1 2014-02-13 Richard Biener <rguenther@suse.de>
4 * doc/install.texi (ISL): Update recommended version to 0.12.2,
5 mention the possibility of an in-tree build.
6 (CLooG): Update recommended version to 0.18.1, mention the
7 possibility of an in-tree build and clarify that the ISL
8 bundled with CLooG does not work.
10 2014-02-13 Jakub Jelinek <jakub@redhat.com>
13 * expr.c (compress_float_constant): If x is a hard register,
14 extend into a pseudo and then move to x.
16 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
18 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
19 caused by bad second argument to warning_at() with -mhotpatch and
20 nested functions (e.g. with gfortran).
22 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
24 * opts.c (option_name): Remove "enabled by default" rider.
26 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
28 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
30 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
31 Uros Bizjak <ubizjak@gmail.com>
34 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU
36 * configure: Regenerated.
38 2014-02-12 Richard Biener <rguenther@suse.de>
40 * vec.c (vec_prefix::calculate_allocation): Move as
41 inline variant to vec.h.
42 (vec_prefix::calculate_allocation_1): New out-of-line version.
43 * vec.h (vec_prefix::calculate_allocation_1): Declare.
44 (vec_prefix::m_has_auto_buf): Rename to ...
45 (vec_prefix::m_using_auto_storage): ... this.
46 (vec_prefix::calculate_allocation): Inline the easy cases
47 and dispatch to calculate_allocation_1 which doesn't need the
49 (va_heap::reserve): Use gcc_checking_assert.
50 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
52 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
54 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
55 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
56 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
58 2014-02-12 Richard Biener <rguenther@suse.de>
60 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
61 when we found a dependence.
63 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
65 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
67 (maybe_fold_stmt): ... into this new function.
68 * omp-low.c (lower_omp): Update comment.
70 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
73 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
76 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
78 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
79 identifiers in comments.
80 (cortexa53_extra_costs): Likewise.
81 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in
83 (cortexa7_extra_costs): Likewise.
84 (cortexa12_extra_costs): Likewise.
85 (cortexa15_extra_costs): Likewise.
86 (v7m_extra_costs): Likewise.
88 2014-02-12 Richard Biener <rguenther@suse.de>
91 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
92 of posix_memalign being successful.
93 (lower_stmt): Restrict lowering of posix_memalign to when
94 -ftree-bit-ccp is enabled.
96 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
98 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
100 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
102 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
104 PR rtl-optimization/60116
105 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
106 other_insn once the combination has been validated.
108 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
111 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
113 * ipa-devirt.c: Include demangle.h
114 (odr_violation_reported): New static variable.
115 (add_type_duplicate): Update odr_violations.
116 (maybe_record_node): Add completep parameter; update it.
117 (record_target_from_binfo): Add COMPLETEP parameter;
119 (possible_polymorphic_call_targets_1): Likewise.
120 (struct polymorphic_call_target_d): Add nonconstruction_targets;
121 rename FINAL to COMPLETE.
122 (record_targets_from_bases): Sanity check we found the binfo;
123 fix COMPLETEP updating.
124 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
125 parameter, fix computing of COMPLETEP.
126 (dump_possible_polymorphic_call_targets): Imrove readability of dump; at
127 LTO time do demangling.
128 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
129 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
131 (gimple_get_virt_method_for_binfo): Likewise.
132 * gimple-fold.h (gimple_get_virt_method_for_binfo,
133 gimple_get_virt_method_for_vtable): Update prototypes.
135 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
138 * genautomata.c (add_presence_absence): Fix typo with
139 {final_}presence_list.
141 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
144 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
145 for VSX/Altivec vectors that land in GPR registers.
147 2014-02-11 Richard Henderson <rth@redhat.com>
148 Jakub Jelinek <jakub@redhat.com>
151 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
152 around drhs if type conversion to lacc->type is not useless.
154 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
156 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
158 (cortex-a57.cortex-a53): Likewise.
159 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
161 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
163 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
166 2014-02-11 Renlin Li <Renlin.Li@arm.com>
168 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
169 add_options_for_arm_vfp3.
171 2014-02-11 Jeff Law <law@redhat.com>
174 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
175 object with an undesirable mode.
177 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
180 * config/i386/sol2-9.h: New file.
181 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
182 *-*-solaris2.9*): Use it.
184 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
186 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
187 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
189 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
191 * config/microblaze/microblaze.c: Extend mcpu version format
193 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
195 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
197 2014-02-10 Richard Henderson <rth@redhat.com>
200 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
201 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
202 ms-abi vs -mno-accumulate-outgoing-args.
203 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
204 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
207 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
210 * cfgexpand.c (expand_asm_operands): Attach source location to
211 ASM_INPUT rtx objects.
212 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
214 2014-02-10 Nick Clifton <nickc@redhat.com>
216 * config/mn10300/mn10300.c (popcount): New function.
