2775 (altivec_expand_vec_perm_const_le): Likewise.
2777 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2779 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2780 (loadsync_<mode>): Change mode.
2781 (load_quadpti, store_quadpti): New.
2782 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2783 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2785 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2788 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2789 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2790 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2792 2014-04-08 Richard Biener <rguenther@suse.de>
2795 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2796 a 64bit widest int print double-int similar to on HWI64 hosts.
2798 2014-04-08 Richard Biener <rguenther@suse.de>
2800 PR tree-optimization/60785
2801 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2802 default defs properly.
2804 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2806 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2807 (Weffc++): Likewise.
2809 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2811 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2812 set completep to false rather than true.
2814 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2817 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2818 ARM_TARGET2_DWARF_FORMAT.
2820 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2823 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2824 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2827 2014-04-07 Richard Biener <rguenther@suse.de>
2829 PR tree-optimization/60766
2830 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2831 (may_eliminate_iv): Convert cand_value_at result to desired type.
2833 2014-04-07 Jason Merrill <jason@redhat.com>
2836 * common.opt (-fno-gnu-unique): Add.
2837 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2839 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2841 * haifa-sched.c: Fix outdated function reference and minor
2842 grammar errors in introductory comment.
2844 2014-04-07 Richard Biener <rguenther@suse.de>
2847 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2849 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2851 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2854 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2855 size accounting for thunks.
2856 (pa_asm_output_mi_thunk): Use final_start_function() and
2857 final_end_function() to output function start and end directives.
2859 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2861 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2862 device specific ISA/ feature information. Remove short_sp and
2863 errata_skip ds. Add avr_device_specific_features enum to have device
2865 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2866 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2867 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2868 updated device specific info.
2869 * config/avr/avr-mcus.def: Merge device specific details to
2870 dev_attribute field.
2871 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2873 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2874 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2875 assembler if RMW isa supported by current device.
2876 * config/avr/genmultilib.awk: Update as device info structure changed.
2877 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2879 2014-04-04 Cong Hou <congh@google.com>
2881 PR tree-optimization/60656
2882 * tree-vect-stmts.c (supportable_widening_operation):
2883 Fix a bug that elements in a vector with vect_used_by_reduction
2884 property are incorrectly reordered when the operation on it is not
2885 consistant with the one in reduction operation.
2887 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
2889 PR rtl-optimization/60155
2890 * gcse.c (record_set_data): New function.
2891 (single_set_gcse): New function.
2892 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2893 (hoist_code): Likewise.
2894 (get_pressure_class_and_nregs): Likewise.
2896 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2898 * explow.c (probe_stack_range): Emit a final optimization blockage.
2900 2014-04-04 Anthony Green <green@moxielogic.com>
2902 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2905 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
2908 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2909 flags are set only during streaming.
2910 * ipa.c (process_references, walk_polymorphic_call_targets,
2911 symtab_remove_unreachable_nodes): Drop bodies of always inline
2912 after early inlining.
2913 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
2915 2014-04-04 Jakub Jelinek <jakub@redhat.com>
2916 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2919 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2922 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2925 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2927 (cortex_a53_fdivd): Likewise.
2929 2014-04-04 Martin Jambor <mjambor@suse.cz>
2932 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2934 * cgraph.c (clone_of_p): Also return true if thunks match.
2935 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2936 cgraph_function_or_thunk_node and an obsolete comment.
2937 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2939 (build_function_decl_skip_args): Likewise.
2940 (set_new_clone_decl_and_node_flags): New function.
2941 (duplicate_thunk_for_node): Likewise.
2942 (redirect_edge_duplicating_thunks): Likewise.
2943 (cgraph_clone_node): New parameter args_to_skip, pass it to
2944 redirect_edge_duplicating_thunks which is called instead of
2945 cgraph_redirect_edge_callee.
2946 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
2947 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2949 2014-04-04 Jeff Law <law@redhat.com>
2952 * config/arm/predicates.md (const_int_I_operand): New predicate.
2953 (const_int_M_operand): Similarly.
2954 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2956 (insv_t2, extv_reg, extzv_t2): Likewise.
2957 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2958 (pop_multiple_with_writeback_and_return): Likewise.
2959 (vfp_pop_multiple_with_writeback): Likewise
2961 2014-04-04 Richard Biener <rguenther@suse.de>
2964 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2965 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2967 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2968 * gimplify.c (gimple_add_tmp_var_fn): New function.
2969 * gimple-expr.h (create_tmp_reg_fn): Declare.
2970 * gimple-expr.c (create_tmp_reg_fn): New function.
2971 * gimple-low.c (record_vars_into): Don't change cfun.
2972 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2973 code generation without cfun.
2975 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2978 * Makefile.in (install-driver): Fix shell scripting.
2980 2014-04-03 Cong Hou <congh@google.com>
2982 PR tree-optimization/60505
2983 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
2984 threshold of number of iterations below which no vectorization
2986 * tree-vect-loop.c (new_loop_vec_info):
2987 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
2988 * tree-vect-loop.c (vect_analyze_loop_operations):
2989 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
2990 * tree-vect-loop.c (vect_transform_loop):
2991 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
2992 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
2993 of iterations of the loop and see if we should build the epilogue.
2995 2014-04-03 Richard Biener <rguenther@suse.de>
2997 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
2998 (streamer_tree_cache_create): Adjust.
2999 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
3000 to allow optional nodes array.
3001 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
3002 (streamer_tree_cache_append): Likewise.
3003 (streamer_tree_cache_create): Create nodes array optionally
3004 as specified by parameter.
3005 * lto-streamer-out.c (create_output_block): Avoid maintaining
3006 the node array in the writer cache.
3007 (DFS_write_tree): Remove assertion.
3008 (produce_asm_for_decls): Free the out decl state hash table early.
3009 * lto-streamer-in.c (lto_data_in_create): Adjust for
3010 streamer_tree_cache_create prototype change.
3012 2014-04-03 Richard Biener <rguenther@suse.de>
3014 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
3015 set TREE_CHAIN to NULL_TREE.
3017 2014-04-03 Richard Biener <rguenther@suse.de>
3019 PR tree-optimization/60740
3020 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
3021 over all GIMPLE_COND operands.
3023 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
3025 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
3026 (Weffc++): Remove Scott's numbering, merge lists and reference
3029 2014-04-03 Nick Clifton <nickc@redhat.com>
3031 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
3034 2014-04-03 Martin Jambor <mjambor@suse.cz>
3036 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
3037 mention gcc_unreachable before failing.
3038 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
3041 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
3044 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
3045 inconsistent code and instead mark the context inconsistent.
3046 (possible_polymorphic_call_targets): For inconsistent contexts
3047 return empty complete list.
3049 2014-04-02 Anthony Green <green@moxielogic.com>
3051 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
3052 (extendqisi2, extendhisi2): Define.
3053 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
3054 (WCHAR_TYPE): Change to unsigned int.
3056 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3058 PR tree-optimization/60733
3059 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
3060 insertion point for PHI candidates to be the end of the feeding
3061 block for the PHI argument.
3063 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
3065 PR rtl-optimization/60650
3066 * lra-constraints.c (process_alt_operands): Decrease reject for
3067 earlyclobber matching.
3069 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3071 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
3073 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3075 * config/spu/spu.c (pad_bb): Do not crash when the last
3076 insn is CODE_FOR_blockage.
3078 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3080 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
3081 lies outside the target mode.
3083 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3086 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
3087 software floating point or no floating point registers, do not
3088 allow any type in the FPRs. Eliminate a test for SPE SIMD types
3089 in GPRs that occurs after we tested for GPRs that would never be
3092 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
3093 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
3094 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
3095 specifically allow DDmode, since that does not use the SPE SIMD
3098 2014-04-02 Richard Biener <rguenther@suse.de>
3101 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
3102 MODE_INTs. Properly use negv_optab.
3103 (expand_abs): Likewise.
3105 2014-04-02 Richard Biener <rguenther@suse.de>
3108 * Makefile.in (install-driver): Guard extra installs with special
3111 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
3113 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3116 2014-04-01 Richard Henderson <rth@redhat.com>
3119 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
3120 alternative enabled before register allocation.
3122 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
3124 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
3125 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
3127 (nios2_large_got_address): Remove unneeded 'sym' parameter.
3128 (nios2_got_address): Update nios2_large_got_address call site.
3129 (nios2_delegitimize_address): New function.
3130 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
3131 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
3132 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
3134 2014-04-01 Martin Husemann <martin@duskware.de>
3136 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
3139 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
3141 PR rtl-optimization/60604
3142 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
3143 check from register_operand.
3144 (register_operand): Redefine in terms of general_operand.
3145 (nonmemory_operand): Use register_operand for the non-constant cases.
3147 2014-04-01 Richard Biener <rguenther@suse.de>
3149 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
3151 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
3153 * doc/invoke.texi (mapp-regs): Clarify.
3155 2014-03-31 Ulrich Drepper <drepper@gmail.com>
3157 * config/i386/avx512fintrin.h (__v32hi): Define type.
3158 (__v64qi): Likewise.
3159 (_mm512_set1_epi8): Define.
3160 (_mm512_set1_epi16): Define.
3161 (_mm512_set4_epi32): Define.
3162 (_mm512_set4_epi64): Define.
3163 (_mm512_set4_pd): Define.
3164 (_mm512_set4_ps): Define.
3165 (_mm512_setr4_epi64): Define.
3166 (_mm512_setr4_epi32): Define.
3167 (_mm512_setr4_pd): Define.
3168 (_mm512_setr4_ps): Define.
3169 (_mm512_setzero_epi32): Define.
3171 2014-03-31 Martin Jambor <mjambor@suse.cz>
3174 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
3175 callsite_arguments_match_p. Updated all callers. Also check types of
3176 corresponding formal parameters and actual arguments.
3177 (not_all_callers_have_enough_arguments_p) Renamed to
3178 some_callers_have_mismatched_arguments_p.
3180 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
3182 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
3184 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
3187 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
3190 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
3192 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
3193 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
3195 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
3196 Use FMAMODE_NOVF512 mode iterator.
3197 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
3198 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
3199 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
3200 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
3202 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
3203 Use VF_128_256 mode iterator.
3204 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
3207 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3209 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
3210 static chain if needed.
3212 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
3215 * lra-constraints.c (index_part_to_reg): New.
3216 (process_address): Use it.
3218 2014-03-27 Jeff Law <law@redhat.com>
3219 Jakub Jelinek <jakub@redhat.com>
3222 * expr.c (do_tablejump): Use simplify_gen_binary rather than
3223 gen_rtx_{PLUS,MULT} to build up the address expression.
3225 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
3226 creating non-canonical RTL.
3228 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3231 * ipa-inline.c (want_inline_small_function_p): Short circuit large
3232 functions; reorganize to make cheap checks first.
3233 (inline_small_functions): Do not estimate growth when dumping;
3235 * ipa-inline.h (inline_summary): Add min_size.
3236 (growth_likely_positive): New function.
3237 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
3238 (set_cond_stmt_execution_predicate): Cleanup.
3239 (estimate_edge_size_and_time): Compute min_size.
3240 (estimate_calls_size_and_time): Likewise.
3241 (estimate_node_size_and_time): Likewise.
3242 (inline_update_overall_summary): Update min_size.
3243 (do_estimate_edge_time): Likewise.
3244 (do_estimate_edge_size): Update.
3245 (do_estimate_edge_hints): Update.
3246 (growth_likely_positive): New function.
3248 2014-03-28 Jakub Jelinek <jakub@redhat.com>
3251 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
3252 also if addr has VOIDmode.
3254 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3256 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
3257 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
3259 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
3260 instructions as well as AdvancedSIMD loads.
3262 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3264 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
3265 Use crypto_aese type.
3266 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
3267 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
3268 crypto_aese, crypto_aesmc. Move to types.md.
3269 * config/arm/types.md (crypto_aes): Split into crypto_aese,
3271 * config/arm/iterators.md (crypto_type): Likewise.
3273 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3275 * cgraph.c: Include expr.h and tree-dfa.h.
3276 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
3279 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
3282 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
3283 regs from checking multi-reg pseudos.
3285 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3287 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
3289 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3291 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
3292 if it would clobber the stack pointer, even temporarily.
3294 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
3296 * mode-switching.c: Make small adjustments to the top comment.
3298 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3300 * config/rs6000/constraints.md (wD constraint): New constraint to
3301 match the constant integer to get the top DImode/DFmode out of a
3302 vector in a VSX register.
3304 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
3305 match the constant integer to get the top DImode/DFmode out of a
3306 vector in a VSX register.
3308 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
3311 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
3314 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
3315 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
3317 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
3318 Optimize vec_extract of 64-bit values, where the value being
3319 extracted is in the top word, where we can use scalar
3320 instructions. Add direct move and store support. Combine the big
3321 endian/little endian vector select load support into a single insn.
3322 (vsx_extract_<mode>_internal1): Likewise.
3323 (vsx_extract_<mode>_internal2): Likewise.
3324 (vsx_extract_<mode>_load): Likewise.
3325 (vsx_extract_<mode>_store): Likewise.
3326 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
3327 combined into vsx_extract_<mode>_load.
3328 (vsx_extract_<mode>_one_le): Likewise.
3330 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
3331 define the top 64-bit vector element.
3333 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
3336 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3337 Document vec_vbpermq builtin.
3340 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
3341 enable use of xxsldwi and xxpermdi builtin functions.
3342 (vec_xxpermdi): Likewise.
3344 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3345 Document use of vec_xxsldwi and vec_xxpermdi builtins.
3347 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
3349 PR rtl-optimization/60650
3350 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
3352 (find_spills_for): New.
3353 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
3354 Spill all pseudos on the second iteration.
3356 2014-03-27 Marek Polacek <polacek@redhat.com>
3359 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
3362 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3364 * config/s390/s390.c (s390_can_use_return_insn): Check for
3365 call-saved FPRs on 31 bit.
3367 2014-03-27 Jakub Jelinek <jakub@redhat.com>
3370 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
3371 if they need regimplification, just drop them instead of
3372 calling gimple_regimplify_operands on them.
3374 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
3377 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
3378 (aarch64_frame_pointer_required): Adjust logic.
3379 (aarch64_can_eliminate): Adjust logic.
3380 (aarch64_override_options_after_change): Adjust logic.
3382 2014-03-27 Dehao Chen <dehao@google.com>
3384 * ipa-inline.c (early_inliner): Update node's inline info.
3386 2014-03-26 Dehao Chen <dehao@google.com>
3388 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
3389 compiler inserted conditional jumps for NAN float check.
3391 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3393 * ubsan.h (ubsan_create_data): Change second argument's type
3394 to const location_t *.
3395 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
3397 (ubsan_create_data): Change second argument to const location_t *PLOC.
3398 Create Loc field whenever PLOC is non-NULL.
3399 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
3400 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
3404 * real.c (real_to_integer2): Change type of low to UHWI.
3406 2014-03-26 Tobias Burnus <burnus@net-b.de>
3408 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
3409 (CILK_SELF_SPECS): New define.
3410 (driver_self_specs): Use it.
3412 2014-03-26 Richard Biener <rguenther@suse.de>
3414 * tree-pretty-print.c (percent_K_format): Implement special
3415 case for LTO and its stripped down BLOCK tree.
3417 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3420 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
3422 * tree-vrp.c (simplify_internal_call_using_ranges): If only
3423 one range is range_int_cst_p, but not both, at least optimize
3424 addition/subtraction of 0 and multiplication by 0 or 1.
3425 * gimple-fold.c (gimple_fold_call): Fold
3426 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
3427 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
3428 INTEGER_CSTs, try to fold at least x * 0 and y - y.
3430 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
3432 PR rtl-optimization/60452
3433 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
3434 <case REG>: Return 1 for invalid offsets from the frame pointer.
3436 2014-03-26 Marek Polacek <polacek@redhat.com>
3439 * doc/extend.texi (C Extensions): Mention variable-length arrays in
3442 2014-03-26 Marek Polacek <polacek@redhat.com>
3445 * doc/extend.texi (Designated Inits): Describe what happens to omitted
3448 2014-03-26 Marek Polacek <polacek@redhat.com>
3451 * ira-color.c (update_conflict_hard_regno_costs): Perform the
3452 multiplication in unsigned type.
3454 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3456 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
3458 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3460 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
3462 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3465 * cif-code.def (UNREACHABLE) New code.
3466 * ipa-inline.c (inline_small_functions): Skip edges to
3467 __builtlin_unreachable.
3468 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
3469 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
3470 predicate to __bulitin_unreachable.
3471 (set_cond_stmt_execution_predicate): Fix issue when
3472 invert_tree_comparison returns ERROR_MARK.
3473 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
3474 propagate to inline clones.
3475 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
3477 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
3478 * cgraphclones.c (cgraph_clone_node): If call destination is already
3479 ureachable, do not redirect it back.
3480 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
3483 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3485 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
3486 Do not modify inline clones.
3488 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3490 * config/i386/i386.md (general_sext_operand): New mode attr.
3491 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
3492 don't generate (sign_extend (const_int)).
3493 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
3494 operands[2]. Use We constraint instead of <i> and
3495 <general_sext_operand> predicate instead of <general_operand>.
3496 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
3497 * config/i386/constraints.md (We): New constraint.
3498 * config/i386/predicates.md (x86_64_sext_operand,
3499 sext_operand): New predicates.
3501 2014-03-25 Martin Jambor <mjambor@suse.cz>
3504 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
3505 inconsistent devirtualizations to __builtin_unreachable.
3507 2014-03-25 Marek Polacek <polacek@redhat.com>
3510 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
3512 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
3514 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
3515 order of elements for big-endian.
3517 2014-03-25 Richard Biener <rguenther@suse.de>
3520 * gimplify-me.c (gimple_regimplify_operands): Update the
3523 2014-03-25 Martin Jambor <mjambor@suse.cz>
3526 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
3527 (lto_output_varpool_node): Likewise.
3528 (input_overwrite_node): Likewise.
3529 (input_varpool_node): Likewise.
3531 2014-03-25 Richard Biener <rguenther@suse.de>
3533 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
3534 (run_gcc): Likewise.
3536 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3538 * combine.c (simplify_compare_const): Add MODE argument.
3539 Handle mode_width 0 as very large mode_width.
3540 (try_combine, simplify_comparison): Adjust callers.
3542 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
3543 type to avoid signed integer overflow.
3544 * explow.c (plus_constant): Likewise.
3546 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
3548 * doc/generic.texi: Correct typos.
3550 2014-03-24 Tobias Burnus <burnus@net-b.de>
3552 * doc/invoke.texi (-flto): Expand section about
3553 using static libraries with LTO.
3555 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3557 PR rtl-optimization/60501
3558 * optabs.def (addptr3_optab): New optab.
3559 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
3560 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
3561 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
3563 * lra.c (emit_add3_insn): Use the addptr pattern if available.
3565 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
3567 2014-03-24 Ulrich Drepper <drepper@gmail.com>
3569 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
3572 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
3573 (_mm256_undefined_ps): Define.
3574 (_mm256_undefined_pd): Define.
3575 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
3576 (_mm_undefined_pd): Define.
3577 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
3578 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
3579 (_mm512_undefined_ps): Define.
3580 (_mm512_undefined_pd): Define.
3581 Use _mm*_undefined_*.
3582 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
3584 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
3586 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
3587 (lshr_simd): DI mode added.
3588 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
3589 (aarch64_ushr_simddi): Likewise.
3590 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
3591 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
3592 (vshrd_n_u64): Likewise.
3594 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3596 * Makefile.in (s-macro_list): Depend on cc1.
3598 2014-03-23 Teresa Johnson <tejohnson@google.com>
3600 * ipa-utils.c (ipa_print_order): Use specified dump file.
3602 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3604 PR rtl-optimization/60601
3605 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3607 * gcc.c (eval_spec_function): Initialize save_growing_value.
3609 2014-03-22 Jakub Jelinek <jakub@redhat.com>
3612 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3613 code == MINUS_EXPR, never swap op0 with op1.
3615 * toplev.c (init_local_tick): Avoid signed integer multiplication
3617 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3618 shift by first operand's bitsize.
3620 2014-03-21 Jakub Jelinek <jakub@redhat.com>
3623 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3625 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3626 TARGET_ISA_64BIT_P(x).
3628 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3630 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3631 pattern for vector nor instead of subtract from splat(-1).