217 (mn10300_expand_prologue): Include saved registers in stack usage
220 2014-02-10 Jeff Law <law@redhat.com>
223 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
224 when changing the SET_DEST of a prior insn to avoid an input
227 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
229 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
230 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
231 -mcall-openbsd, or -mcall-linux.
232 (CC1_ENDIAN_BIG_SPEC): Remove.
233 (CC1_ENDIAN_LITTLE_SPEC): Remove.
234 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
235 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
236 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
237 and %cc1_endian_default.
238 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
240 2014-02-10 Richard Biener <rguenther@suse.de>
242 PR tree-optimization/60115
243 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
244 MEM_REF handling. Properly verify that the accesses are not
245 out of the objects bound.
247 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
249 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
252 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
254 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
255 proper constants and fix formatting.
256 (possible_polymorphic_call_targets): Fix formatting.
258 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
259 Ilya Tocar <ilya.tocar@intel.com>
261 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
262 (_mm512_loadu_epi32): Renamed into...
263 (_mm512_loadu_si512): This.
264 (_mm512_storeu_epi32): Renamed into...
265 (_mm512_storeu_si512): This.
266 (_mm512_maskz_ceil_ps): Removed.
267 (_mm512_maskz_ceil_pd): Ditto.
268 (_mm512_maskz_floor_ps): Ditto.
269 (_mm512_maskz_floor_pd): Ditto.
270 (_mm512_floor_round_ps): Ditto.
271 (_mm512_floor_round_pd): Ditto.
272 (_mm512_ceil_round_ps): Ditto.
273 (_mm512_ceil_round_pd): Ditto.
274 (_mm512_mask_floor_round_ps): Ditto.
275 (_mm512_mask_floor_round_pd): Ditto.
276 (_mm512_mask_ceil_round_ps): Ditto.
277 (_mm512_mask_ceil_round_pd): Ditto.
278 (_mm512_maskz_floor_round_ps): Ditto.
279 (_mm512_maskz_floor_round_pd): Ditto.
280 (_mm512_maskz_ceil_round_ps): Ditto.
281 (_mm512_maskz_ceil_round_pd): Ditto.
282 (_mm512_expand_pd): Ditto.
283 (_mm512_expand_ps): Ditto.
284 * config/i386/i386.c (ix86_builtins): Remove
285 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
287 * config/i386/predicates.md (const1256_operand): New.
288 (const_1_to_2_operand): Ditto.
289 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
290 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
291 (*avx512pf_gatherpf<mode>sf): Ditto.
292 (avx512pf_gatherpf<mode>df): Ditto.
293 (*avx512pf_gatherpf<mode>df_mask): Ditto.
294 (*avx512pf_gatherpf<mode>df): Ditto.
295 (avx512pf_scatterpf<mode>sf): Ditto.
296 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
297 (*avx512pf_scatterpf<mode>sf): Ditto.
298 (avx512pf_scatterpf<mode>df): Ditto.
299 (*avx512pf_scatterpf<mode>df_mask): Ditto.
300 (*avx512pf_scatterpf<mode>df): Ditto.
301 (avx512f_expand<mode>): Removed.
302 (<shift_insn><mode>3<mask_name>): Change predicate type.
304 2014-02-08 Jakub Jelinek <jakub@redhat.com>
306 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
307 not at the end of datarefs vector use ordered_remove to avoid
308 reordering datarefs vector.
311 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
312 mark local addressable non-static vars as GOVD_PRIVATE
313 instead of GOVD_LOCAL.
314 * omp-low.c (lower_omp_for): Move gimple_bind_vars
315 and BLOCK_VARS of gimple_bind_block to new_stmt rather
319 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
320 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
321 assume_aligned or alloc_align attributes.
322 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
323 arguments. Handle also assume_aligned and alloc_align attributes.
324 (evaluate_stmt): Adjust bit_value_assume_aligned caller.
325 Handle calls to functions with assume_aligned or alloc_align
327 * doc/extend.texi: Document assume_aligned and alloc_align
330 2014-02-08 Terry Guo <terry.guo@arm.com>
332 * doc/invoke.texi: Document ARM -march=armv7e-m.
334 2014-02-08 Jakub Jelinek <jakub@redhat.com>
336 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
337 flag on __cilkrts_rethrow builtin.
340 * ipa-cp.c (determine_versionability): Fail at -O0
341 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
342 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
345 2014-02-04 Jakub Jelinek <jakub@redhat.com>
348 * tree-inline.c (copy_forbidden): Fail for
349 __attribute__((optimize (0))) functions.
351 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
353 * varpool.c: Include pointer-set.h.
354 (varpool_remove_unreferenced_decls): Variables in other partitions
355 will not be output; be however careful to not lose information
358 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
360 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
361 lookup in the vtable constructor.