3632 (altivec_expand_vec_perm_const_le): Likewise.
3634 2014-03-21 Richard Henderson <rth@twiddle.net>
3637 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3638 related insns after epilogue_completed.
3640 2014-03-21 Martin Jambor <mjambor@suse.cz>
3643 * cgraph.h (symtab_node): New flag body_removed.
3644 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3645 when removing bodies.
3646 * symtab.c (dump_symtab_base): Dump body_removed flag.
3647 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3648 had their bodies removed.
3650 2014-03-21 Martin Jambor <mjambor@suse.cz>
3653 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3656 2014-03-21 Richard Biener <rguenther@suse.de>
3658 PR tree-optimization/60577
3659 * tree-core.h (struct tree_base): Document nothrow_flag use
3660 in VAR_DECL_NONALIASED.
3661 * tree.h (VAR_DECL_NONALIASED): New.
3662 (may_be_aliased): Adjust.
3663 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3665 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3667 * expr.c (expand_expr_real_1): Remove outdated comment.
3669 2014-03-20 Jakub Jelinek <jakub@redhat.com>
3672 * ira.c (adjust_cleared_regs): Call copy_rtx on
3673 *reg_equiv[REGNO (loc)].src_p before passing it to
3674 simplify_replace_fn_rtx.
3677 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3678 into CONST, put pic register as first operand of PLUS. Use
3679 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3681 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3683 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3685 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3687 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3688 around for store forwarding issue in the FPU on the UT699.
3689 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3690 loads and operations if -mfix-ut699 is specified.
3691 (divtf3_hq): Tweak attribute.
3692 (sqrttf2_hq): Likewise.
3694 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3696 * calls.c (store_one_arg): Remove incorrect const qualification on the
3697 type of the temporary.
3698 * cfgexpand.c (expand_return): Likewise.
3699 * expr.c (expand_constructor): Likewise.
3700 (expand_expr_real_1): Likewise.
3702 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3704 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3707 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3710 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3712 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3714 * config/arm/aarch-common-protos.h
3715 (alu_cost_table): Fix spelling of "extend".
3716 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3718 2014-03-19 Richard Biener <rguenther@suse.de>
3721 * tree-core.h (tree_type_common): Re-order pointer members
3722 to reduce recursion depth during GC walks.
3724 2014-03-19 Marek Polacek <polacek@redhat.com>
3727 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3728 before accessing it.
3730 2014-03-19 Richard Biener <rguenther@suse.de>
3733 * lto-streamer-in.c (input_function): In WPA stage do not drop
3736 2014-03-19 Jakub Jelinek <jakub@redhat.com>
3738 PR tree-optimization/60559
3739 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3740 with build_zero_cst assignment.
3742 2014-03-18 Kai Tietz <ktietz@redhat.com>
3744 PR rtl-optimization/56356
3745 * sdbout.c (sdbout_parms): Verify that parms'
3746 incoming argument is valid.
3747 (sdbout_reg_parms): Likewise.
3749 2014-03-18 Richard Henderson <rth@redhat.com>
3752 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3753 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
3754 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
3756 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3758 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
3759 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
3760 Italicize plugin event names in description. Explain that
3761 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3762 Remind that no GCC functions should be called after PLUGIN_FINISH.
3763 Explain what pragmas with expansion are.
3765 2014-03-18 Martin Liska <mliska@suse.cz>
3767 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3768 gimple call statement is update.
3769 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3770 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3772 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3775 * ubsan.c (ubsan_instrument_unreachable): Call
3776 initialize_sanitizer_builtins.
3777 (ubsan_pass): Likewise.
3780 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3781 varpool_finalize_decl instead of rest_of_decl_compilation.
3783 2014-03-18 Richard Biener <rguenther@suse.de>
3785 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3786 by using bitmap_and_compl instead of bitmap_and_compl_into.
3787 (df_rd_transfer_function): Likewise.
3789 2014-03-18 Richard Biener <rguenther@suse.de>
3791 * doc/lto.texi (fresolution): Fix typo.
3793 2014-03-18 Richard Biener <rguenther@suse.de>
3795 * doc/invoke.texi (flto): Update for changes in 4.9.
3797 2014-03-18 Richard Biener <rguenther@suse.de>
3799 * doc/loop.texi: Remove section on the removed lambda framework.
3800 Update loop docs with recent changes in preserving loop structure.
3802 2014-03-18 Richard Biener <rguenther@suse.de>
3804 * doc/lto.texi (-fresolution): Document.
3806 2014-03-18 Richard Biener <rguenther@suse.de>
3808 * doc/contrib.texi: Adjust my name.
3810 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3813 * internal-fn.c: Include diagnostic-core.h.
3814 (expand_BUILTIN_EXPECT): New function.
3815 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3816 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3817 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3818 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3820 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3821 Revert 3 argument __builtin_expect code.
3822 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3823 * gimple-fold.c (gimple_fold_call): Likewise.
3824 * tree.h (fold_builtin_expect): New prototype.
3825 * builtins.c (build_builtin_expect_predicate): Add predictor
3826 argument, if non-NULL, create 3 argument __builtin_expect.
3827 (fold_builtin_expect): No longer static. Add ARG2 argument,
3828 pass it through to build_builtin_expect_predicate.
3829 (fold_builtin_2): Adjust caller.
3830 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3831 * internal-fn.def (BUILTIN_EXPECT): New.
3833 2014-03-18 Tobias Burnus <burnus@net-b.de>
3836 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3837 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3838 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3840 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3843 * predict.c (combine_predictions_for_bb): Fix up formatting.
3844 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3845 fill what it points to if non-NULL.
3846 (tree_predict_by_opcode): Adjust caller, use the predictor.
3847 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3849 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3851 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3852 proper constant for the store mode.
3854 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3856 * symtab.c (change_decl_assembler_name): Fix transparent alias
3859 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3861 * config/aarch64/aarch64.c: Correct the comments about the
3862 aarch64 stack layout.
3864 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3866 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3867 check for GF_OMP_FOR_KIND_FOR.
3869 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3871 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3872 ymm and zmm register names.
3874 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3877 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3878 note creation for the 2010-08-31 changes.
3880 2014-03-17 Marek Polacek <polacek@redhat.com>
3883 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3884 as -fno-tree-loop-vectorize.
3885 (expand_omp_simd): Likewise.
3887 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3889 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3890 (eligible_for_call_delay): New prototype.
3891 * config/sparc/sparc.c (tls_call_delay): Rename into...
3892 (eligible_for_call_delay): ...this. Return false if the instruction
3893 cannot be put in the delay slot of a branch.
3894 (eligible_for_restore_insn): Simplify.
3895 (eligible_for_return_delay): Return false if the instruction cannot be
3896 put in the delay slot of a branch and simplify.
3897 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3898 put in the delay slot of a branch.
3899 * config/sparc/sparc.md (fix_ut699): New attribute.
3900 (tls_call_delay): Delete.
3901 (in_call_delay): Reimplement.
3902 (eligible_for_sibcall_delay): Rename into...
3903 (in_sibcall_delay): ...this.
3904 (eligible_for_return_delay): Rename into...
3905 (in_return_delay): ...this.
3906 (in_branch_delay): Reimplement.
3907 (in_uncond_branch_delay): Delete.
3908 (in_annul_branch_delay): Delete.
3910 2014-03-14 Richard Henderson <rth@redhat.com>
3913 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3914 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3915 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3916 (floathi splitters): Remove.
3917 (float<SWI48x>xf2): New pattern.
3918 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3919 code that tried to handle DImode for 32-bit, but which was excluded
3920 by the pattern's condition. Drop allocation of stack temporary.
3921 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3922 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3923 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3924 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3925 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3926 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3927 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3928 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3929 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3930 (*float<SWI48x><X87MODEF>2_i387): Remove.
3931 (all float _with_temp splitters): Remove.
3932 (*float<SWI48x><MODEF>2_i387): New pattern.
3933 (*float<SWI48><MODEF>2_sse): New pattern.
3934 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3935 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3937 2014-03-14 Jakub Jelinek <jakub@redhat.com>
3938 Marek Polacek <polacek@redhat.com>
3941 * common.opt (dump_base_name_prefixed): New Variable.
3942 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3943 if x_dump_base_name_prefixed is already set, set it at the end.
3945 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3947 PR rtl-optimization/60508
3948 * lra-constraints.c (get_reload_reg): Add new parameter
3950 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3951 Pass the new parameter values.
3953 2014-03-14 Richard Biener <rguenther@suse.de>
3955 * common.opt: Revert unintented changes from r205065.
3958 2014-03-14 Richard Biener <rguenther@suse.de>
3961 * cfghooks.c (split_block): Properly adjust all loops the
3962 block was a latch of.
3964 2014-03-14 Martin Jambor <mjambor@suse.cz>
3967 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3970 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
3973 * config/avr/avr.c (avr_set_current_function): Pass function name
3974 through default_strip_name_encoding before sanity checking instead
3975 of skipping the first char of the assembler name.
3977 2014-03-13 Richard Henderson <rth@redhat.com>
3980 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
3981 (ix86_force_to_memory, ix86_free_from_memory): Remove.
3982 * config/i386/i386-protos.h: Likewise.
3983 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
3984 in the expander instead of a splitter.
3985 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
3986 any possibility of requiring a memory.
3987 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
3988 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
3989 (fp branch splitters): Update for ix86_split_fp_branch.
3990 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
3991 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
3992 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
3993 (*fop_<MODEF>_2_i387): Remove f/r alternative.
3994 (*fop_<MODEF>_3_i387): Likewise.
3995 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
3996 (splitters for the fop_* register patterns): Remove.
3997 (fscalexf4_i387): Rename from *fscalexf4_i387.
3998 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
4000 2014-03-13 Jakub Jelinek <jakub@redhat.com>
4002 PR tree-optimization/59779
4003 * tree-dfa.c (get_ref_base_and_extent): Use double_int
4004 type for bitsize and maxsize instead of HOST_WIDE_INT.
4006 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
4008 PR rtl-optimization/57320
4009 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
4010 the CFG after thread_prologue_and_epilogue_insns.
4012 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
4014 PR rtl-optimization/57189
4015 * lra-constraints.c (process_alt_operands): Disfavor spilling
4018 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
4020 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
4022 2014-03-13 Jakub Jelinek <jakub@redhat.com>
4024 PR tree-optimization/59025
4026 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
4027 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
4029 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
4032 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
4033 calls of avr_out_plus_1.
4035 2014-03-13 Bin Cheng <bin.cheng@arm.com>
4037 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
4038 BB's single pred and update the father loop's latch info later.
4040 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
4042 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
4046 (VEC_base): Likewise.
4047 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
4048 registers, we need to swap double words in little endian mode.
4050 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
4051 to be a container mode for 128-bit integer operations added in ISA
4052 2.07. Unlike TImode and PTImode, the preferred register set is
4053 the Altivec/VMX registers for the 128-bit operations.
4055 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
4057 (rs6000_split_128bit_ok_p): Likewise.
4059 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
4060 macros for creating ISA 2.07 normal and overloaded builtin
4061 functions with 3 arguments.
4062 (BU_P8V_OVERLOAD_3): Likewise.
4063 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
4064 for use as overloaded functions.
4065 (VPERM_1TI_UNS): Likewise.
4066 (VSEL_1TI): Likewise.
4067 (VSEL_1TI_UNS): Likewise.
4068 (ST_INTERNAL_1ti): Likewise.
4069 (LD_INTERNAL_1ti): Likewise.
4070 (XXSEL_1TI): Likewise.
4071 (XXSEL_1TI_UNS): Likewise.
4072 (VPERM_1TI): Likewise.
4073 (VPERM_1TI_UNS): Likewise.
4074 (XXPERMDI_1TI): Likewise.
4075 (SET_1TI): Likewise.
4076 (LXVD2X_V1TI): Likewise.
4077 (STXVD2X_V1TI): Likewise.
4078 (VEC_INIT_V1TI): Likewise.
4079 (VEC_SET_V1TI): Likewise.
4080 (VEC_EXT_V1TI): Likewise.
4081 (EQV_V1TI): Likewise.
4082 (NAND_V1TI): Likewise.
4083 (ORC_V1TI): Likewise.
4084 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
4085 added in ISA 2.07. Add both normal 'altivec' builtins, and the
4087 (VADDUQM): Likewise.
4088 (VSUBCUQ): Likewise.
4089 (VADDEUQM): Likewise.
4090 (VADDECUQ): Likewise.
4091 (VSUBEUQM): Likewise.
4092 (VSUBECUQ): Likewise.
4094 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
4095 __int128_t and __uint128_t types.
4096 (__uint128_type): Likewise.
4097 (altivec_categorize_keyword): Add support for vector __int128_t,
4098 vector __uint128_t, vector __int128, and vector unsigned __int128
4099 as a container type for TImode operations that need to be done in
4100 VSX/Altivec registers.
4101 (rs6000_macro_to_expand): Likewise.
4102 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
4103 to support 128-bit integer instructions vaddcuq, vadduqm,
4104 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
4105 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
4107 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
4108 for V1TImode, and set up preferences to use VSX/Altivec registers.
4109 Setup VSX reload handlers.
4110 (rs6000_debug_reg_global): Likewise.
4111 (rs6000_init_hard_regno_mode_ok): Likewise.
4112 (rs6000_preferred_simd_mode): Likewise.
4113 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
4114 (easy_altivec_constant): Likewise.
4115 (output_vec_const_move): Likewise.
4116 (rs6000_expand_vector_set): Convert V1TImode set and extract to
4118 (rs6000_expand_vector_extract): Likewise.
4119 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
4121 (rs6000_const_vec): Add support for V1TImode.
4122 (rs6000_emit_le_vsx_load): Swap double words when loading or
4123 storing TImode/V1TImode.
4124 (rs6000_emit_le_vsx_store): Likewise.
4125 (rs6000_emit_le_vsx_move): Likewise.
4126 (rs6000_emit_move): Add support for V1TImode.
4127 (altivec_expand_ld_builtin): Likewise.
4128 (altivec_expand_st_builtin): Likewise.
4129 (altivec_expand_vec_init_builtin): Likewise.
4130 (altivec_expand_builtin): Likewise.
4131 (rs6000_init_builtins): Add support for V1TImode type. Add
4132 support for ISA 2.07 128-bit integer builtins. Define type names
4133 for the VSX/Altivec vector types.
4134 (altivec_init_builtins): Add support for overloaded vector
4135 functions with V1TImode type.
4136 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
4137 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
4139 (rs6000_split_128bit_ok_p): Likewise.
4140 (rs6000_handle_altivec_attribute): Create V1TImode from vector
4141 __int128_t and vector __uint128_t.
4143 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
4144 and mode attributes.
4151 (VS_scalar): Likewise.
4152 (VS_double): Likewise.
4153 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
4155 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
4156 we support the ISA 2.07 128-bit integer arithmetic instructions.
4157 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
4158 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
4159 and TImode types for use with the builtin functions.
4160 (V1TI_type_node): Likewise.
4161 (unsigned_V1TI_type_node): Likewise.
4162 (intTI_type_internal_node): Likewise.
4163 (uintTI_type_internal_node): Likewise.
4165 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
4166 128-bit builtin functions.
4167 (UNSPEC_VADDEUQM): Likewise.
4168 (UNSPEC_VADDECUQ): Likewise.
4169 (UNSPEC_VSUBCUQ): Likewise.
4170 (UNSPEC_VSUBEUQM): Likewise.
4171 (UNSPEC_VSUBECUQ): Likewise.
4172 (VM): Add V1TImode to vector mode iterators.
4174 (VI_unit): Likewise.
4175 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
4176 (altivec_vaddcuq): Likewise.
4177 (altivec_vsubuqm): Likewise.
4178 (altivec_vsubcuq): Likewise.
4179 (altivec_vaddeuqm): Likewise.
4180 (altivec_vaddecuq): Likewise.
4181 (altivec_vsubeuqm): Likewise.
4182 (altivec_vsubecuq): Likewise.
4184 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
4186 (BOOL_128): Likewise.
4187 (BOOL_REGS_OUTPUT): Likewise.
4188 (BOOL_REGS_OP1): Likewise.
4189 (BOOL_REGS_OP2): Likewise.
4190 (BOOL_REGS_UNARY): Likewise.
4191 (BOOL_REGS_AND_CR0): Likewise.
4193 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
4194 128-bit integer builtin support.
4195 (vec_vadduqm): Likewise.
4196 (vec_vaddecuq): Likewise.
4197 (vec_vaddeuqm): Likewise.
4198 (vec_vsubecuq): Likewise.
4199 (vec_vsubeuqm): Likewise.
4200 (vec_vsubcuq): Likewise.
4201 (vec_vsubuqm): Likewise.
4203 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4204 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
4205 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
4206 128-bit integer add/subtract to ISA 2.07.
4208 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
4210 * config/arc/arc.c (arc_predicate_delay_insns):
4211 Fix third argument passed to conditionalize_nonjump.
4213 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
4215 * config/aarch64/aarch64-builtins.c
4216 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
4217 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
4218 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
4219 instead of __builtin_lfloor.
4220 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
4222 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4224 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
4225 (tree_ssa_ifcombine_bb_1): New function.
4226 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
4227 is an empty forwarder block to then_bb or vice versa and then_bb
4228 and else_bb are effectively swapped.
4230 2014-03-12 Christian Bruel <christian.bruel@st.com>
4233 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
4234 REG_CFA_DEF_CFA note.
4235 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
4236 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
4238 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
4240 PR tree-optimization/60454
4241 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
4243 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4245 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
4246 Do not define target_cpu_default2 to generic.
4247 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
4248 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
4249 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
4251 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4252 Marc Glisse <marc.glisse@inria.fr>
4254 PR tree-optimization/60502
4255 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
4256 instead of build_low_bits_mask.
4258 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4261 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
4262 if there are multiple uses, but op doesn't live on E edge.
4263 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
4264 clobber stmts before __builtin_unreachable.
4266 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
4268 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
4269 hard_frame_pointer_rtx.
4270 * cse.c (cse_insn): Remove volatile check.
4271 * cselib.c (cselib_process_insn): Likewise.
4272 * dse.c (scan_insn): Likewise.
4274 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4276 * config/arc/arc.c (conditionalize_nonjump): New function,
4278 (arc_ifcvt): ... this.
4279 (arc_predicate_delay_insns): Use it.
4281 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4283 * config/arc/predicates.md (extend_operand): During/after reload,
4284 allow const_int_operand.
4285 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
4286 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
4287 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
4289 (umulsi3_highpart_i): Likewise.
4291 2014-03-11 Richard Biener <rguenther@suse.de>
4293 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
4294 Add asserts to guard possible wrong-code bugs.
4296 2014-03-11 Richard Biener <rguenther@suse.de>
4298 PR tree-optimization/60429
4299 PR tree-optimization/60485
4300 * tree-ssa-structalias.c (set_union_with_increment): Properly
4301 take into account all fields that overlap the shifted vars.
4302 (do_sd_constraint): Likewise.
4303 (do_ds_constraint): Likewise.
4304 (get_constraint_for_ptr_offset): Likewise.
4306 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
4308 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
4309 (nios2_compute_frame_layout):
4310 Add calculation of cfun->machine->fp_save_offset.
4311 (nios2_expand_prologue): Correct setting of frame pointer register
4313 (nios2_expand_epilogue): Update recovery of stack pointer from
4314 frame pointer accordingly.
4315 (nios2_initial_elimination_offset): Update calculation of offset
4316 for eliminating to HARD_FRAME_POINTER_REGNUM.
4318 2014-03-10 Jakub Jelinek <jakub@redhat.com>
4321 * ipa.c (symtab_remove_unreachable_nodes): Don't call
4322 cgraph_get_create_node on VAR_DECLs.
4324 2014-03-10 Richard Biener <rguenther@suse.de>
4327 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
4329 2014-03-08 Douglas B Rupp <rupp@gnat.com>
4331 * config/vms/vms.opt (vms_float_format): New variable.
4333 2014-03-08 Tobias Burnus <burnus@net-b.de>
4335 * doc/invoke.texi (-fcilkplus): Update implementation status.
4337 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
4338 Richard Biener <rguenther@suse.de>
4340 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
4341 consistently accross all TUs.
4342 (run_gcc): Enable -fshort-double automatically at link at link-time
4343 and disallow override.
4345 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
4348 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
4349 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
4350 if they can't be used.
4352 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4354 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
4355 for Solaris 11/x86 ld.
4356 * configure: Regenerate.
4358 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4360 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
4361 (LIB_TLS_SPEC): Save as ld_tls_libs.
4362 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
4363 (HAVE_AS_IX86_TLSLDM): New test.
4364 * configure, config.in: Regenerate.
4365 * config/i386/i386.c (legitimize_tls_address): Fall back to
4366 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
4367 cannot support TLS_MODEL_LOCAL_DYNAMIC.
4368 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
4369 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
4371 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
4373 * common.opt (fira-loop-pressure): Mark as optimization.
4375 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
4377 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
4378 an OpenMP mappable type.
4380 2014-03-06 Matthias Klose <doko@ubuntu.com>
4382 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
4383 MULTILIB_OSDIRNAMES is not defined.
4385 2014-03-06 Jakub Jelinek <jakub@redhat.com>
4386 Meador Inge <meadori@codesourcery.com>
4389 * config/arm/arm.c (arm_tls_symbol_p): Remove.
4390 (arm_legitimize_address): Call legitimize_tls_address for any
4391 arm_tls_referenced_p expression, handle constant addend. Call it
4392 before testing for !TARGET_ARM.
4393 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
4395 2014-03-06 Richard Biener <rguenther@suse.de>
4401 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4403 * tree-streamer.c (record_common_node): Assert we don't record
4404 nodes with type double.
4405 (preload_common_node): Skip type double, complex double and double
4406 pointer since it is now frontend dependent due to fshort-double option.
4408 2014-03-06 Richard Biener <rguenther@suse.de>
4410 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
4411 or -fno-lto is specified and the linker has full plugin support.
4412 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
4413 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
4414 * lto-wrapper.c (merge_and_complain): Merge compile-time
4415 optimization levels.
4416 (run_gcc): And pass it through to the link options.
4418 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
4422 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4424 * cselib.c (remove_useless_values): Skip to avoid quadratic
4425 behavior if the condition moved from...
4426 (cselib_process_insn): ... here holds.
4428 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4431 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
4432 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
4435 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
4436 (TM_H): Add x86-tune.def.
4438 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4440 * config/aarch64/aarch64.c (generic_tunings):
4441 Use cortexa57_extra_costs.
4443 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4446 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
4447 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
4448 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
4451 2014-03-04 Heiher <r@hev.cc>
4453 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
4454 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
4456 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
4458 * config/i386/predicates.md (const2356_operand): Change to ...
4459 (const2367_operand): ... this.
4460 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
4462 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4463 (*avx512pf_scatterpf<mode>sf): Ditto.
4464 (avx512pf_scatterpf<mode>df): Ditto.
4465 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4466 (*avx512pf_scatterpf<mode>df): Ditto.
4467 * config/i386/i386.c (ix86_expand_builtin): Update
4468 incorrect hint operand error message.
4470 2014-03-04 Richard Biener <rguenther@suse.de>
4472 * lto-section-in.c (lto_get_section_data): Fix const cast.
4474 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4476 * tree-streamer.c (record_common_node): Assert we don't record
4477 nodes with type double.
4478 (preload_common_node): Skip type double, complex double and double
4479 pointer since it is now frontend dependent due to fshort-double option.
4481 2014-03-04 Richard Biener <rguenther@suse.de>
4484 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
4485 (lto_input_toplevel_asms): Likewise.
4486 * lto-section-in.c (lto_get_section_data): Instead do it here
4489 2014-03-04 Richard Biener <rguenther@suse.de>
4491 PR tree-optimization/60382
4492 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
4493 dead PHIs a reduction.
4495 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
4497 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
4499 (_mm_prefetch): Move out of GCC target("sse") pragma.
4500 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
4501 GCC target("prfchw") pragma.
4502 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
4504 * config/i386/i386.c (ix86_option_override_internal): Enable
4505 -mprfchw with -mprefetchwt1.
4507 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4509 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
4512 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4514 * opts.h (CL_PCH_IGNORE): Define.
4515 * targhooks.c (option_affects_pch_p):
4516 Return false for options that have CL_PCH_IGNORE set.
4517 * opt-functions.awk: Process PchIgnore.
4518 * doc/options.texi: Document PchIgnore.
4520 * config/arc/arc.opt (misize): Add PchIgnore property.
4522 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4524 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
4525 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
4526 constraint on constants to permit them being loaded into
4527 GENERAL_REGS or BASE_REGS.
4529 2014-03-03 Nick Clifton <nickc@redhat.com>
4531 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
4532 anti-cacnonical alternatives.
4533 (negandhi3_real): New pattern.
4534 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
4536 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4538 * config/avr/avr-mcus.def: Remove atxmega16x1.
4539 * config/avr/avr-tables.opt: Regenerate.
4540 * config/avr/t-multilib: Regenerate.
4541 * doc/avr-mmcu.texi: Regenerate.
4543 2014-03-03 Tobias Grosser <tobias@grosser.es>
4544 Mircea Namolaru <mircea.namolaru@inria.fr>
4546 PR tree-optimization/58028
4547 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
4550 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4552 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
4553 not handled by recognizers.
4555 2014-03-03 Jakub Jelinek <jakub@redhat.com>
4558 * function.c (expand_function_end): Don't emit
4559 clobber_return_register sequence if clobber_after is a BARRIER.
4560 * cfgexpand.c (construct_exit_block): Append instructions before
4561 return_label to prev_bb.
4563 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4565 * config/rs6000/constraints.md: Document reserved use of "wc".
4567 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4570 * ipa.c (function_and_variable_visibility): When dissolving comdat
4571 group, also set all symbols to local.
4573 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4578 2013-12-14 Jan Hubicka <jh@suse.cz>
4580 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
4582 2014-03-02 Jon Beniston <jon@beniston.com>
4588 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
4589 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
4590 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
4591 (simple_return, *simple_return): New patterns
4592 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
4593 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
4595 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
4597 * dwarf2out.c (gen_subprogram_die): Tidy.
4599 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4602 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4603 (*mov_t_msb_neg_negc): ... this new insn.
4605 2014-02-28 Jason Merrill <jason@redhat.com>
4608 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4611 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4614 * dwarf2out.c (decltype_auto_die): New static.
4615 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4616 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4617 (is_cxx_auto): Likewise.
4619 2014-02-28 Ian Bolton <ian.bolton@arm.com>
4621 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4622 we are not using general regs only.
4624 2014-02-28 Richard Biener <rguenther@suse.de>
4627 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4628 previous fix and only allow to remove trivial pre-headers
4629 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
4630 (remove_forwarder_block): Properly update the latch of a loop.
4632 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4635 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4636 (cselib_preserved_hash_table): New.
4637 (preserve_constants_and_equivs): Move preserved vals to it.
4638 (cselib_find_slot): Look it up first.
4639 (cselib_init): Initialize it.
4640 (cselib_finish): Release it.
4641 (dump_cselib_table): Dump it.
4643 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4646 * cselib.c (remove_useless_values): Skip to avoid quadratic
4647 behavior if the condition moved from...
4648 (cselib_process_insn): ... here holds.
4650 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4653 * var-tracking.c (vt_initialize): Apply the same condition to
4654 preserve the CFA base value.
4656 2014-02-28 Joey Ye <joey.ye@arm.com>
4659 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4660 if reload in progress or completed.
4662 2014-02-28 Tobias Burnus <burnus@net-b.de>
4665 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4668 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4670 * doc/tm.texi.in (Condition Code Status): Update documention for
4671 relative locations of cc0-setter and cc0-user.
4673 2014-02-27 Jeff Law <law@redhat.com>
4675 PR rtl-optimization/52714
4676 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4677 into two independent simple sets, if I3 is a jump, ensure the
4678 pattern we place into I3 is a (set (pc) ...).
4680 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4681 Jeff Law <law@redhat.com>
4683 PR rtl-optimization/49847
4684 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4685 are in different blocks.
4686 * doc/tm.texi (Condition Code Status): Update documention for
4687 relative locations of cc0-setter and cc0-user.
4689 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4692 * lra.c (lra_emit_add): Check SUBREG too.
4694 2014-02-27 Andreas Schwab <schwab@suse.de>
4696 * config/m68k/m68k.c (m68k_option_override): Disable
4697 -flive-range-shrinkage for classic m68k.
4698 (m68k_override_options_after_change): Likewise.
4700 2014-02-27 Marek Polacek <polacek@redhat.com>
4703 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4704 -Wmaybe-uninitialized.
4706 2014-02-27 Alan Modra <amodra@gmail.com>
4709 * reload1.c (emit_input_reload_insns): When reload_override_in,
4710 set old to rl->in_reg when rl->in_reg is a subreg.
4712 2014-02-26 Richard Biener <rguenther@suse.de>
4715 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4717 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4719 * common/config/i386/predicates.md (const1256_operand): Remove.
4720 (const2356_operand): New.
4721 (const_1_to_2_operand): Remove.
4722 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4723 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4724 (*avx512pf_gatherpf<mode>sf): Ditto.
4725 (avx512pf_gatherpf<mode>df): Ditto.
4726 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4727 (*avx512pf_gatherpf<mode>df): Ditto.
4728 (avx512pf_scatterpf<mode>sf): Ditto.
4729 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4730 (*avx512pf_scatterpf<mode>sf): Ditto.
4731 (avx512pf_scatterpf<mode>df): Ditto.
4732 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4733 (*avx512pf_scatterpf<mode>df): Ditto.
4734 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4736 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4738 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4739 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4740 (_mm512_mask_testn_epi64_mask): Move to ...
4741 * config/i386/avx512cdintrin.h: Here.
4742 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4743 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4744 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4745 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4746 TARGET_AVX512F from TARGET_AVX512CD.
4748 2014-02-26 Richard Biener <rguenther@suse.de>
4751 * ipa.c (walk_polymorphic_call_targets): Properly guard
4752 call to inline_update_overall_summary.
4754 2014-02-26 Bin Cheng <bin.cheng@arm.com>
4757 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4758 and latches only if requested. Fix latch if it is removed.
4759 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4760 LOOPS_HAVE_PREHEADERS.
4762 2014-02-25 Andrew Pinski <apinski@cavium.com>
4764 * builtins.c (expand_builtin_thread_pointer): Create a new target
4765 when the target is NULL.
4767 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4769 PR rtl-optimization/60317
4770 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4771 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4772 * lra-assigns.c: Include params.h.
4773 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4774 other reload pseudos considerations.
4776 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4778 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4779 to use canonical form for nor<mode>3.
4781 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4784 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4787 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4789 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4790 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4791 (ix86_handle_option): Handle OPT_mprefetchwt1.
4792 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4793 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4795 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4796 OPTION_MASK_ISA_PREFETCHWT1.
4797 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4798 (PTA_PREFETCHWT1): New.
4799 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4800 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4801 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4802 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4803 (*prefetch_avx512pf_<mode>_: Change into ...
4804 (*prefetch_prefetchwt1_<mode>: This.
4805 * config/i386/i386.opt (mprefetchwt1): New.
4806 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4807 (_mm_prefetch): Handle intent to write.
4808 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4810 2014-02-25 Richard Biener <rguenther@suse.de>
4813 * emit-rtl.c (mem_attrs_htab): Remove.
4814 (mem_attrs_htab_hash): Likewise.
4815 (mem_attrs_htab_eq): Likewise.
4816 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4817 (init_emit_once): Do not allocate mem_attrs_htab.
4819 2014-02-25 Richard Biener <rguenther@suse.de>
4822 * lto-opts.c (lto_write_options): Output non-explicit conservative
4823 -fwrapv, -fno-trapv and -fno-strict-overflow.
4824 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4825 (run_gcc): And pass them through.
4827 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4829 * sel-sched.c (calculate_new_fences): New parameter ptime.
4830 Calculate it as a maximum over all fence cycles.
4831 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4832 Print the final schedule timing when sched_verbose.
4834 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4836 PR rtl-optimization/60292
4837 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4838 bit fot the fence instruction.
4840 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4842 * calls.h: Fix typo in comment.
4844 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4846 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4847 adjusting offsetable addresses.
4849 2014-02-24 Guozhi Wei <carrot@google.com>
4851 * sparseset.h (sparseset_pop): Fix the wrong index.
4853 2014-02-24 Walter Lee <walt@tilera.com>
4855 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4856 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4858 * common/config/tilegx/tilegx-common.c
4859 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4860 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4862 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4863 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4864 (tilegx_gimplify_va_arg_expr): Handle big endian.
4865 (tilegx_expand_unaligned_load): Ditto.
4866 (tilegx_expand_unaligned_store): Ditto.
4867 (TARGET_RETURN_IN_MSB): New.
4868 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4869 (TARGET_ENDIAN_DEFAULT): New.
4870 (TARGET_BIG_ENDIAN): Handle big endian.
4871 (BYTES_BIG_ENDIAN): Ditto.
4872 (WORDS_BIG_ENDIAN): Ditto.
4873 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4876 * config/tilegx/tilegx.md (extv): Handle big endian.
4878 (insn_st<n>): Ditto.
4879 (insn_st<n>_add<bitsuffix>): Ditto.
4880 (insn_stnt<n>): Ditto.
4881 (insn_stnt<n>_add<bitsuffix>):Ditto.
4882 (vec_interleave_highv8qi): Handle big endian.
4883 (vec_interleave_highv8qi_be): New.
4884 (vec_interleave_highv8qi_le): New.
4885 (insn_v1int_h): Handle big endian.
4886 (vec_interleave_lowv8qi): Handle big endian.
4887 (vec_interleave_lowv8qi_be): New.
4888 (vec_interleave_lowv8qi_le): New.
4889 (insn_v1int_l): Handle big endian.
4890 (vec_interleave_highv4hi): Handle big endian.
4891 (vec_interleave_highv4hi_be): New.
4892 (vec_interleave_highv4hi_le): New.
4893 (insn_v2int_h): Handle big endian.
4894 (vec_interleave_lowv4hi): Handle big endian.
4895 (vec_interleave_lowv4hi_be): New.
4896 (vec_interleave_lowv4hi_le): New.
4897 (insn_v2int_l): Handle big endian.
4898 (vec_interleave_highv2si): Handle big endian.
4899 (vec_interleave_highv2si_be): New.
4900 (vec_interleave_highv2si_le): New.
4901 (insn_v4int_h): Handle big endian.
4902 (vec_interleave_lowv2si): Handle big endian.
4903 (vec_interleave_lowv2si_be): New.
4904 (vec_interleave_lowv2si_le): New.
4905 (insn_v4int_l): Handle big endian.
4906 * config/tilegx/tilegx.opt (mbig-endian): New option.
4907 (mlittle-endian): New option.
4908 * doc/install.texi: Document tilegxbe-linux.
4909 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4911 2014-02-24 Martin Jambor <mjambor@suse.cz>
4914 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4915 there are no parameter descriptors.
4917 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
4919 PR rtl-optimization/60268
4920 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4921 initialization to ...
4922 (sched_rgn_init): ... here.
4923 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4925 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4927 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4930 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4932 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4935 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4937 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4938 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
4940 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4942 * config/microblaze/predicates.md: Add cmp_op predicate.
4943 * config/microblaze/microblaze.md: Add branch_compare instruction
4944 which uses cmp_op predicate and emits cmp insn before branch.
4945 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
4946 to microblaze_expand_conditional_branch and consolidate logic.
4947 (microblaze_expand_conditional_branch): emit branch_compare
4948 insn instead of handling cmp op separate from branch insn.
4950 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4952 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4955 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4957 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4958 define_insn with define_expand and new define_insn
4959 *altivec_lve<VI_char>x_internal.
4960 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4961 and new define_insn *altivec_stve<VI_char>x_internal.
4962 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4964 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4966 (altivec_expand_stvex_be): New function.
4968 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4970 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
4971 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
4972 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4973 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4975 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4978 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
4979 instead of emit_move_insn.
4981 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4983 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
4985 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
4986 gen_altivec_vsumsws.
4988 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4990 * config/rs6000/altivec.md (altivec_lvxl): Rename as
4991 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
4992 (altivec_lvxl_<mode>): New define_expand incorporating
4993 -maltivec=be semantics where needed.
4994 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
4995 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
4996 semantics where needed.
4997 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
4998 (altivec_stvx_<mode>): New define_expand incorporating
4999 -maltivec=be semantics where needed.
5000 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
5001 VM2 iterator instead of V4SI.
5002 (altivec_stvxl_<mode>): New define_expand incorporating
5003 -maltivec=be semantics where needed.
5004 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
5005 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
5006 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
5007 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
5008 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
5009 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
5010 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
5011 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
5012 ALTIVEC_BUILTIN_STVXL.
5013 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
5014 (altivec_expand_stvx_be): Likewise.
5015 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
5016 (altivec_expand_lvx_be): Likewise.
5017 (altivec_expand_stvx_be): Likewise.
5018 (altivec_expand_builtin): Add cases for
5019 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
5020 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
5021 (altivec_init_builtins): Add definitions for
5022 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
5023 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
5025 2014-02-21 Catherine Moore <clm@codesourcery.com>
5027 * doc/invoke.texi (mvirt, mno-virt): Document.
5028 * config/mips/mips.opt (mvirt): New option.
5029 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
5031 2014-02-21 Richard Biener <rguenther@suse.de>
5033 PR tree-optimization/60276
5034 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
5035 (STMT_VINFO_MIN_NEG_DIST): New macro.
5036 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
5037 STMT_VINFO_MIN_NEG_DIST.
5038 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
5039 made for negative dependence distances still hold.
5041 2014-02-21 Richard Biener <rguenther@suse.de>
5044 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
5045 DECL_INITIAL for globals not in the current function context.
5047 2014-02-21 Jakub Jelinek <jakub@redhat.com>
5049 PR tree-optimization/56490
5050 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
5051 * tree-ssa-uninit.c: Include params.h.
5052 (compute_control_dep_chain): Add num_calls argument, return false
5053 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
5054 num_calls to recursive call.
5055 (find_predicates): Change dep_chain into normal array,
5056 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
5057 variable and adjust compute_control_dep_chain caller.
5058 (find_def_preds): Likewise.
5060 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
5062 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
5063 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
5065 2014-02-21 Nick Clifton <nickc@redhat.com>
5067 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
5068 (pushhi1): Likewise.
5069 (popqi1): Add mode to pre_dec.
5072 2014-02-21 Jakub Jelinek <jakub@redhat.com>
5074 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
5075 mode for mask of V8SFmode permutation.
5077 2014-02-20 Richard Henderson <rth@redhat.com>
5080 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
5081 a new pseudo for OLDVAL.
5083 2014-02-20 Jakub Jelinek <jakub@redhat.com>
5086 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
5087 gen_reg_rtx if d->testing_p.
5088 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
5089 if d->testing_p and we will certainly return true.
5090 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
5093 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
5095 * emit-rtl.c (gen_reg_rtx): Assert that
5096 crtl->emit.regno_pointer_align_length is non-zero.
5098 2014-02-20 Richard Henderson <rth@redhat.com>
5101 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
5102 on failure the store back into EXPECT.
5104 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
5105 Sandra Loosemore <sandra@codesourcery.com>
5107 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
5108 * config/nios2/nios2.c (nios2_function_profiler): Add
5109 -fPIC (flag_pic == 2) support.
5110 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
5111 (nios2_large_offset_p): New function.
5112 (nios2_unspec_reloc_p): Move up position, update to use
5113 nios2_large_offset_p.
5114 (nios2_unspec_address): Remove function.
5115 (nios2_unspec_offset): New function.
5116 (nios2_large_got_address): New function.
5117 (nios2_got_address): Add large offset support.
5118 (nios2_legitimize_tls_address): Update usage of removed and new
5120 (nios2_symbol_binds_local_p): New function.
5121 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
5122 (nios2_legitimize_address): Update to use nios2_large_offset_p.
5123 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
5124 (nios2_print_operand): Merge H/L processing, add hiadj/lo
5125 processing for (const (unspec ...)).
5126 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
5128 2014-02-20 Richard Biener <rguenther@suse.de>
5130 * tree-cfg.c (replace_uses_by): Mark altered BBs before
5131 doing the substitution.
5132 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
5134 2014-02-20 Martin Jambor <mjambor@suse.cz>
5137 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
5138 info when checking whether lattices are bottom.
5140 2014-02-20 Richard Biener <rguenther@suse.de>
5143 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
5146 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
5149 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
5150 parameter specifying the scaling.
5151 (inline_call): Update.
5152 (want_inline_recursively): Guard division by zero.
5153 (recursive_inlining): Update.
5154 * ipa-inline.h (clone_inlined_nodes): Update.
5156 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
5159 * config/i386/i386.c (classify_argument): Pass structures of size
5160 64 bytes or less in register.
5162 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
5163 Kirill Yukhin <kirill.yukhin@intel.com>
5165 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
5166 (_mm_rcp28_round_ss): Ditto.
5167 (_mm_rsqrt28_round_sd): Ditto.
5168 (_mm_rsqrt28_round_ss): Ditto.
5169 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
5170 (_mm_rcp14_round_ss): Ditto.
5171 (_mm_rsqrt14_round_sd): Ditto.
5172 (_mm_rsqrt14_round_ss): Ditto.
5173 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
5174 the first input operand, get rid of match_dup.
5175 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
5177 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
5179 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
5180 operand as the first input operand, set type attribute.
5181 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
5183 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
5184 operand as the first input operand, set type attribute.
5186 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5188 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
5191 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
5194 * config/i386/i386.c (construct_container): Remove TFmode check
5195 for X86_64_INTEGER_CLASS.
5197 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
5200 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
5201 only when -Wpsabi is enabled.
5203 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
5206 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
5207 passing arrays in registers are the same as for structs, so remove the
5208 special case for them.
5210 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
5212 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
5213 destination type, extract only the valid bits if the source type is not
5214 integral and has a different mode.
5216 2014-02-19 Richard Biener <rguenther@suse.de>
5219 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
5222 2014-02-19 Richard Biener <rguenther@suse.de>
5225 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
5226 (ipa_modify_call_arguments): Emit an argument load explicitely and
5227 preserve virtual SSA form there and for the replacement call.
5228 Do not update SSA form nor free dominance info.
5230 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5232 * ipa.c (function_and_variable_visibility): Also clear WEAK
5233 flag when disolving COMDAT_GROUP.
5235 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5237 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
5238 * ipa-prop.c (ipa_set_jf_known_type): Return early when
5240 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
5241 do more sanity checks.
5242 (detect_type_change): Return true when giving up early.
5243 (compute_complex_assign_jump_func): Fix type parameter of
5244 ipa_set_ancestor_jf.
5245 (compute_complex_ancestor_jump_func): Likewise.
5246 (update_jump_functions_after_inlining): Fix updating of
5248 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
5250 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5252 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
5253 inline clones when edge disappears.
5255 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
5258 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
5259 Split 64-bit moves into 2 patterns. Do not allow the use of
5260 direct move for TDmode in little endian, since the decimal value
5261 has little endian bytes within a word, but the 64-bit pieces are
5262 ordered in a big endian fashion, and normal subreg's of TDmode are
5264 (mov<mode>_64bit_dm): Likewise.
5265 (movtd_64bit_nodm): Likewise.
5267 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5269 PR tree-optimization/60174
5270 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
5271 statement of an SSA_NAME that occurs in an abnormal PHI node.
5273 2014-02-18 Jakub Jelinek <jakub@redhat.com>
5276 * final.c (SEEN_BB): Remove.
5277 (SEEN_NOTE, SEEN_EMITTED): Renumber.
5278 (final_scan_insn): Don't force_source_line on second
5279 NOTE_INSN_BASIC_BLOCK.
5281 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
5284 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
5285 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
5286 (type_natural_mode): Warn ABI change when %zmm register is not
5287 available for AVX512F vector value passing.
5289 2014-02-18 Kai Tietz <ktietz@redhat.com>
5292 * config/i386/i386.c (ix86_expand_prologue): Use value in
5293 rax register as displacement when restoring %r10 or %rax.
5294 Fix wrong offset when restoring both registers.
5296 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5298 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
5299 assertion with conditional return.
5301 2014-02-18 Jakub Jelinek <jakub@redhat.com>
5302 Uros Bizjak <ubizjak@gmail.com>
5305 * config/i386/driver-i386.c (host_detect_local_cpu): If
5306 YMM state is not saved by the OS, also clear has_f16c. Move
5307 CPUID 0x80000001 handling before YMM state saving checking.
5309 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
5311 PR rtl-optimization/58960
5312 * haifa-sched.c (alloc_global_sched_pressure_data): New,
5313 factored out from ...
5314 (sched_init): ... here.
5315 (free_global_sched_pressure_data): New, factored out from ...
5316 (sched_finish): ... here.
5317 * sched-int.h (free_global_sched_pressure_data): Declare.
5318 * sched-rgn.c (nr_regions_initial): New static global.
5319 (haifa_find_rgns): Initialize it.
5320 (schedule_region): Disable sched-pressure for the newly
5323 2014-02-17 Richard Biener <rguenther@suse.de>
5325 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
5326 release SSA defs of pattern stmts.
5328 2014-02-17 Richard Biener <rguenther@suse.de>
5330 * tree-inline.c (expand_call_inline): Release the virtual
5331 operand defined by the call we are about to inline.
5333 2014-02-17 Richard Biener <rguenther@suse.de>
5335 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
5337 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
5338 Ilya Tocar <ilya.tocar@intel.com>
5340 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
5341 arguments order in builtin.
5342 (_mm512_permutexvar_epi64): Ditto.
5343 (_mm512_mask_permutexvar_epi64): Ditto
5344 (_mm512_maskz_permutexvar_epi32): Ditto
5345 (_mm512_permutexvar_epi32): Ditto
5346 (_mm512_mask_permutexvar_epi32): Ditto
5348 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5350 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
5351 (p8_vmrgow): Likewise.
5353 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5355 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
5358 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
5361 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
5362 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
5363 into 64-bit and 32-bit moves. On 64-bit moves, add support for
5364 using direct move instructions on ISA 2.07. Also adjust
5365 instruction length for 64-bit.
5366 (mov<mode>_64bit, TFmode/TDmode): Likewise.
5367 (mov<mode>_32bit, TFmode/TDmode): Likewise.
5369 2014-02-15 Alan Modra <amodra@gmail.com>
5373 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
5374 find_replacement on parts of insn rtl that might be reloaded.
5376 2014-02-15 Richard Biener <rguenther@suse.de>
5378 PR tree-optimization/60183
5379 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
5380 (tree_ssa_phiprop): Calculate and free post-dominators.
5382 2014-02-14 Jeff Law <law@redhat.com>
5384 PR rtl-optimization/60131
5385 * ree.c (get_extended_src_reg): New function.
5386 (combine_reaching_defs): Use it rather than assuming location of REG.
5387 (find_and_remove_re): Verify first operand of extension is
5388 a REG before adding the insns to the copy list.
5390 2014-02-14 Roland McGrath <mcgrathr@google.com>
5392 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
5393 * configure: Regenerated.
5394 * config.in: Regenerated.
5395 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
5396 instead of ASM_SHORT.
5398 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
5399 Richard Earnshaw <rearnsha@arm.com>
5401 PR rtl-optimization/59535
5402 * lra-constraints.c (process_alt_operands): Encourage alternative
5403 when unassigned pseudo class is superset of the alternative class.
5404 (inherit_reload_reg): Don't inherit when optimizing for code size.
5405 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
5406 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
5407 modes not less than 4 for Thumb1.
5409 2014-02-14 Kyle McMartin <kyle@redhat.com>
5412 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
5414 2014-02-14 Richard Biener <rguenther@suse.de>
5416 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
5417 (get_frame_arg): Drop the assert with langhook types_compatible_p.
5418 Do not strip INDIRECT_REFs.
5420 2014-02-14 Richard Biener <rguenther@suse.de>
5423 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
5424 DECL_FUNCTION_SPECIFIC_TARGET.
5425 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
5426 * tree-streamer-out.c (pack_ts_target_option): Remove.
5427 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
5428 (write_ts_function_decl_tree_pointers): Do not stream
5429 DECL_FUNCTION_SPECIFIC_TARGET.
5430 * tree-streamer-in.c (unpack_ts_target_option): Remove.
5431 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
5432 (lto_input_ts_function_decl_tree_pointers): Do not stream
5433 DECL_FUNCTION_SPECIFIC_TARGET.
5435 2014-02-14 Jakub Jelinek <jakub@redhat.com>
5437 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
5438 (get_initial_def_for_induction, vectorizable_induction): Ignore
5439 debug stmts when looking for exit_phi.
5440 (vectorizable_live_operation): Fix up condition.
5442 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5444 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
5445 nreverse() because it changes the content of original tree list.
5447 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5449 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
5450 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
5452 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5454 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
5455 GNU coding standards.
5457 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5460 * dwarf2out.c (gen_subprogram_die): Don't call
5461 add_calling_convention_attribute if subr_die is old_die.
5463 2014-02-13 Sharad Singhai <singhai@google.com>
5465 * doc/optinfo.texi: Fix order of nodes.
5467 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
5469 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
5470 operands[2], not operands[3].
5472 2014-02-13 Richard Biener <rguenther@suse.de>
5475 * doc/install.texi (ISL): Update recommended version to 0.12.2,
5476 mention the possibility of an in-tree build.
5477 (CLooG): Update recommended version to 0.18.1, mention the
5478 possibility of an in-tree build and clarify that the ISL
5479 bundled with CLooG does not work.
5481 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5484 * expr.c (compress_float_constant): If x is a hard register,
5485 extend into a pseudo and then move to x.
5487 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
5489 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
5490 caused by bad second argument to warning_at() with -mhotpatch and
5491 nested functions (e.g. with gfortran).
5493 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
5495 * opts.c (option_name): Remove "enabled by default" rider.
5497 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
5499 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
5501 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
5502 Uros Bizjak <ubizjak@gmail.com>
5505 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
5506 * configure: Regenerated.
5508 2014-02-12 Richard Biener <rguenther@suse.de>
5510 * vec.c (vec_prefix::calculate_allocation): Move as
5511 inline variant to vec.h.
5512 (vec_prefix::calculate_allocation_1): New out-of-line version.
5513 * vec.h (vec_prefix::calculate_allocation_1): Declare.
5514 (vec_prefix::m_has_auto_buf): Rename to ...
5515 (vec_prefix::m_using_auto_storage): ... this.
5516 (vec_prefix::calculate_allocation): Inline the easy cases
5517 and dispatch to calculate_allocation_1 which doesn't need the
5519 (va_heap::reserve): Use gcc_checking_assert.
5520 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
5521 m_using_auto_storage.
5522 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
5524 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
5525 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
5526 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
5528 2014-02-12 Richard Biener <rguenther@suse.de>
5530 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
5531 when we found a dependence.
5533 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
5535 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
5537 (maybe_fold_stmt): ... into this new function.
5538 * omp-low.c (lower_omp): Update comment.
5540 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
5543 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
5546 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
5548 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
5549 identifiers in comments.
5550 (cortexa53_extra_costs): Likewise.
5551 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
5552 (cortexa7_extra_costs): Likewise.
5553 (cortexa12_extra_costs): Likewise.
5554 (cortexa15_extra_costs): Likewise.
5555 (v7m_extra_costs): Likewise.
5557 2014-02-12 Richard Biener <rguenther@suse.de>
5560 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
5561 of posix_memalign being successful.
5562 (lower_stmt): Restrict lowering of posix_memalign to when
5563 -ftree-bit-ccp is enabled.
5565 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5567 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
5569 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
5571 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
5573 PR rtl-optimization/60116
5574 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
5575 other_insn once the combination has been validated.
5577 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
5580 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
5582 * ipa-devirt.c: Include demangle.h
5583 (odr_violation_reported): New static variable.
5584 (add_type_duplicate): Update odr_violations.
5585 (maybe_record_node): Add completep parameter; update it.
5586 (record_target_from_binfo): Add COMPLETEP parameter;
5587 update it as needed.
5588 (possible_polymorphic_call_targets_1): Likewise.
5589 (struct polymorphic_call_target_d): Add nonconstruction_targets;
5590 rename FINAL to COMPLETE.
5591 (record_targets_from_bases): Sanity check we found the binfo;
5592 fix COMPLETEP updating.
5593 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
5594 parameter, fix computing of COMPLETEP.
5595 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
5596 at LTO time do demangling.
5597 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
5598 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
5600 (gimple_get_virt_method_for_binfo): Likewise.
5601 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5602 gimple_get_virt_method_for_vtable): Update prototypes.
5604 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5607 * genautomata.c (add_presence_absence): Fix typo with
5608 {final_}presence_list.
5610 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5613 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5614 for VSX/Altivec vectors that land in GPR registers.
5616 2014-02-11 Richard Henderson <rth@redhat.com>
5617 Jakub Jelinek <jakub@redhat.com>
5620 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5621 around drhs if type conversion to lacc->type is not useless.
5623 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5625 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5627 (cortex-a57.cortex-a53): Likewise.
5628 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5630 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5632 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5635 2014-02-11 Renlin Li <Renlin.Li@arm.com>
5637 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5638 add_options_for_arm_vfp3.
5640 2014-02-11 Jeff Law <law@redhat.com>
5643 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5644 object with an undesirable mode.
5646 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5649 * config/i386/sol2-9.h: New file.
5650 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5651 *-*-solaris2.9*): Use it.
5653 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5655 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5656 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5658 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5660 * config/microblaze/microblaze.c: Extend mcpu version format
5662 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
5664 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5666 2014-02-10 Richard Henderson <rth@redhat.com>
5669 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5670 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5671 ms-abi vs -mno-accumulate-outgoing-args.
5672 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5673 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5676 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5679 * cfgexpand.c (expand_asm_operands): Attach source location to
5680 ASM_INPUT rtx objects.
5681 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5683 2014-02-10 Nick Clifton <nickc@redhat.com>
5685 * config/mn10300/mn10300.c (popcount): New function.
5686 (mn10300_expand_prologue): Include saved registers in stack usage
5689 2014-02-10 Jeff Law <law@redhat.com>
5692 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
5693 when changing the SET_DEST of a prior insn to avoid an input reload.
5695 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5697 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5698 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5699 -mcall-openbsd, or -mcall-linux.
5700 (CC1_ENDIAN_BIG_SPEC): Remove.
5701 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5702 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5703 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5704 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5705 and %cc1_endian_default.
5706 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5708 2014-02-10 Richard Biener <rguenther@suse.de>
5710 PR tree-optimization/60115
5711 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5712 MEM_REF handling. Properly verify that the accesses are not
5713 out of the objects bound.
5715 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5717 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5720 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5722 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5723 proper constants and fix formatting.
5724 (possible_polymorphic_call_targets): Fix formatting.
5726 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5727 Ilya Tocar <ilya.tocar@intel.com>
5729 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5730 (_mm512_loadu_epi32): Renamed into...
5731 (_mm512_loadu_si512): This.
5732 (_mm512_storeu_epi32): Renamed into...
5733 (_mm512_storeu_si512): This.
5734 (_mm512_maskz_ceil_ps): Removed.
5735 (_mm512_maskz_ceil_pd): Ditto.
5736 (_mm512_maskz_floor_ps): Ditto.
5737 (_mm512_maskz_floor_pd): Ditto.
5738 (_mm512_floor_round_ps): Ditto.
5739 (_mm512_floor_round_pd): Ditto.
5740 (_mm512_ceil_round_ps): Ditto.
5741 (_mm512_ceil_round_pd): Ditto.
5742 (_mm512_mask_floor_round_ps): Ditto.
5743 (_mm512_mask_floor_round_pd): Ditto.
5744 (_mm512_mask_ceil_round_ps): Ditto.
5745 (_mm512_mask_ceil_round_pd): Ditto.
5746 (_mm512_maskz_floor_round_ps): Ditto.
5747 (_mm512_maskz_floor_round_pd): Ditto.
5748 (_mm512_maskz_ceil_round_ps): Ditto.
5749 (_mm512_maskz_ceil_round_pd): Ditto.
5750 (_mm512_expand_pd): Ditto.
5751 (_mm512_expand_ps): Ditto.
5752 * config/i386/i386.c (ix86_builtins): Remove
5753 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5754 (bdesc_args): Ditto.
5755 * config/i386/predicates.md (const1256_operand): New.
5756 (const_1_to_2_operand): Ditto.
5757 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5758 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5759 (*avx512pf_gatherpf<mode>sf): Ditto.
5760 (avx512pf_gatherpf<mode>df): Ditto.
5761 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5762 (*avx512pf_gatherpf<mode>df): Ditto.
5763 (avx512pf_scatterpf<mode>sf): Ditto.
5764 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5765 (*avx512pf_scatterpf<mode>sf): Ditto.
5766 (avx512pf_scatterpf<mode>df): Ditto.
5767 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5768 (*avx512pf_scatterpf<mode>df): Ditto.
5769 (avx512f_expand<mode>): Removed.
5770 (<shift_insn><mode>3<mask_name>): Change predicate type.
5772 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5774 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5775 not at the end of datarefs vector use ordered_remove to avoid
5776 reordering datarefs vector.
5779 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5780 mark local addressable non-static vars as GOVD_PRIVATE
5781 instead of GOVD_LOCAL.
5782 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5783 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5787 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5788 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5789 assume_aligned or alloc_align attributes.
5790 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5791 arguments. Handle also assume_aligned and alloc_align attributes.
5792 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5793 calls to functions with assume_aligned or alloc_align attributes.
5794 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
5796 2014-02-08 Terry Guo <terry.guo@arm.com>
5798 * doc/invoke.texi: Document ARM -march=armv7e-m.
5800 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5802 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5803 flag on __cilkrts_rethrow builtin.
5806 * ipa-cp.c (determine_versionability): Fail at -O0
5807 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5808 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5811 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5814 * tree-inline.c (copy_forbidden): Fail for
5815 __attribute__((optimize (0))) functions.
5817 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5819 * varpool.c: Include pointer-set.h.
5820 (varpool_remove_unreferenced_decls): Variables in other partitions
5821 will not be output; be however careful to not lose information
5824 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5826 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5827 lookup in the vtable constructor.
5829 2014-02-07 Jeff Law <law@redhat.com>
5832 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5833 define_insn_and_split.
5835 * ipa-inline.c (inline_small_functions): Fix typos.
5837 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5839 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5840 (s390_can_use_return_insn): Declare.
5841 * config/s390/s390.h (EPILOGUE_USES): Define.
5842 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5844 (s390_chunkify_start): Handle return JUMP_LABELs.
5845 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5846 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5847 (s390_can_use_return_insn): New functions.
5848 (s390_fix_long_loop_prediction): Handle conditional returns.
5849 (TARGET_SET_UP_BY_PROLOGUE): Define.
5850 * config/s390/s390.md (ANY_RETURN): New code iterator.
5851 (*creturn, *csimple_return, return, simple_return): New patterns.
5853 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5855 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5856 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5857 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5858 REG_CFA_RESTORE list when deciding not to restore a register.
5860 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5862 * config/s390/s390.c: Include tree-pass.h and context.h.
5863 (s390_early_mach): New function, split out from...
5864 (s390_emit_prologue): ...here.
5865 (pass_data_s390_early_mach): New pass structure.
5866 (pass_s390_early_mach): New class.
5867 (s390_option_override): Create and register early_mach pass.
5868 Move to end of file.
5870 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5872 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5873 to match for the exit block.
5875 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5877 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5878 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5879 Reject misaligned operands.
5881 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5883 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
5885 2014-02-07 Richard Biener <rguenther@suse.de>
5888 * gimple-low.c (lower_builtin_posix_memalign): New function.
5889 (lower_stmt): Call it to lower posix_memalign in a way
5890 to make alignment info accessible.
5892 2014-02-07 Jakub Jelinek <jakub@redhat.com>
5895 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5896 __builtin_setjmp_receiver.
5898 2014-02-07 Richard Biener <rguenther@suse.de>
5901 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5902 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5903 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5904 Handle BUILT_IN_POSIX_MEMALIGN.
5905 (find_func_clobbers): Likewise.
5906 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5907 (call_may_clobber_ref_p_1): Likewise.
5909 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5912 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5915 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5918 * lto-cgraph.c (lto_output_node): Use
5919 symtab_get_symbol_partitioning_class.
5920 (lto_output_varpool_node): likewise.
5921 (symtab_get_symbol_partitioning_class): Move here from
5923 * cgraph.h (symbol_partitioning_class): Likewise.
5924 (symtab_get_symbol_partitioning_class): Declare.
5926 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5928 * ggc.h (ggc_internal_cleared_alloc): New macro.
5929 * vec.h (vec_safe_copy): Handle memory stats.
5930 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5931 * target-globals.c (save_target_globals): Likewise.
5933 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5936 * expr.c (emit_move_resolve_push): Export; be bit more selective
5937 on when to clear alias set.
5938 * expr.h (emit_move_resolve_push): Declare.
5939 * function.h (struct function): Add tail_call_marked.
5940 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5941 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5942 * config/i386/i386.md (TImode move expander): De not call
5944 (FP push expanders): Preserve memory attributes.
5945 * config/i386/sse.md (push<mode>1): Remove.
5946 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
5947 (ix86_expand_push): Remove.
5948 * config/i386/mmx.md (push<mode>1): Remove.
5950 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5952 PR rtl-optimization/60030
5953 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5954 lopart with paradoxical subreg before shifting it up by hprec.
5956 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5958 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5959 Remove extra newline at end of file.
5960 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5961 (arm_issue_rate): Handle cortexa57.
5962 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5963 (cortex-a57.cortex-a53): Likewise.
5965 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5968 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5969 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5971 (arm_expand_prologue): Adjust all callers.
5972 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5973 info, registers also at the lowest numbered registers side. Use
5974 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5978 * var-tracking.c (adjust_mems): Before adding a SET to
5979 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
5981 2014-02-06 Alan Modra <amodra@gmail.com>
5984 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
5985 change SDmode to DDmode when lra_in_progress.
5987 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5990 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
5991 free_data_ref on the dr first, and before goto again also set dr
5992 to the next dr. For simd_lane_access, free old datarefs[i] before
5993 overwriting it. For get_vectype_for_scalar_type failure, don't
5994 free_data_ref if simd_lane_access.
5996 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
5999 * tree.h (opts_for_fn): New inline function.
6000 (opt_for_fn): Define.
6001 * config/i386/i386.c (ix86_function_regparm): Use
6002 opt_for_fn (decl, optimize) instead of optimize.
6004 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
6006 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
6007 for SYMBOL_REF in large memory model.
6009 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6011 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
6013 (cortex-a57): Likewise.
6014 (cortex-a57.cortex-a53): Likewise.
6016 2014-02-06 Yury Gribov <y.gribov@samsung.com>
6017 Kugan Vivekanandarajah <kuganv@linaro.org>
6019 * config/arm/arm.c (arm_vector_alignment_reachable): Check
6021 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
6023 2014-02-06 Richard Biener <rguenther@suse.de>
6025 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
6026 set_loop_copy and initialize_original_copy_tables.
6028 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
6030 * config/aarch64/aarch64-simd.md
6031 (aarch64_ashr_simddi): Change QI to SI.
6033 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
6034 Jakub Jelinek <jakub@redhat.com>
6037 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
6040 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6042 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
6043 CODE_FOR_altivec_vpku[hw]um to
6044 CODE_FOR_altivec_vpku[hw]um_direct.
6045 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
6046 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
6047 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
6048 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
6050 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6052 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
6053 generation for -maltivec=be.
6054 (altivec_vsumsws): Simplify redundant test.
6056 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6058 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
6059 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
6060 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
6061 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
6062 gen_altivec_vpkuwum.
6063 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
6065 (altivec_vpks<VI_char>ss): Likewise.
6066 (altivec_vpks<VI_char>us): Likewise.
6067 (altivec_vpku<VI_char>us): Likewise.
6068 (altivec_vpku<VI_char>um): Likewise.
6069 (altivec_vpku<VI_char>um_direct): New (copy of
6070 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
6072 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
6073 target is little endian and -maltivec=be is not specified.
6074 (*altivec_vupkhs<VU_char>_direct): New (copy of
6075 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
6076 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
6077 target is little endian and -maltivec=be is not specified.
6078 (*altivec_vupkls<VU_char>_direct): New (copy of
6079 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
6080 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
6081 little endian and -maltivec=be is not specified.
6082 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
6083 little endian and -maltivec=be is not specified.
6085 2014-02-05 Richard Henderson <rth@redhat.com>
6088 * combine-stack-adj.c: Revert r206943.
6089 * sched-int.h (struct deps_desc): Add last_args_size.
6090 * sched-deps.c (init_deps): Initialize it.
6091 (sched_analyze_insn): Add OUTPUT dependencies between insns that
6092 contain REG_ARGS_SIZE notes.
6094 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
6096 * lto-cgraph.c (asm_nodes_output): Make global.
6097 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
6098 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
6099 (driver_handle_option): Handle OPT_fwpa.
6101 2014-02-05 Jakub Jelinek <jakub@redhat.com>
6104 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
6105 a comment typo and formatting issue. If odr_hash hasn't been
6106 created, return vNULL and set *completep to false.
6109 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
6110 bb with no successors.
6112 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
6115 * doc/invoke.texi (-march): Clarify documentation for ARM.
6119 2014-02-05 Richard Biener <rguenther@suse.de>
6121 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
6122 when not vectorizing because of too many alias checks.
6123 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
6124 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
6126 2014-02-05 Nick Clifton <nickc@redhat.com>
6128 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
6129 accept extended registers in any mode when compiling for the MN10300.
6131 2014-02-05 Yury Gribov <y.gribov@samsung.com>
6133 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
6134 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
6135 sanitization attributes.
6136 (can_inline_edge_p): Likewise.
6137 (sanitize_attrs_match_for_inline_p): New function.
6139 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
6141 * ipa-prop.c (detect_type_change): Shor circuit testing of
6142 type changes on THIS pointer.
6144 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
6147 * config/pa/pa.c (legitimize_tls_address): Return original address
6148 if not passed a SYMBOL_REF rtx.
6149 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
6151 (pa_emit_move_sequence): Simplify TLS source operands.
6152 (pa_legitimate_constant_p): Reject all TLS constants.
6153 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
6154 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
6156 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
6158 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
6159 groups when we know they are controlled by LTO.
6160 * varasm.c (default_binds_local_p_1): If object is in other partition,
6161 it will be resolved locally.
6163 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
6165 * config/host-linux.c (linux_gt_pch_use_address): Don't
6166 use SSIZE_MAX because it is not always defined.
6168 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
6171 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
6172 threshold for pseudo splitting.
6173 (update_ebb_live_info): Process call argument hard registers and
6174 hard registers from insn definition too.
6175 (max_small_class_regs_num): New constant.
6176 (inherit_in_ebb): Update live hard regs through EBBs. Update
6177 reloads_num only for small register classes. Don't split for
6180 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
6183 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
6186 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
6188 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
6191 2014-02-04 Marek Polacek <polacek@redhat.com>
6193 * gdbinit.in (pel): Define.
6195 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
6197 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
6200 2014-02-04 Richard Biener <rguenther@suse.de>
6203 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
6204 in function context local.
6205 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
6206 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
6207 similar to LTO_imported_decl_ref.
6209 2014-02-04 Jakub Jelinek <jakub@redhat.com>
6211 PR tree-optimization/60002
6212 * cgraphclones.c (build_function_decl_skip_args): Clear
6215 PR tree-optimization/60023
6216 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
6217 false to gsi_replace.
6218 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
6219 has been in some EH region and vec_stmt could throw, add
6220 vec_stmt into the same EH region.
6221 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
6222 has no lhs, ignore it.
6223 * internal-fn.c (expand_MASK_LOAD): Likewise.
6226 * tree-inline.c (copy_forbidden): Fail for
6227 __attribute__((optimize (0))) functions.
6230 * omp-low.c (simd_clone_struct_copy): If from->inbranch
6231 is set, copy one less argument.
6232 (expand_simd_clones): Don't subtract clone_info->inbranch
6233 from simd_clone_struct_alloc argument.
6235 PR rtl-optimization/57915
6236 * recog.c (simplify_while_replacing): If all unary/binary/relational
6237 operation arguments are constant, attempt to simplify those.
6240 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
6241 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
6243 2014-02-04 Richard Biener <rguenther@suse.de>
6245 PR tree-optimization/60012
6246 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
6247 TBAA disambiguation to all DDRs.
6249 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6252 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
6253 (LINK_SPEC): Use it for -shared, -shared-libgcc.
6255 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6258 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
6260 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6262 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
6263 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
6265 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6268 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
6269 to figure out targets of polymorphic calls with known decl.
6270 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6271 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
6272 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
6273 (get_polymorphic_call_info): ... here.
6274 (get_polymorphic_call_info_from_invariant): New function.
6276 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6278 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
6279 lookup via vtable pointer; check for type consistency
6280 and turn inconsitent facts into UNREACHABLE.
6281 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6282 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
6283 type inconsistent querries; return UNREACHABLE instead.
6285 2014-02-03 Richard Henderson <rth@twiddle.net>
6288 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
6289 already processed this node.
6290 (normalize_one_pred_1): Pass along mark_set.
6291 (normalize_one_pred): Create and destroy a pointer_set_t.
6292 (normalize_one_pred_chain): Likewise.
6294 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
6296 PR gcov-profile/58602
6297 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
6299 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6302 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
6303 -fno-devirtualize; try to devirtualize by the knowledge of
6304 virtual table pointer given by aggregate propagation.
6305 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6306 (ipa_print_node_jump_functions): Dump also offset that
6307 is relevant for polymorphic calls.
6308 (determine_known_aggregate_parts): Add arg_type parameter; use it
6309 instead of determining the type from pointer type.
6310 (ipa_compute_jump_functions_for_edge): Update call of
6311 determine_known_aggregate_parts.
6312 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
6313 (gimple_get_virt_method_for_binfo): ... here; simplify using
6314 vtable_pointer_value_to_vtable.
6315 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
6316 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
6317 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
6318 (vtable_pointer_value_to_vtable): Break out from ...; handle also
6320 (vtable_pointer_value_to_binfo): ... here.
6321 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
6323 2014-02-03 Teresa Johnson <tejohnson@google.com>
6325 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
6326 redef of outer loop index variable.
6328 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
6332 * doc/extend.texi (Function Attributes): Typo.
6334 2014-02-03 Cong Hou <congh@google.com>
6336 PR tree-optimization/60000
6337 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
6338 if the vectorized statement is a store. A store statement can only
6339 appear at the end of pattern statements.
6341 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6343 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
6344 (ix86_option_override_internal): Default long double to 64-bit for
6345 32-bit Bionic and to 128-bit for 64-bit Bionic.
6347 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
6348 TARGET_LONG_DOUBLE_128 is true.
6349 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
6351 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
6352 (mlong-double-64): Negate -mlong-double-128.
6353 (mlong-double-128): New option.
6355 * config/i386/i386-c.c (ix86_target_macros): Define
6356 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
6358 * doc/invoke.texi: Document -mlong-double-128.
6360 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6362 PR rtl-optimization/60024
6363 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
6365 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
6367 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
6369 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6371 PR rtl-optimization/57662
6372 * sel-sched.c (code_motion_path_driver): Do not mark already not
6373 existing blocks in the visiting bitmap.
6375 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6377 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
6378 on the insn being emitted.
6380 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
6381 Will Deacon <will.deacon@arm.com>
6383 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
6385 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6387 * config/arm/arm-tables.opt: Regenerate.
6389 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6391 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
6392 for vector types other than V16QImode.
6393 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
6394 define_expand, and call altivec_expand_vec_perm_le when producing
6395 code with little endian element order.
6396 (*altivec_vperm_<mode>_internal): New insn having previous
6397 behavior of altivec_vperm_<mode>.
6398 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
6399 altivec_expand_vec_perm_le when producing code with little endian
6401 (*altivec_vperm_<mode>_uns_internal): New insn having previous
6402 behavior of altivec_vperm_<mode>_uns.
6404 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6406 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
6407 (altivec_vsumsws): Add handling for -maltivec=be with a little
6409 (altivec_vsumsws_direct): New.
6410 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
6411 gen_altivec_vsumsws.
6413 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6415 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
6416 vtable_pointer_value_to_binfo): New functions.
6417 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
6418 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
6420 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
6422 * config/nios2/nios2.md (load_got_register): Initialize GOT
6423 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
6424 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
6426 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6428 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
6429 preserverd by passthrough, do not propagate the type.
6431 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6433 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
6434 (mips_atomic_assign_expand_fenv): New function.
6435 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
6437 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6439 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
6440 (__builtin_mips_set_fcsr): Likewise.
6441 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
6442 MIPS_USI_FTYPE_VOID.
6443 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
6444 (mips16_expand_set_fcsr): Likewise.
6445 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
6446 (mips16_set_fcsr_stub): Likewise.
6447 (mips16_get_fcsr_one_only_stub): New class.
6448 (mips16_set_fcsr_one_only_stub): Likewise.
6449 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
6450 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
6451 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
6452 (hard_float): New availability predicate.
6453 (mips_builtins): Add get_fcsr and set_fcsr.
6454 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
6455 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
6456 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
6457 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
6458 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
6461 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6463 * config/mips/mips.c (mips_one_only_stub): New class.
6464 (mips_need_mips16_rdhwr_p): Replace with...
6465 (mips16_rdhwr_stub): ...this new variable.
6466 (mips16_stub_call_address): New function.
6467 (mips16_rdhwr_one_only_stub): New class.
6468 (mips_expand_thread_pointer): Use mips16_stub_call_address.
6469 (mips_output_mips16_rdhwr): Delete.
6470 (mips_finish_stub): New function.
6471 (mips_code_end): Use it to handle rdhwr stubs.
6473 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
6476 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
6477 when calculating size of integer atomic types.
6479 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
6481 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
6483 2014-02-01 Jakub Jelinek <jakub@redhat.com>
6485 PR tree-optimization/60003
6486 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
6487 * profile.c (branch_prob): Use gimple_call_builtin_p
6488 to check for BUILT_IN_SETJMP_RECEIVER.
6489 * tree-inline.c (copy_bb): Call notice_special_calls.
6491 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
6494 * lra-constraints.c (process_alt_operands): Update reload_sum only
6497 2014-01-31 Richard Henderson <rth@redhat.com>
6500 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
6501 until after else_eh is processed.
6503 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6505 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
6506 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
6507 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
6508 in smmintrin.h, remove them.
6509 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
6510 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
6511 * config/i386/i386.md (ROUND_SAE): Fix value.
6512 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
6513 (const48_operand): New.
6514 * config/i386/subst.md (round), (round_expand): Use
6515 const_4_or_8_to_11_operand.
6516 (round_saeonly), (round_saeonly_expand): Use const48_operand.
6518 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6520 * config/i386/constraints.md (Yk): Swap meaning with k.
6521 * config/i386/i386.md (movhi_internal): Change Yk to k.
6522 (movqi_internal): Ditto.
6523 (*k<logic><mode>): Ditto.
6526 (kandn<mode>): Ditto.
6527 (*<code>hi_1): Ditto.
6528 (*<code>qi_1): Ditto.
6529 (kxnor<mode>): Ditto.
6530 (kortestzhi): Ditto.
6531 (kortestchi): Ditto.
6533 (*one_cmplhi2_1): Ditto.
6534 (*one_cmplqi2_1): Ditto.
6535 * config/i386/sse.md (): Change k to Yk.
6536 (avx512f_load<mode>_mask): Ditto.
6537 (avx512f_blendm<mode>): Ditto.
6538 (avx512f_store<mode>_mask): Ditto.
6539 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
6540 (avx512f_storedqu<mode>_mask): Ditto.
6541 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
6543 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
6544 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
6545 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
6546 (avx512f_maskcmp<mode>3): Ditto.
6547 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
6548 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
6549 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
6550 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
6551 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
6552 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
6553 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
6554 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
6555 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
6556 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
6557 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
6558 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
6559 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
6560 (vec_extract_lo_<mode>_maskm): Ditto.
6561 (vec_extract_hi_<mode>_maskm): Ditto.
6562 (avx512f_vternlog<mode>_mask): Ditto.
6563 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
6564 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
6565 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
6566 (avx512f_<code>v8div16qi2_mask): Ditto.
6567 (avx512f_<code>v8div16qi2_mask_store): Ditto.
6568 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
6569 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
6570 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
6571 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
6572 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6573 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6574 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6575 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6576 (avx512cd_maskb_vec_dupv8di): Ditto.
6577 (avx512cd_maskw_vec_dupv16si): Ditto.
6578 (avx512f_vpermi2var<mode>3_maskz): Ditto.
6579 (avx512f_vpermi2var<mode>3_mask): Ditto.
6580 (avx512f_vpermi2var<mode>3_mask): Ditto.
6581 (avx512f_vpermt2var<mode>3_maskz): Ditto.
6582 (*avx512f_gathersi<mode>): Ditto.
6583 (*avx512f_gathersi<mode>_2): Ditto.
6584 (*avx512f_gatherdi<mode>): Ditto.
6585 (*avx512f_gatherdi<mode>_2): Ditto.
6586 (*avx512f_scattersi<mode>): Ditto.
6587 (*avx512f_scatterdi<mode>): Ditto.
6588 (avx512f_compress<mode>_mask): Ditto.
6589 (avx512f_compressstore<mode>_mask): Ditto.
6590 (avx512f_expand<mode>_mask): Ditto.
6591 * config/i386/subst.md (mask): Change k to Yk.
6592 (mask_scalar_merge): Ditto.
6595 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
6597 * doc/extend.texi (Vector Extensions): Document ?: in C++.
6599 2014-01-31 Richard Biener <rguenther@suse.de>
6602 * builtins.c (fold_builtin_memory_op): Make sure to not
6603 use a floating-point mode or a boolean or enumeral type for
6606 2014-01-30 DJ Delorie <dj@redhat.com>
6608 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6609 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6610 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6611 whenever main() has an epilogue.
6613 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6615 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6616 unused variable "field".
6617 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6618 (vsx_mergeh_<mode>): Likewise.
6619 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6620 (altivec_vmrghh): Likewise.
6621 (altivec_vmrghw): Likewise.
6622 (altivec_vmrglb): Likewise.
6623 (altivec_vmrglh): Likewise.
6624 (altivec_vmrglw): Likewise.
6625 (altivec_vspltb): Add missing uses.
6626 (altivec_vsplth): Likewise.
6627 (altivec_vspltw): Likewise.
6628 (altivec_vspltsf): Likewise.
6630 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6633 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6634 frame related instructions.
6636 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6638 PR rtl-optimization/59959
6639 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6640 any reload of register whose subreg is invalid.
6642 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6644 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6645 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6646 Add missing return type - void.
6648 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6650 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6651 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6652 remove element index adjustment for endian (now handled in vsx.md
6654 (altivec_expand_vec_perm_const): Use
6655 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6656 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6657 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6658 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6659 define_expand and a new define_insn *altivec_vspltb_internal;
6660 adjust for -maltivec=be on a little endian target.
6661 (altivec_vspltb_direct): New.
6662 (altivec_vsplth): Divide into a define_expand and a new
6663 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6664 little endian target.
6665 (altivec_vsplth_direct): New.
6666 (altivec_vspltw): Divide into a define_expand and a new
6667 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6668 little endian target.
6669 (altivec_vspltw_direct): New.
6670 (altivec_vspltsf): Divide into a define_expand and a new
6671 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6672 a little endian target.
6674 2014-01-30 Richard Biener <rguenther@suse.de>
6676 PR tree-optimization/59993
6677 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6678 can propagate form the earlier stmt and avoid the transform
6679 when the intermediate result is needed.
6681 2014-01-30 Alangi Derick <alangiderick@gmail.com>
6683 * README.Portability: Fix typo.
6685 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
6687 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6688 comparison_operator with ordered_comparison_operator.
6690 2014-01-30 Nick Clifton <nickc@redhat.com>
6692 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6693 Rename to mn10300_store_multiple_regs.
6694 * config/mn10300/mn10300.c: Likewise.
6695 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6696 store_multiple_regs.
6697 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6698 Call mn10300_store_multiple_regs.
6700 2014-01-30 Nick Clifton <nickc@redhat.com>
6701 DJ Delorie <dj@redhat.com>
6703 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6704 %fp 2 to keep registers after it properly word-aligned.
6705 (rl78_alloc_physical_registers_umul): Handle the case where both
6706 input operands are the same.
6708 2014-01-30 Richard Biener <rguenther@suse.de>
6710 PR tree-optimization/59903
6711 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6714 2014-01-30 Jason Merrill <jason@redhat.com>
6717 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6720 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6722 2014-01-30 Richard Biener <rguenther@suse.de>
6724 PR tree-optimization/59951
6725 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
6727 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6730 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6731 SFmode to DFmode case.
6733 2014-01-29 DJ Delorie <dj@redhat.com>
6735 * config/msp430/msp430.opt (-minrt): New.
6736 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6738 (ENDFILE_SPEC): Likewise.
6740 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
6742 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6743 (estimate_function_body_sizes): Use it.
6745 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6748 * dwarf2out.c (is_cxx_auto): New.
6749 (is_base_type): Use it.
6750 (gen_type_die_with_usage): Likewise.
6752 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6754 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6755 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6756 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6757 -maltivec=be with LE targets.
6758 (vsx_mergeh_<mode>): Likewise.
6759 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
6760 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6761 (altivec_vmrghb): Replace with define_expand and new
6762 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
6763 (altivec_vmrghb_direct): New define_insn.
6764 (altivec_vmrghh): Replace with define_expand and new
6765 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
6766 (altivec_vmrghh_direct): New define_insn.
6767 (altivec_vmrghw): Replace with define_expand and new
6768 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
6769 (altivec_vmrghw_direct): New define_insn.
6770 (*altivec_vmrghsf): Adjust for endianness.
6771 (altivec_vmrglb): Replace with define_expand and new
6772 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
6773 (altivec_vmrglb_direct): New define_insn.
6774 (altivec_vmrglh): Replace with define_expand and new
6775 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
6776 (altivec_vmrglh_direct): New define_insn.
6777 (altivec_vmrglw): Replace with define_expand and new
6778 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
6779 (altivec_vmrglw_direct): New define_insn.
6780 (*altivec_vmrglsf): Adjust for endianness.
6781 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6782 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6783 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6784 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6785 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6786 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6787 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6788 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6790 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6792 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6793 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6796 2014-01-29 Richard Biener <rguenther@suse.de>
6798 PR tree-optimization/58742
6799 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6800 associate_pointerplus_align.
6801 (associate_pointerplus_diff): New function.
6802 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6803 and associate_pointerplus_diff.
6805 2014-01-29 Richard Biener <rguenther@suse.de>
6807 * lto-streamer.h (LTO_major_version): Bump to 3.
6808 (LTO_minor_version): Reset to 0.
6810 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6812 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6813 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6814 (arm_file_start): Generate correct asm header for armv7ve.
6815 * config/arm/bpabi.h: Add multilib support for armv7ve.
6816 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6817 and cortex-a15 to armv7ve.
6818 * config/arm/t-aprofile: Add multilib support for armv7ve.
6819 * doc/invoke.texi: Document -march=armv7ve.
6821 2014-01-29 Richard Biener <rguenther@suse.de>
6823 PR tree-optimization/58742
6824 * tree-ssa-forwprop.c (associate_plusminus): Return true
6825 if we changed sth, defer EH cleanup to ...
6826 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6827 (simplify_mult): New function.
6829 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6832 PR tree-optimization/59920
6833 * tree.c (build_common_builtin_nodes): Remove
6834 __builtin_setjmp_dispatcher initialization.
6835 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6836 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6837 instead of gsi_after_labels + manually skipping debug stmts.
6838 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6839 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6840 * tree-inline.c (copy_edges_for_bb): Remove
6841 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6842 argument. Ignore computed_goto_p stmts. Don't call
6843 make_abnormal_goto_edges. If a call might need abnormal edges
6844 for non-local gotos, see if it already has an edge to
6845 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6846 with true argument, don't do anything then, otherwise add
6847 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6848 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6850 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6851 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6852 (lower_stmt): Don't set data->calls_builtin_setjmp.
6853 (lower_builtin_setjmp): Adjust comment.
6854 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6855 * tree-cfg.c (found_computed_goto): Remove.
6856 (factor_computed_gotos): Remove.
6857 (make_goto_expr_edges): Return bool, true for computed gotos.
6858 Don't call make_abnormal_goto_edges.
6859 (build_gimple_cfg): Don't set found_computed_goto, don't call
6860 factor_computed_gotos.
6861 (computed_goto_p): No longer static.
6862 (make_blocks): Don't set found_computed_goto.
6863 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6864 (make_edges): If make_goto_expr_edges returns true, push bb
6865 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6866 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6867 vector. Record mapping between bbs and OpenMP regions if there
6868 are any, adjust make_gimple_omp_edges caller. Call
6869 handle_abnormal_edges.
6870 (make_abnormal_goto_edges): Remove.
6871 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6872 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6873 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6874 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6875 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6876 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6877 filling *region also set *region_idx to (*region)->entry->index.
6880 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6881 For REGs set ORIGINAL_REGNO.
6883 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
6885 * doc/md.texi: Mention that a target shouldn't implement
6886 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6889 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6891 PR tree-optimization/59594
6892 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6893 a copy of the datarefs vector rather than the vector itself.
6895 2014-01-28 Jason Merrill <jason@redhat.com>
6898 * dwarf2out.c (auto_die): New static.
6899 (gen_type_die_with_usage): Handle C++1y 'auto'.
6900 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6903 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6906 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6907 (SPEC_X32): Likewise.
6908 (SPEC_64): Likewise.
6909 * config/i386/i386.c (ix86_option_override_internal): Turn off
6910 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6912 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6913 * config/i386/i386.h (TARGET_16BIT): New macro.
6914 (TARGET_16BIT_P): Likewise.
6915 * config/i386/i386.opt: Add m16.
6916 * doc/invoke.texi: Document -m16.
6918 2014-01-28 Jakub Jelinek <jakub@redhat.com>
6920 PR preprocessor/59935
6921 * input.c (location_get_source_line): Bail out on when line number
6922 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
6924 2014-01-28 Richard Biener <rguenther@suse.de>
6926 PR tree-optimization/58742
6927 * tree-ssa-forwprop.c (associate_plusminus): Handle
6928 pointer subtraction of the form (T)(P + A) - (T)P.
6930 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6932 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6935 2014-01-28 Richard Biener <rguenther@suse.de>
6938 2014-01-28 Richard Biener <rguenther@suse.de>
6940 PR rtl-optimization/45364
6941 PR rtl-optimization/59890
6942 * var-tracking.c (local_get_addr_clear_given_value): Handle
6943 already cleared slot.
6944 (val_reset): Handle not allocated local_get_addr_cache.
6945 (vt_find_locations): Use post-order on the inverted CFG.
6947 2014-01-28 Richard Biener <rguenther@suse.de>
6949 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6951 2014-01-28 Richard Biener <rguenther@suse.de>
6953 PR rtl-optimization/45364
6954 PR rtl-optimization/59890
6955 * var-tracking.c (local_get_addr_clear_given_value): Handle
6956 already cleared slot.
6957 (val_reset): Handle not allocated local_get_addr_cache.
6958 (vt_find_locations): Use post-order on the inverted CFG.
6960 2014-01-28 Alan Modra <amodra@gmail.com>
6962 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6963 * configure.ac <recursive call for build != host>: Define
6964 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6965 and LD_FOR_BUILD too.
6966 * configure: Regenerate.
6968 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6970 * config/i386/i386.c (get_builtin_code_for_version): Separate
6971 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6972 Broadwell from Haswell.
6974 2014-01-27 Steve Ellcey <sellcey@mips.com>
6976 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6977 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6978 * config/mips/mips.c (mips_option_override): Change setting
6980 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
6981 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
6982 Change from Mask to Var.
6984 2014-01-27 Jeff Law <law@redhat.com>
6986 * ipa-inline.c (inline_small_functions): Fix typo.
6988 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
6990 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
6991 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
6992 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
6993 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
6994 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
6995 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
6996 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
6997 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
6998 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
6999 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
7000 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
7001 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
7002 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
7003 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
7004 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
7005 (_mm512_storeu_epi64): Ditto.
7006 (_mm512_cmpge_epi32_mask): Ditto.
7007 (_mm512_cmpge_epu32_mask): Ditto.
7008 (_mm512_cmpge_epi64_mask): Ditto.
7009 (_mm512_cmpge_epu64_mask): Ditto.
7010 (_mm512_cmple_epi32_mask): Ditto.
7011 (_mm512_cmple_epu32_mask): Ditto.
7012 (_mm512_cmple_epi64_mask): Ditto.
7013 (_mm512_cmple_epu64_mask): Ditto.
7014 (_mm512_cmplt_epi32_mask): Ditto.
7015 (_mm512_cmplt_epu32_mask): Ditto.
7016 (_mm512_cmplt_epi64_mask): Ditto.
7017 (_mm512_cmplt_epu64_mask): Ditto.
7018 (_mm512_cmpneq_epi32_mask): Ditto.
7019 (_mm512_cmpneq_epu32_mask): Ditto.
7020 (_mm512_cmpneq_epi64_mask): Ditto.
7021 (_mm512_cmpneq_epu64_mask): Ditto.
7022 (_mm512_expand_pd): Ditto.
7023 (_mm512_expand_ps): Ditto.
7024 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
7025 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
7026 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
7027 * config/i386/i386.c (ix86_builtins): Add
7028 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
7029 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
7030 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
7031 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
7032 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
7033 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
7034 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
7035 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
7036 IX86_BUILTIN_PMOVUSQW512_MEM.
7037 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
7038 __builtin_ia32_pmovsqd512mem_mask,
7039 __builtin_ia32_pmovqd512mem_mask,
7040 __builtin_ia32_pmovusqw512mem_mask,
7041 __builtin_ia32_pmovsqw512mem_mask,
7042 __builtin_ia32_pmovqw512mem_mask,
7043 __builtin_ia32_pmovusdw512mem_mask,
7044 __builtin_ia32_pmovsdw512mem_mask,
7045 __builtin_ia32_pmovdw512mem_mask,
7046 __builtin_ia32_pmovqb512mem_mask,
7047 __builtin_ia32_pmovusqb512mem_mask,
7048 __builtin_ia32_pmovsqb512mem_mask,
7049 __builtin_ia32_pmovusdb512mem_mask,
7050 __builtin_ia32_pmovsdb512mem_mask,
7051 __builtin_ia32_pmovdb512mem_mask.
7052 (bdesc_args): Add __builtin_ia32_expanddf512,
7053 __builtin_ia32_expandsf512.
7054 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
7055 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
7056 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
7057 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
7058 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
7059 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
7060 (avx512f_<code>v8div16qi2_mask_store): This.
7061 (avx512f_expand<mode>): New.
7063 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
7065 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
7067 (_mm512_mask_prefetch_i64gather_pd): Ditto.
7068 (_mm512_prefetch_i32scatter_pd): Ditto.
7069 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
7070 (_mm512_prefetch_i64scatter_pd): Ditto.
7071 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
7072 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
7073 (_mm512_mask_prefetch_i64gather_ps): Ditto.
7074 (_mm512_prefetch_i32scatter_ps): Ditto.
7075 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
7076 (_mm512_prefetch_i64scatter_ps): Ditto.
7077 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
7078 * config/i386/i386-builtin-types.def: Define
7079 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
7080 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
7081 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
7082 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
7083 IX86_BUILTIN_SCATTERPFQPD.
7084 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
7085 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
7086 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
7087 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
7088 __builtin_ia32_scatterpfqps.
7089 (ix86_expand_builtin): Expand new built-ins.
7090 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
7091 fix memory access data type.
7092 (*avx512pf_gatherpf<mode>_mask): Ditto.
7093 (*avx512pf_gatherpf<mode>): Ditto.
7094 (avx512pf_scatterpf<mode>): Ditto.
7095 (*avx512pf_scatterpf<mode>_mask): Ditto.
7096 (*avx512pf_scatterpf<mode>): Ditto.
7097 (GATHER_SCATTER_SF_MEM_MODE): New.
7098 (avx512pf_gatherpf<mode>df): Ditto.
7099 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7100 (*avx512pf_scatterpf<mode>df): Ditto.
7102 2014-01-27 Jakub Jelinek <jakub@redhat.com>
7105 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
7106 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
7109 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
7111 * common/config/arm/arm-common.c
7112 (arm_rewrite_mcpu): Handle multiple names.
7114 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
7116 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
7118 * gimple-builder.h (create_gimple_tmp): Delete.
7120 2014-01-27 Christian Bruel <christian.bruel@st.com>
7122 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
7125 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
7127 * config/pa/pa.md (call): Generate indirect long calls to non-local
7128 functions when outputing 32-bit code.
7129 (call_value): Likewise except for special call to buggy powf function.
7131 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
7132 portable runtime and PIC indirect calls.
7133 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
7134 and PIC call sequences. Use ldo instead of blr to set return register
7135 in PIC call sequence.
7137 2014-01-25 Walter Lee <walt@tilera.com>
7139 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
7140 avoid clobbering a live register.
7142 2014-01-25 Walter Lee <walt@tilera.com>
7144 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
7145 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
7146 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
7147 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
7149 2014-01-25 Walter Lee <walt@tilera.com>
7151 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
7152 arguments on even registers.
7153 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
7155 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
7156 (BIGGEST_ALIGNMENT): Ditto.
7157 (BIGGEST_FIELD_ALIGNMENT): Ditto.
7159 2014-01-25 Walter Lee <walt@tilera.com>
7161 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
7162 insns before bundling.
7163 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
7165 2014-01-25 Walter Lee <walt@tilera.com>
7167 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
7168 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
7169 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
7171 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
7173 * config/mips/constraints.md (kl): Delete.
7174 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
7175 define expands, using...
7176 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
7177 instructions for MIPS16.
7178 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
7179 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
7181 2014-01-25 Walter Lee <walt@tilera.com>
7183 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
7187 2014-01-25 Walter Lee <walt@tilera.com>
7189 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
7190 (TARGET_EXPAND_TO_RTL_HOOK): Define.
7192 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
7194 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
7197 2014-01-25 Jakub Jelinek <jakub@redhat.com>
7199 * print-rtl.c (in_call_function_usage): New var.
7200 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
7201 EXPR_LIST mode as mode and not as reg note name.
7204 * cfgloopmanip.c (copy_loop_info): If
7205 loop->warned_aggressive_loop_optimizations, make sure
7206 the flag is set in target loop too.
7208 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
7210 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
7212 * builtins.def: Likewise.
7213 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
7214 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
7215 * ira.c (ira_setup_eliminable_regset): Likewise.
7216 * omp-low.c (gate_expand_omp): Likewise.
7217 (execute_lower_omp): Likewise.
7218 (diagnose_sb_0): Likewise.
7219 (gate_diagnose_omp_blocks): Likewise.
7220 (simd_clone_clauses_extract): Likewise.
7223 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7225 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
7226 correction for little endian...
7227 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
7230 2014-01-24 Jeff Law <law@redhat.com>
7232 PR tree-optimization/59919
7233 * tree-vrp.c (find_assert_locations_1): Do not register asserts
7234 for non-returning calls.
7236 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
7238 * common/config/aarch64/aarch64-common.c
7239 (aarch64_rewrite_mcpu): Handle multiple names.
7240 * config/aarch64/aarch64.h
7241 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
7243 2014-01-24 Dodji Seketeli <dodji@redhat.com>
7245 * input.c (add_file_to_cache_tab): Handle the case where fopen
7248 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
7251 * config/i386/i386.md (pushsf splitter): Get stack adjustment
7252 from push operand if code of push isn't PRE_DEC.
7254 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7257 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
7258 -mquad-memory-atomic. Update -mquad-memory documentation to say
7259 it is only used for non-atomic loads/stores.
7261 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
7262 -mquad-memory or -mquad-memory-atomic switches.
7264 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
7265 -mquad-memory-atomic to ISA 2.07 support.
7267 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
7268 to separate support of normal quad word memory operations (ldq, stq)
7269 from the atomic quad word memory operations.
7271 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
7272 support to separate non-atomic quad word operations from atomic
7273 quad word operations. Disable non-atomic quad word operations in
7274 little endian mode so that we don't have to swap words after the
7275 load and before the store.
7276 (quad_load_store_p): Add comment about atomic quad word support.
7277 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
7278 options printed with -mdebug=reg.
7280 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
7281 -mquad-memory-atomic as the test for whether we have quad word
7282 atomic instructions.
7283 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
7284 or -mp8-vector are used, allow byte/half-word atomic operations.
7286 * config/rs6000/sync.md (load_lockedti): Insure that the address
7287 is a proper indexed or indirect address for the lqarx instruction.
7288 On little endian systems, swap the hi/lo registers after the lqarx
7290 (load_lockedpti): Use indexed_or_indirect_operand predicate to
7291 insure the address is valid for the lqarx instruction.
7292 (store_conditionalti): Insure that the address is a proper indexed
7293 or indirect address for the stqcrx. instruction. On little endian
7294 systems, swap the hi/lo registers before doing the stqcrx.
7296 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
7297 insure the address is valid for the stqcrx. instruction.
7299 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
7300 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
7301 type of quad memory support is available.
7303 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
7306 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
7307 there is a danger of looping.
7309 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
7311 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
7312 force flag_ira_loop_pressure if set via command line.
7314 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7316 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
7317 (ashr_simd): New builtin handling DI mode.
7318 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
7319 (aarch64_sshr_simddi): New match pattern.
7320 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
7321 (vshrd_n_s64): Likewise.
7322 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
7324 2014-01-23 Nick Clifton <nickc@redhat.com>
7326 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
7327 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
7328 favour of mcu specific scripts.
7329 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
7332 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7333 Alex Velenko <Alex.Velenko@arm.com>
7335 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
7336 (vaddv_s16): Likewise.
7337 (vaddv_s32): Likewise.
7338 (vaddv_u8): Likewise.
7339 (vaddv_u16): Likewise.
7340 (vaddv_u32): Likewise.
7341 (vaddvq_s8): Likewise.
7342 (vaddvq_s16): Likewise.
7343 (vaddvq_s32): Likewise.
7344 (vaddvq_s64): Likewise.
7345 (vaddvq_u8): Likewise.
7346 (vaddvq_u16): Likewise.
7347 (vaddvq_u32): Likewise.
7348 (vaddvq_u64): Likewise.
7349 (vaddv_f32): Likewise.
7350 (vaddvq_f32): Likewise.
7351 (vaddvq_f64): Likewise.
7352 (vmaxv_f32): Likewise.
7353 (vmaxv_s8): Likewise.
7354 (vmaxv_s16): Likewise.
7355 (vmaxv_s32): Likewise.
7356 (vmaxv_u8): Likewise.
7357 (vmaxv_u16): Likewise.
7358 (vmaxv_u32): Likewise.
7359 (vmaxvq_f32): Likewise.
7360 (vmaxvq_f64): Likewise.
7361 (vmaxvq_s8): Likewise.
7362 (vmaxvq_s16): Likewise.
7363 (vmaxvq_s32): Likewise.
7364 (vmaxvq_u8): Likewise.
7365 (vmaxvq_u16): Likewise.
7366 (vmaxvq_u32): Likewise.
7367 (vmaxnmv_f32): Likewise.
7368 (vmaxnmvq_f32): Likewise.
7369 (vmaxnmvq_f64): Likewise.
7370 (vminv_f32): Likewise.
7371 (vminv_s8): Likewise.
7372 (vminv_s16): Likewise.
7373 (vminv_s32): Likewise.
7374 (vminv_u8): Likewise.
7375 (vminv_u16): Likewise.
7376 (vminv_u32): Likewise.
7377 (vminvq_f32): Likewise.
7378 (vminvq_f64): Likewise.
7379 (vminvq_s8): Likewise.
7380 (vminvq_s16): Likewise.
7381 (vminvq_s32): Likewise.
7382 (vminvq_u8): Likewise.
7383 (vminvq_u16): Likewise.
7384 (vminvq_u32): Likewise.
7385 (vminnmv_f32): Likewise.
7386 (vminnmvq_f32): Likewise.
7387 (vminnmvq_f64): Likewise.
7389 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7391 * config/aarch64/aarch64-simd.md
7392 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
7393 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
7394 (*aarch64_mul3_elt<mode>): Likewise.
7395 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
7396 (*aarch64_mul3_elt_to_64v2df): Likewise.
7397 (*aarch64_mla_elt<mode>): Likewise.
7398 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7399 (*aarch64_mls_elt<mode>): Likewise.
7400 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7401 (*aarch64_fma4_elt<mode>): Likewise.
7402 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7403 (*aarch64_fma4_elt_to_64v2df): Likewise.
7404 (*aarch64_fnma4_elt<mode>): Likewise.
7405 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7406 (*aarch64_fnma4_elt_to_64v2df): Likewise.
7407 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
7408 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
7409 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7410 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7411 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
7412 (aarch64_sqdmull_lane<mode>_internal): Likewise.
7413 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
7415 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
7417 * config/aarch64/aarch64-simd.md
7418 (aarch64_be_checked_get_lane<mode>): New define_expand.
7419 * config/aarch64/aarch64-simd-builtins.def
7420 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
7421 New builtin definition.
7422 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
7423 Use new safe be builtin.
7425 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7427 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
7429 (aarch64_be_st1<mode>): Likewise.
7430 (aarch_ld1<VALL:mode>): Define_expand modified.
7431 (aarch_st1<VALL:mode>): Likewise.
7432 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
7433 (UNSPEC_ST1): Likewise.
7435 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
7437 * config/microblaze/microblaze.md: Add trap insn and attribute
7439 2014-01-23 Dodji Seketeli <dodji@redhat.com>
7441 PR preprocessor/58580
7442 * input.h (location_get_source_line): Take an additional line_size
7444 (void diagnostics_file_cache_fini): Declare new function.
7445 * input.c (struct fcache): New type.
7446 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
7447 New static constants.
7448 (diagnostic_file_cache_init, total_lines_num)
7449 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
7450 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
7451 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
7452 (get_next_line, read_next_line, goto_next_line, read_line_num):
7453 New static function definitions.
7454 (diagnostic_file_cache_fini): New function.
7455 (location_get_source_line): Take an additional output line_len
7456 parameter. Re-write using lookup_or_add_file_to_cache_tab and
7458 * diagnostic.c (diagnostic_finish): Call
7459 diagnostic_file_cache_fini.
7460 (adjust_line): Take an additional input parameter for the length
7461 of the line, rather than calculating it with strlen.
7462 (diagnostic_show_locus): Adjust the use of
7463 location_get_source_line and adjust_line with respect to their new
7464 signature. While displaying a line now, do not stop at the first
7465 null byte. Rather, display the zero byte as a space and keep
7466 going until we reach the size of the line.
7467 * Makefile.in: Add vec.o to OBJS-libcommon
7469 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7470 Ilya Tocar <ilya.tocar@intel.com>
7472 * config/i386/avx512fintrin.h (_mm512_kmov): New.
7473 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
7474 (__builtin_ia32_kmov16): Ditto.
7475 * config/i386/i386.md (UNSPEC_KMOV): New.
7478 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7480 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
7481 (_mm512_storeu_si512): Ditto.
7483 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
7486 * rtl.h (get_referenced_operands): Declare.
7487 * recog.c (get_referenced_operands): New function.
7488 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
7489 operands have been referenced when recording LO_SUM references.
7491 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
7493 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
7495 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7497 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
7498 Enable for generic and recent AMD targets.
7500 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7502 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
7503 ARG_SIZE note when adjustment was eliminated.
7505 2014-01-22 Jeff Law <law@redhat.com>
7507 PR tree-optimization/59597
7508 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
7509 in file. Accept new argument REGISTERING and use it to modify
7510 dump output appropriately.
7511 (register_jump_thread): Corresponding changes.
7512 (mark_threaded_blocks): Reinstate code to cancel unprofitable
7513 thread paths involving joiner blocks. Add code to dump cancelled
7514 jump threading paths.
7516 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
7518 PR rtl-optimization/59477
7519 * lra-constraints.c (inherit_in_ebb): Process call for living hard
7520 regs. Update reloads_num and potential_reload_hard_regs for all insns.
7522 2014-01-22 Tom Tromey <tromey@redhat.com>
7524 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
7526 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
7528 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7530 PR rtl-optimization/59896
7531 * lra-constraints.c (process_alt_operands): Check unused note for
7532 matched operands of insn with no output reloads.
7534 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
7536 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
7537 (mips_move_from_gpr_cost): Likewise.
7539 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7541 PR rtl-optimization/59858
7542 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
7543 ira_class_hard_regs_num.
7544 (process_alt_operands): Increase reject for dying matched operand.
7546 2014-01-21 Jakub Jelinek <jakub@redhat.com>
7549 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
7550 smaller than size, perform several stores or loads and stores
7551 at dst + count - size to store or copy all of size bytes, rather
7552 than just last modesize bytes.
7554 2014-01-20 DJ Delorie <dj@redhat.com>
7556 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
7557 that CLOBBERs are REGs before propogating their values.
7559 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
7562 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
7563 (cgraph_inline_failed_type): New function.
7564 * cgraph.h (DEFCIFCODE): Add type.
7565 (cgraph_inline_failed_type_t): New enum.
7566 (cgraph_inline_failed_type): New prototype.
7567 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
7568 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
7569 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
7570 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
7571 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
7572 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
7573 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
7574 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
7575 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
7576 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
7577 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
7578 OPTIMIZATION_MISMATCH.
7579 * tree-inline.c (expand_call_inline): Emit errors during
7580 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
7582 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7585 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
7586 mode attribute in insn output.
7588 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
7590 * output.h (output_constant): Delete.
7591 * varasm.c (output_constant): Make private.
7593 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
7595 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
7597 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7600 * tree.h (fold_builtin_strcat): New prototype.
7601 * builtins.c (fold_builtin_strcat): No longer static. Add len
7602 argument, if non-NULL, don't call c_strlen. Optimize
7603 directly into __builtin_memcpy instead of __builtin_strcpy.
7604 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7605 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7607 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7609 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7610 for SImode_address_operand operands, having only a REG argument.
7612 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7614 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7615 loader name using mbig-endian.
7616 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7618 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7620 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7624 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7626 * config/aarch64/aarch64-protos.h
7627 (aarch64_cannot_change_mode_class_ptr): Declare.
7628 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7629 aarch64_cannot_change_mode_class_ptr): New.
7630 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7631 backend hook aarch64_cannot_change_mode_class.
7633 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7635 * common/config/aarch64/aarch64-common.c
7636 (aarch64_handle_option): Don't handle any option order logic here.
7637 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7638 selected_cpu, warn on architecture version mismatch.
7639 (aarch64_override_options): Fix parsing order for option strings.
7641 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7642 Iain Sandoe <iain@codesourcery.com>
7645 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7646 warning. Amend comment to reflect current functionality.
7648 2014-01-20 Richard Biener <rguenther@suse.de>
7651 * builtins.c (fold_builtin_strcat): Remove case better handled
7652 by tree-ssa-strlen.c.
7654 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7656 * config/aarch64/aarch64.opt
7657 (mcpu, march, mtune): Make case-insensitive.
7659 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7662 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7663 if operands[1] is a REG or ZERO_EXTEND of a REG.
7665 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
7667 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7669 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
7671 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7672 long non-pic millicode calls.
7674 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7676 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7678 2014-01-19 Kito Cheng <kito@0xlab.org>
7680 * builtins.c (expand_movstr): Check movstr expand done or fail.
7682 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7683 H.J. Lu <hongjiu.lu@intel.com>
7686 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7687 to DImode for zero-extended addresses.
7689 2014-01-19 Jakub Jelinek <jakub@redhat.com>
7691 PR rtl-optimization/57763
7692 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7693 on the new indirect jump_insn and increment LABEL_NUSES (label).
7695 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
7699 * config.gcc (x86_archs): New variable.
7700 (x86_64_archs): Likewise.
7701 (x86_cpus): Likewise.
7702 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7703 --with-arch/--with-cpu= options.
7704 Support --with-arch=/--with-cpu={nehalem,westmere,
7705 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7707 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7709 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7710 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7712 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7714 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7716 2014-01-18 Jakub Jelinek <jakub@redhat.com>
7719 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7720 clear cpp_get_options (parse_in)->warn_unused_macros for
7721 ix86_target_macros_internal with cpp_define.
7723 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7725 * jump.c (delete_related_insns): Keep (use (insn))s.
7726 * reorg.c (redundant_insn): Check for barriers too.
7728 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7730 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
7732 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
7734 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7735 call to $$dyncall when TARGET_LONG_CALLS is true.
7737 2014-01-17 Jeff Law <law@redhat.com>
7739 * ree.c (combine_set_extension): Temporarily disable test for
7740 changing number of hard registers.
7742 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7745 * ipa-inline-analysis.c (inline_free_summary):
7746 Do not free summary of aliases.
7748 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7751 * gimplify.c (gimplify_expr): Use create_tmp_var
7752 instead of create_tmp_var_raw. If cond doesn't have
7753 integral type, don't add the IFN_ANNOTATE builtin at all.
7755 2014-01-17 Martin Jambor <mjambor@suse.cz>
7758 * ipa-cp.c (prev_edge_clone): New variable.
7759 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7760 Also resize prev_edge_clone vector.
7761 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7762 (ipcp_edge_removal_hook): New function.
7763 (ipcp_driver): Register ipcp_edge_removal_hook.
7765 2014-01-17 Andrew Pinski <apinski@cavium.com>
7766 Steve Ellcey <sellcey@mips.com>
7769 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7772 2014-01-17 Jeff Law <law@redhat.com>
7775 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7776 so that pass_ccp runs first.
7778 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7780 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7781 (ix86_adjust_cost): Use !TARGET_XXX.
7782 (do_reorder_for_imul): Likewise.
7783 (swap_top_of_ready_list): Likewise.
7784 (ix86_sched_reorder): Likewise.
7786 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7788 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7789 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7790 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7791 (intel_memset): New. Duplicate slm_memset.
7792 (intel_cost): New. Duplicate slm_cost.
7793 (m_INTEL): New macro.
7794 (processor_target_table): Add "intel".
7795 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7796 with PROCESSOR_INTEL for "intel".
7797 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7798 PROCESSOR_SILVERMONT.
7799 (ix86_issue_rate): Likewise.
7800 (ix86_adjust_cost): Likewise.
7801 (ia32_multipass_dfa_lookahead): Likewise.
7802 (swap_top_of_ready_list): Likewise.
7803 (ix86_sched_reorder): Likewise.
7804 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7805 instead of TARGET_OPT_AGU.
7806 * config/i386/i386.h (TARGET_INTEL): New.
7807 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7808 (processor_type): Add PROCESSOR_INTEL.
7809 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7810 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7812 2014-01-17 Marek Polacek <polacek@redhat.com>
7815 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7818 2014-01-17 Richard Biener <rguenther@suse.de>
7820 PR tree-optimization/46590
7821 * opts.c (default_options_table): Add entries for
7822 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7823 all enabled at -O1 but not for -Og.
7824 * common.opt (fbranch-count-reg): Remove Init(1).
7825 (fmove-loop-invariants): Likewise.
7826 (ftree-pta): Likewise.
7828 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7830 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7831 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7832 decls to at least the GCC 4.8 used alignments.
7835 * tree-nested.c (convert_nonlocal_reference_stmt,
7836 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7837 of GIMPLE_BIND stmts, adjust associated decls.
7839 2014-01-17 Richard Biener <rguenther@suse.de>
7841 PR tree-optimization/46590
7842 * vec.h (vec<>::bseach): New member function implementing
7843 binary search according to C89 bsearch.
7844 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7845 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7846 bitmap pointer again. Make accesses_in_loop a flat array.
7847 (mem_ref_obstack): New global.
7848 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7849 (mark_ref_stored): Likewise.
7850 (ref_indep_loop_p_2): Likewise.
7851 (set_ref_stored_in_loop): New helper function.
7852 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7853 (memref_free): Adjust.
7854 (record_mem_ref_loc): Simplify.
7855 (gather_mem_refs_stmt): Adjust.
7856 (sort_locs_in_loop_postorder_cmp): New function.
7857 (analyze_memory_references): Sort accesses_in_loop after
7858 loop postorder number.
7859 (find_ref_loc_in_loop_cmp): New function.
7860 (for_all_locs_in_loop): Find relevant cluster of locs in
7861 accesses_in_loop and iterate without recursion.
7862 (execute_sm): Avoid uninit warning.
7863 (struct ref_always_accessed): Simplify.
7864 (ref_always_accessed::operator ()): Likewise.
7865 (ref_always_accessed_p): Likewise.
7866 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7867 loop postorder numbers here.
7868 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7871 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7874 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7875 on decls for which assemble_alias has been called.
7877 2014-01-17 Nick Clifton <nickc@redhat.com>
7879 * config/msp430/msp430.opt: (mcpu): New option.
7880 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7881 (msp430_option_override): Parse target_cpu. If the MCU name
7882 matches a generic string, clear target_mcu.
7883 (msp430_attr): Allow numeric interrupt values up to 63.
7884 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7885 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7887 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7889 * config/msp430/msp430.md (popm): Use %J rather than %I.
7890 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7891 (addhi_cy_i): Use immediate_operand for operand 2.
7892 * doc/invoke.texi: Document -mcpu option.
7894 2014-01-17 Richard Biener <rguenther@suse.de>
7896 PR rtl-optimization/38518
7897 * df.h (df_analyze_loop): Declare.
7898 * df-core.c: Include cfgloop.h.
7899 (df_analyze_1): Split out main part of df_analyze.
7900 (df_analyze): Adjust.
7901 (loop_inverted_post_order_compute): New function.
7902 (loop_post_order_compute): Likewise.
7903 (df_analyze_loop): New function avoiding whole-function
7905 * loop-invariant.c (find_defs): Use df_analyze_loop.
7906 (find_invariants): Adjust.
7907 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7909 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7911 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7912 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7914 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7916 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7917 traversal when removing references.
7919 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
7922 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7924 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7927 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7928 pushing a reload for an autoinc when we had previously reloaded an
7929 inner part of the address.
7931 2014-01-16 Jakub Jelinek <jakub@redhat.com>
7933 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7935 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7936 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7937 when not giving up or versioning for alias only because of
7939 (vect_analyze_data_ref_dependences): Set to true.
7940 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7942 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7943 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7947 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7950 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7951 operand 0 predicate for gathers, use a new pseudo as subtarget.
7953 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7956 * lra-constraints.c (process_alt_operands): Add printing debug info.
7957 Check absence of input/output reloads for matched operands too.
7959 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7961 PR rtl-optimization/59835
7962 * ira.c (ira_init_register_move_cost): Increase cost for
7965 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7967 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
7969 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
7972 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7973 non-register objects. Use gen_(high/low)part more consistently.
7976 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7979 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
7980 endian support, remove tests for WORDS_BIG_ENDIAN.
7981 (p8_mfvsrd_3_<mode>): Likewise.
7982 (reload_gpr_from_vsx<mode>): Likewise.
7983 (reload_gpr_from_vsxsf): Likewise.
7984 (p8_mfvsrd_4_disf): Likewise.
7986 2014-01-16 Richard Biener <rguenther@suse.de>
7988 PR rtl-optimization/46590
7989 * lcm.c (compute_antinout_edge): Use postorder iteration.
7990 (compute_laterin): Use inverted postorder iteration.
7992 2014-01-16 Nick Clifton <nickc@redhat.com>
7995 * varasm.c (output_constant): Return the number of bytes actually
7997 (output_constructor_array_range): Update the field size with the
7998 number of bytes emitted by output_constant.
7999 (output_constructor_regular_field): Likewise. Also do not
8000 complain if the total number of bytes emitted is now greater
8001 than the expected fieldpos.
8002 * output.h (output_constant): Update prototype and descriptive comment.
8004 2014-01-16 Marek Polacek <polacek@redhat.com>
8007 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
8008 it is error_mark_node.
8010 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
8012 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
8013 VALID_AVX256_REG_OR_OI_MODE.
8015 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
8017 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
8018 current procedure should be profiled.
8020 2014-01-15 Andrew Pinski <apinski@cavium.com>
8022 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
8023 of moving from/to the STACK_REG register class.
8025 2014-01-15 Richard Henderson <rth@redhat.com>
8028 * reginfo.c (global_regs_decl): Globalize.
8029 * rtl.h (global_regs_decl): Declare.
8030 * ira.c (do_reload): Diagnose frame_pointer_needed and it
8031 reserved via global_regs.
8033 2014-01-15 Teresa Johnson <tejohnson@google.com>
8035 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
8037 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8039 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
8040 and vmulosh rather than call gen_vec_widen_smult_*.
8041 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
8042 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
8043 (vec_widen_smult_even_v16qi): Likewise.
8044 (vec_widen_umult_even_v8hi): Likewise.
8045 (vec_widen_smult_even_v8hi): Likewise.
8046 (vec_widen_umult_odd_v16qi): Likewise.
8047 (vec_widen_smult_odd_v16qi): Likewise.
8048 (vec_widen_umult_odd_v8hi): Likewise.
8049 (vec_widen_smult_odd_v8hi): Likewise.
8050 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
8051 vmuloub rather than call gen_vec_widen_umult_*.
8052 (vec_widen_umult_lo_v16qi): Likewise.
8053 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
8054 vmulosb rather than call gen_vec_widen_smult_*.
8055 (vec_widen_smult_lo_v16qi): Likewise.
8056 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
8057 rather than call gen_vec_widen_umult_*.
8058 (vec_widen_umult_lo_v8hi): Likewise.
8059 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
8060 rather than call gen_vec_widen_smult_*.
8061 (vec_widen_smult_lo_v8hi): Likewise.
8063 2014-01-15 Jeff Law <law@redhat.com>
8065 PR tree-optimization/59747
8066 * ree.c (find_and_remove_re): Properly handle case where a second
8067 eliminated extension requires widening a copy created for elimination
8068 of a prior extension.
8069 (combine_set_extension): Ensure that the number of hard regs needed
8070 for a destination register does not change when we widen it.
8072 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
8074 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
8075 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
8076 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
8077 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
8078 (avr-*-rtems*): Likewise.
8079 (bfin*-rtems*): Likewise.
8080 (moxie-*-rtems*): Likewise.
8081 (h8300-*-rtems*): Likewise.
8082 (i[34567]86-*-rtems*): Likewise.
8083 (lm32-*-rtems*): Likewise.
8084 (m32r-*-rtems*): Likewise.
8085 (m68k-*-rtems*): Likewise.
8086 (microblaze*-*-rtems*): Likewise.
8087 (mips*-*-rtems*): Likewise.
8088 (powerpc-*-rtems*): Likewise.
8089 (sh-*-rtems*): Likewise.
8090 (sparc-*-rtems*): Likewise.
8091 (sparc64-*-rtems*): Likewise.
8092 (v850-*-rtems*): Likewise.
8093 (m32c-*-rtems*): Likewise.
8095 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
8097 PR rtl-optimization/59511
8098 * ira.c (ira_init_register_move_cost): Use memory costs for some
8099 cases of register move cost calculations.
8100 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
8101 instead of BB frequency.
8102 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
8103 * lra-assigns.c (find_hard_regno_for): Ditto.
8105 2014-01-15 Richard Biener <rguenther@suse.de>
8107 PR tree-optimization/59822
8108 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
8109 (vectorizable_load): Use it to hoist defs of uses of invariant
8110 loads out of the loop.
8112 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8113 Kugan Vivekanandarajah <kuganv@linaro.org>
8116 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
8119 2014-01-15 Richard Biener <rguenther@suse.de>
8121 PR rtl-optimization/59802
8122 * lcm.c (compute_available): Use inverted postorder to seed
8123 the initial worklist.
8125 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8128 * config/s390/s390.c (s390_preferred_reload_class): Don't return
8129 ADDR_REGS for invalid symrefs in non-PIC code.
8131 2014-01-15 Jakub Jelinek <jakub@redhat.com>
8134 * builtins.c (determine_block_size): Initialize *probable_max_size
8135 even if len_rtx is CONST_INT.
8137 2014-01-14 Andrew Pinski <apinski@cavium.com>
8139 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
8140 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
8141 (cortexa53_tunings): Likewise.
8142 (aarch64_sched_issue_rate): New function.
8143 (TARGET_SCHED_ISSUE_RATE): Define.
8145 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
8147 * ira-costs.c (find_costs_and_classes): Add missed
8148 ira_init_register_move_cost_if_necessary.
8150 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
8153 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
8155 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
8158 * config/i386/i386.c (type_natural_mode): Add a bool parameter
8159 to indicate if type is used for function return value. Warn ABI
8160 change if the vector mode isn't available for function return value.
8161 (ix86_function_arg_advance): Pass false to type_natural_mode.
8162 (ix86_function_arg): Likewise.
8163 (ix86_gimplify_va_arg): Likewise.
8164 (function_arg_32): Don't warn ABI change.
8165 (ix86_function_value): Pass true to type_natural_mode.
8166 (ix86_return_in_memory): Likewise.
8167 (ix86_struct_value_rtx): Removed.
8168 (TARGET_STRUCT_VALUE_RTX): Likewise.
8170 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8172 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
8173 converting a conditional jump into a conditional return.
8175 2014-01-14 Richard Biener <rguenther@suse.de>
8177 PR tree-optimization/58921
8178 PR tree-optimization/59006
8179 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
8180 hoisting invariant stmts.
8181 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
8182 invariant loads on the preheader edge if possible.
8184 2014-01-14 Joey Ye <joey.ye@arm.com>
8186 * doc/plugin.texi (Building GCC plugins): Update to C++.
8188 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
8190 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
8191 (_mm_rcp28_round_ss): Ditto.
8192 (_mm_rsqrt28_round_sd): Ditto.
8193 (_mm_rsqrt28_round_ss): Ditto.
8194 (_mm_rcp28_sd): Ditto.
8195 (_mm_rcp28_ss): Ditto.
8196 (_mm_rsqrt28_sd): Ditto.
8197 (_mm_rsqrt28_ss): Ditto.
8198 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
8199 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
8200 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
8201 (IX86_BUILTIN_RCP28SD): Ditto.
8202 (IX86_BUILTIN_RCP28SS): Ditto.
8203 (IX86_BUILTIN_RSQRT28SD): Ditto.
8204 (IX86_BUILTIN_RSQRT28SS): Ditto.
8205 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
8206 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
8207 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
8208 (ix86_expand_special_args_builtin): Expand new FTYPE.
8209 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
8210 (srcp14<mode>): Make insn unary.
8211 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
8212 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
8213 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
8214 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
8215 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
8216 Fix rounding: make it SAE only.
8217 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
8219 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
8221 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
8222 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
8223 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
8224 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
8225 (round_saeonly_mask_scalar_operand4): Ditto.
8226 (round_saeonly_mask_scalar_op3): Ditto.
8227 (round_saeonly_mask_scalar_op4): Ditto.
8229 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8231 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
8232 Implement -maltivec=be for vec_insert and vec_extract.
8234 2014-01-10 DJ Delorie <dj@redhat.com>
8236 * config/msp430/msp430.md (call_internal): Don't allow memory
8237 references with SP as the base register.
8238 (call_value_internal): Likewise.
8239 * config/msp430/constraints.md (Yc): New. For memory references
8240 that don't use SP as a base register.
8242 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
8243 "an integer without a # prefix"
8244 * config/msp430/msp430.md (epilogue_helper): Use it.
8246 2014-01-13 Jakub Jelinek <jakub@redhat.com>
8249 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
8250 AVX512F gather builtins.
8251 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
8252 on gather decls with INTEGER_TYPE masktype.
8253 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
8254 directly into the builtin rather than hoisting it before loop.
8256 PR tree-optimization/59387
8257 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
8258 (scev_const_prop): If folded_casts and type has undefined overflow,
8259 use force_gimple_operand instead of force_gimple_operand_gsi and
8260 for each added stmt if it is assign with
8261 arith_code_with_undefined_signed_overflow, call
8262 rewrite_to_defined_overflow.
8263 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
8264 gimple-fold.h instead.
8265 (arith_code_with_undefined_signed_overflow,
8266 rewrite_to_defined_overflow): Moved to ...
8267 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
8268 rewrite_to_defined_overflow): ... here. No longer static.
8269 Include gimplify-me.h.
8270 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
8271 rewrite_to_defined_overflow): New prototypes.
8273 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8275 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
8277 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
8279 * builtins.c (get_object_alignment_2): Minor tweak.
8280 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
8282 2014-01-13 Christian Bruel <christian.bruel@st.com>
8284 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
8285 optimized non constant lengths.
8287 2014-01-13 Jakub Jelinek <jakub@redhat.com>
8290 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
8291 load as __atomic_load_N if possible.
8293 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8295 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
8297 (rs6000_expand_builtin): Adjust call.
8299 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8302 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
8303 * config/rs6000/rs6000.c: Include target-globals.h.
8304 (rs6000_set_current_function): Instead of doing target_reinit
8305 unconditionally, use save_target_globals_default_opts and
8306 restore_target_globals.
8308 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
8310 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
8311 (rs6000_expand_builtin): Handle mffs and mtfsf.
8312 (rs6000_init_builtins): Define mffs and mtfsf.
8313 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
8314 (rs6000_mffs): New pattern.
8315 (rs6000_mtfsf): New pattern.
8317 2014-01-11 Bin Cheng <bin.cheng@arm.com>
8319 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
8320 Start narrowing with START. Apply candidate-use pair
8321 and check overall cost in narrowing.
8322 (iv_ca_prune): Pass new argument.
8324 2014-01-10 Jeff Law <law@redhat.com>
8327 * ree.c (combine_reaching_defs): Ensure the defining statement
8328 occurs before the extension when optimizing extensions with
8329 different source and destination hard registers.
8331 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8334 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
8335 vtables into the type inheritance graph.
8337 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8339 PR rtl-optimization/59754
8340 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
8341 modes in the REGNO != REGNO case.
8343 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8345 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
8347 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8349 PR tree-optimization/59745
8350 * tree-predcom.c (tree_predictive_commoning_loop): Call
8351 free_affine_expand_cache if giving up because components is NULL.
8353 * target-globals.c (save_target_globals): Allocate < 4KB structs using
8354 GC in payload of target_globals struct instead of allocating them on
8355 the heap and the larger structs separately using GC.
8356 * target-globals.h (struct target_globals): Make regs, hard_regs,
8357 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
8358 of GTY((skip)) and change type to void *.
8359 (reset_target_globals): Cast loads from those fields to corresponding
8362 2014-01-10 Steve Ellcey <sellcey@mips.com>
8365 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
8366 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
8367 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
8369 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
8372 * aarch64-modes.def (CC_Zmode): New flags mode.
8373 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
8374 represents an equality.
8375 (aarch64_get_condition_code): Handle CC_Zmode.
8376 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
8378 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8380 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
8381 extraction in good case.
8383 2014-01-10 Richard Biener <rguenther@suse.de>
8385 PR tree-optimization/59374
8386 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
8387 checking after SLP discovery. Mark stmts not participating
8388 in any SLP instance properly.
8390 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8392 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
8393 when handling a SET rtx.
8395 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8397 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
8398 (cortex-a57): Likewise.
8399 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
8401 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8403 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
8404 non-iwmmxt builtins.
8406 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8410 * ipa-devirt.c record_target_from_binfo): Take as argument
8411 stack of binfos and lookup matching one for virtual inheritance.
8412 (possible_polymorphic_call_targets_1): Update.
8414 2014-01-10 Huacai Chen <chenhc@lemote.com>
8416 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
8417 kernel strings for Loongson-2E/2F/3A.
8419 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8422 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
8423 is_gimple_call before calling gimple_call_internal_p.
8425 2014-01-09 Steve Ellcey <sellcey@mips.com>
8427 * Makefile.in (TREE_FLOW_H): Remove.
8428 (TREE_SSA_H): Add file names from tree-flow.h.
8429 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
8430 * tree.h: Remove tree-flow.h reference.
8431 * hash-table.h: Remove tree-flow.h reference.
8432 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
8433 reference with tree-ssa-loop.h.
8435 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8437 * doc/invoke.texi: Add -maltivec={be,le} options, and document
8438 default element-order behavior for -maltivec.
8439 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
8440 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
8441 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
8442 when targeting big endian, at least for now.
8443 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
8445 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8448 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
8449 var satisfies use_register_for_decl, just take into account type
8450 alignment, rather than decl alignment.
8452 PR tree-optimization/59622
8453 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
8454 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
8455 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
8456 Don't devirtualize for inplace at all. For targets.length () == 1,
8457 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
8459 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8461 * config/i386/i386.md (cpu): Remove the unused btver1.
8463 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8465 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
8467 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8470 * tree-core.h (struct target_globals): New forward declaration.
8471 (struct tree_target_option): Add globals field.
8472 * tree.h (TREE_TARGET_GLOBALS): Define.
8473 (prepare_target_option_nodes_for_pch): New prototype.
8474 * target-globals.h (struct target_globals): Define even if
8476 * tree.c (prepare_target_option_node_for_pch,
8477 prepare_target_option_nodes_for_pch): New functions.
8478 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
8479 * config/i386/i386.c: Include target-globals.h.
8480 (ix86_set_current_function): Instead of doing target_reinit
8481 unconditionally, use save_target_globals_default_opts and
8482 restore_target_globals.
8484 2014-01-09 Richard Biener <rguenther@suse.de>
8486 PR tree-optimization/59715
8487 * tree-cfg.h (split_critical_edges): Declare.
8488 * tree-cfg.c (split_critical_edges): Export.
8489 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
8491 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
8493 * cfgexpand.c (expand_stack_vars): Optionally disable
8494 asan stack protection.
8495 (expand_used_vars): Likewise.
8496 (partition_stack_vars): Likewise.
8497 * asan.c (asan_emit_stack_protection): Optionally disable
8498 after return stack usage.
8499 (instrument_derefs): Optionally disable memory access instrumentation.
8500 (instrument_builtin_call): Likewise.
8501 (instrument_strlen_call): Likewise.
8502 (asan_protect_global): Optionally disable global variables protection.
8503 * doc/invoke.texi: Added doc for new options.
8504 * params.def: Added new options.
8505 * params.h: Likewise.
8507 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8509 PR rtl-optimization/59724
8510 * ifcvt.c (cond_exec_process_if_block): Don't call
8511 flow_find_head_matching_sequence with 0 longest_match.
8512 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
8513 non-active insns if !stop_after.
8514 (try_head_merge_bb): Revert 2014-01-07 changes.
8516 2014-01-08 Jeff Law <law@redhat.com>
8518 * ree.c (get_sub_rtx): New function, extracted from...
8519 (merge_def_and_ext): Here.
8520 (combine_reaching_defs): Use get_sub_rtx.
8522 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
8524 * cgraph.h (varpool_variable_node): Do not choke on null node.
8526 2014-01-08 Catherine Moore <clm@codesourcery.com>
8528 * config/mips/mips.md (simple_return): Attempt to use JRC
8530 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
8532 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8534 PR rtl-optimization/59137
8535 * reorg.c (steal_delay_list_from_target): Call update_block for
8537 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
8539 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8541 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
8542 two duplicate entries.
8544 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8547 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
8549 * config/mips/mips.c (mips_truncated_op_cost): New function.
8550 (mips_rtx_costs): Adjust test for BADDU.
8551 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
8553 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
8555 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
8556 (*baddu_si): ...this new pattern.
8558 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8561 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
8563 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8566 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
8568 (expand_expr, expand_normal): Adjust.
8569 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
8570 inner_reference_p. Use inner_reference_p to expand inner references.
8571 (store_expr): Adjust.
8572 * cfgexpand.c (expand_call_stmt): Adjust.
8574 2014-01-08 Rong Xu <xur@google.com>
8576 * gcov-io.c (gcov_var): Move from gcov-io.h.
8577 (gcov_position): Ditto.
8578 (gcov_is_error): Ditto.
8579 (gcov_rewrite): Ditto.
8580 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
8581 only part to libgcc/libgcov.h.
8583 2014-01-08 Marek Polacek <polacek@redhat.com>
8586 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
8588 2014-01-08 Marek Polacek <polacek@redhat.com>
8591 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
8593 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8595 PR rtl-optimization/59649
8596 * stor-layout.c (get_mode_bounds): For BImode return
8597 0 and STORE_FLAG_VALUE.
8599 2014-01-08 Richard Biener <rguenther@suse.de>
8602 * gimple.h (is_gimple_builtin_call): Remove.
8603 (gimple_builtin_call_types_compatible_p): New.
8604 (gimple_call_builtin_p): New overload.
8605 * gimple.c (is_gimple_builtin_call): Remove.
8606 (validate_call): Rename to ...
8607 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8609 (validate_type): New static function.
8610 (gimple_call_builtin_p): New overload and adjust.
8611 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8612 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8613 (gimple_fold_stmt_to_constant_1): Likewise.
8614 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8616 2014-01-08 Richard Biener <rguenther@suse.de>
8619 * gimplify.c (gimplify_expr): Gimplify register-register type
8620 VIEW_CONVERT_EXPRs to separate stmts.
8622 2014-01-07 Jeff Law <law@redhat.com>
8625 * ree.c (combine_set_extension): Handle case where source
8626 and destination registers in an extension insn are different.
8627 (combine_reaching_defs): Allow source and destination registers
8628 in extension to be different under limited circumstances.
8629 (add_removable_extension): Remove restriction that the
8630 source and destination registers in the extension are the same.
8631 (find_and_remove_re): Emit a copy from the extension's
8632 destination to its source after the defining insn if
8633 the source and destination registers are different.
8636 * ifcvt.c (merge_if_block): If we are merging a block with more than
8637 one successor with a block with no successors, remove any BARRIER
8638 after the second block.
8640 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
8642 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8644 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
8647 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8648 for 14-bit register offsets when INT14_OK_STRICT is false.
8650 2014-01-07 Roland Stigge <stigge@antcom.de>
8651 Michael Meissner <meissner@linux.vnet.ibm.com>
8654 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
8655 Only check TFmode for SPE constants. Don't check TImode or TDmode.
8657 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8659 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8662 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8664 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8665 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8666 rtx is const0_rtx or not.
8668 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8671 * target-globals.c (save_target_globals): Remove this_fn_optab
8673 * toplev.c: Include optabs.h.
8674 (target_reinit): Temporarily restore the global options if another
8675 set of options are in force.
8677 2014-01-07 Jakub Jelinek <jakub@redhat.com>
8679 PR rtl-optimization/58668
8680 * cfgcleanup.c (flow_find_cross_jump): Don't count
8681 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8682 to determine what is counted.
8683 (flow_find_head_matching_sequence): Use active_insn_p to determine
8685 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8687 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8688 determine what is counted.
8690 PR tree-optimization/59643
8691 * tree-predcom.c (split_data_refs_to_components): If one dr is
8692 read and one write, determine_offset fails and the write isn't
8693 in the bad component, just put the read into the bad component.
8695 2014-01-07 Mike Stump <mikestump@comcast.net>
8696 Jakub Jelinek <jakub@redhat.com>
8699 * tree-core.h (struct tree_optimization_option): Change optabs
8700 type from unsigned char * to void *.
8701 * optabs.c (init_tree_optimization_optabs): Adjust
8702 TREE_OPTIMIZATION_OPTABS initialization.
8704 2014-01-06 Jakub Jelinek <jakub@redhat.com>
8707 * config/i386/i386.h (struct machine_function): Add
8708 no_drap_save_restore field.
8709 * config/i386/i386.c (ix86_save_reg): Use
8710 !cfun->machine->no_drap_save_restore instead of
8711 crtl->stack_realign_needed.
8712 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8713 this function clears frame_pointer_needed. Set
8714 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8715 and DRAP reg is needed.
8717 2014-01-06 Marek Polacek <polacek@redhat.com>
8720 * doc/implement-c.texi: Mention that other integer types are
8721 permitted as bit-field types in strictly conforming mode.
8723 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
8725 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8728 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
8730 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8732 2014-01-06 Martin Jambor <mjambor@suse.cz>
8735 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8737 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8739 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8743 * var-tracking.c (add_stores): Preserve the value of the source even if
8744 we don't record the store.
8746 2014-01-06 Terry Guo <terry.guo@arm.com>
8748 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8750 2014-01-05 Iain Sandoe <iain@codesourcery.com>
8753 * config/darwin.c (darwin_function_section): Adjust return values to
8754 correspond to optimisation changes made in r206070.
8756 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
8758 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8759 from prefetch_block tune setting.
8760 (nocona_cost): Correct size of prefetch block to 64.
8762 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8764 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8765 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8766 used to save the static chain register in the computation of the offset
8767 from which the FP registers need to be restored.
8769 2014-01-04 Jakub Jelinek <jakub@redhat.com>
8771 PR tree-optimization/59519
8772 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8773 ICE if get_current_def (current_new_name) is already non-NULL, as long
8774 as it is a phi result of some other phi in *new_exit_bb that has
8777 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8778 or vmovdqu* for misaligned_operand.
8779 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8780 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8781 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8782 aligned_mem for AVX512F masked aligned load and store builtins and for
8785 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
8787 PR tree-optimization/59651
8788 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
8789 Address range for negative step should be added by TYPE_SIZE_UNIT.
8791 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8793 * config/m68k/m68k.c (handle_move_double): Handle pushes with
8794 overlapping registers also for registers other than the stack pointer.
8796 2014-01-03 Marek Polacek <polacek@redhat.com>
8799 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8802 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8805 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8808 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8809 (push splitter): Use <P:MODE_SIZE> instead of
8810 GET_MODE_SIZE (<P:MODE>mode).
8811 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8812 (mov -1, reg peephole2): Likewise.
8813 * config/i386/sse.md (*mov<mode>_internal,
8814 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8815 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8816 *<code><mode>3, *andnot<mode>3<mask_name>,
8817 <mask_codefor><code><mode>3<mask_name>): Likewise.
8818 * config/i386/subst.md (mask_mode512bit_condition,
8819 sd_mask_mode512bit_condition): Likewise.
8821 2014-01-02 Xinliang David Li <davidxl@google.com>
8823 PR tree-optimization/59303
8824 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8825 (dump_predicates): Better output format.
8826 (pred_equal_p): New function.
8827 (is_neq_relop_p): Ditto.
8828 (is_neq_zero_form_p): Ditto.
8829 (pred_expr_equal_p): Ditto.
8830 (pred_neg_p): Ditto.
8831 (simplify_pred): Ditto.
8832 (simplify_preds_2): Ditto.
8833 (simplify_preds_3): Ditto.
8834 (simplify_preds_4): Ditto.
8835 (simplify_preds): Ditto.
8837 (push_to_worklist): Ditto.
8838 (get_pred_info_from_cmp): Ditto.
8839 (is_degenerated_phi): Ditto.
8840 (normalize_one_pred_1): Ditto.
8841 (normalize_one_pred): Ditto.
8842 (normalize_one_pred_chain): Ditto.
8843 (normalize_preds): Ditto.
8844 (normalize_cond_1): Remove function.
8845 (normalize_cond): Ditto.
8846 (is_gcond_subset_of): Ditto.
8847 (is_subset_of_any): Ditto.
8848 (is_or_set_subset_of): Ditto.
8849 (is_and_set_subset_of): Ditto.
8850 (is_norm_cond_subset_of): Ditto.
8851 (pred_chain_length_cmp): Ditto.
8852 (convert_control_dep_chain_into_preds): Type change.
8853 (find_predicates): Ditto.
8854 (find_def_preds): Ditto.
8855 (destroy_predicates_vecs): Ditto.
8856 (find_matching_predicates_in_rest_chains): Ditto.
8857 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8858 (is_pred_expr_subset): Ditto.
8859 (is_pred_chain_subset_of): Ditto.
8860 (is_included_in): Ditto.
8861 (is_superset_of): Ditto.
8863 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8865 Update copyright years.
8867 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8869 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8870 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8871 config/arc/arc.md, config/arc/arc.opt,
8872 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8873 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8874 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8875 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8876 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8877 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8878 vtable-verify.c, vtable-verify.h: Use the standard form for the
8881 2014-01-02 Tobias Burnus <burnus@net-b.de>
8883 * gcc.c (process_command): Update copyright notice dates.
8884 * gcov-dump.c: Ditto.
8886 * doc/cpp.texi: Bump @copying's copyright year.
8887 * doc/cppinternals.texi: Ditto.
8888 * doc/gcc.texi: Ditto.
8889 * doc/gccint.texi: Ditto.
8890 * doc/gcov.texi: Ditto.
8891 * doc/install.texi: Ditto.
8892 * doc/invoke.texi: Ditto.
8894 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8896 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8898 2014-01-01 Jakub Jelinek <jakub@redhat.com>
8900 * config/i386/sse.md (*mov<mode>_internal): Guard
8901 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8903 PR rtl-optimization/59647
8904 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8905 new_rtx into UNSIGNED_FLOAT rtxes.
8907 Copyright (C) 2014 Free Software Foundation, Inc.
8909 Copying and distribution of this file, with or without modification,
8910 are permitted in any medium without royalty provided the copyright
8911 notice and this notice are preserved.