363 2014-02-07 Jeff Law <law@redhat.com>
366 * config/m68k/m68k.md (ashldi_extsi): Turn into a
367 define_insn_and_split.
369 * ipa-inline.c (inline_small_functions): Fix typos.
371 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
373 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
374 (s390_can_use_return_insn): Declare.
375 * config/s390/s390.h (EPILOGUE_USES): Define.
376 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
378 (s390_chunkify_start): Handle return JUMP_LABELs.
379 (s390_early_mach): Emit a main_pool instruction on the entry edge.
380 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
381 (s390_can_use_return_insn): New functions.
382 (s390_fix_long_loop_prediction): Handle conditional returns.
383 (TARGET_SET_UP_BY_PROLOGUE): Define.
384 * config/s390/s390.md (ANY_RETURN): New code iterator.
385 (*creturn, *csimple_return, return, simple_return): New patterns.
387 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
389 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
390 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
391 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
392 REG_CFA_RESTORE list when deciding not to restore a register.
394 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
396 * config/s390/s390.c: Include tree-pass.h and context.h.
397 (s390_early_mach): New function, split out from...
398 (s390_emit_prologue): ...here.
399 (pass_data_s390_early_mach): New pass structure.
400 (pass_s390_early_mach): New class.
401 (s390_option_override): Create and register early_mach pass.
404 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
406 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
407 to match for the exit block.
409 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
411 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
412 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
413 Reject misaligned operands.
415 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
417 * optabs.c (expand_atomic_compare_and_swap): Allow expander to
420 2014-02-07 Richard Biener <rguenther@suse.de>
423 * gimple-low.c (lower_builtin_posix_memalign): New function.
424 (lower_stmt): Call it to lower posix_memalign in a way
425 to make alignment info accessible.
427 2014-02-07 Jakub Jelinek <jakub@redhat.com>
430 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
431 __builtin_setjmp_receiver.
433 2014-02-07 Richard Biener <rguenther@suse.de>
436 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
437 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
438 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
439 Handle BUILT_IN_POSIX_MEMALIGN.
440 (find_func_clobbers): Likewise.
441 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
442 (call_may_clobber_ref_p_1): Likewise.
444 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
447 * ipa-devirt.c (record_target_from_binfo): Remove overactive sanity check.
449 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
452 * lto-cgraph.c (lto_output_node): Use
453 symtab_get_symbol_partitioning_class.
454 (lto_output_varpool_node): likewise.
455 (symtab_get_symbol_partitioning_class): Move here from
457 * cgraph.h (symbol_partitioning_class): Likewise.
458 (symtab_get_symbol_partitioning_class): Declare.
460 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
462 * ggc.h (ggc_internal_cleared_alloc): New macro.
463 * vec.h (vec_safe_copy): Handle memory stats.
464 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
465 * target-globals.c (save_target_globals): Likewise.
467 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
470 * expr.c (emit_move_resolve_push): Export; be bit more selective
471 on when to clear alias set.
472 * expr.h (emit_move_resolve_push): Declare.
473 * function.h (struct function): Add tail_call_marked.
474 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
475 * config/i386/i386-protos.h (ix86_expand_push): Remove.
476 * config/i386/i386.md (TImode move expander): De not call
478 (FP push expanders): Preserve memory attributes.
479 * config/i386/sse.md (push<mode>1): Remove.
480 * config/i386/i386.c (ix86_expand_vector_move): Handle push
482 (ix86_expand_push): Remove.
483 * config/i386/mmx.md (push<mode>1): Remove.
485 2014-02-06 Jakub Jelinek <jakub@redhat.com>
487 PR rtl-optimization/60030
488 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
489 lopart with paradoxical subreg before shifting it up by hprec.
491 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
493 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
494 Remove extra newline at end of file.
495 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
496 (arm_issue_rate): Handle cortexa57.
497 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
498 (cortex-a57.cortex-a53): Likewise.
500 2014-02-06 Jakub Jelinek <jakub@redhat.com>
503 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
504 don't record in REG_FRAME_RELATED_EXPR registers not set in that
506 (arm_expand_prologue): Adjust all callers.
507 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
508 info, registers also at the lowest numbered registers side. Use
509 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
513 * var-tracking.c (adjust_mems): Before adding a SET
514 to amd->side_effects, adjust it's SET_SRC using
515 simplify_replace_fn_rtx.
517 2014-02-06 Alan Modra <amodra@gmail.com>
520 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
521 change SDmode to DDmode when lra_in_progress.
523 2014-02-06 Jakub Jelinek <jakub@redhat.com>
526 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
527 free_data_ref on the dr first, and before goto again also set dr
528 to the next dr. For simd_lane_access, free old datarefs[i] before
529 overwriting it. For get_vectype_for_scalar_type failure, don't
530 free_data_ref if simd_lane_access.
532 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
535 * tree.h (opts_for_fn): New inline function.
536 (opt_for_fn): Define.
537 * config/i386/i386.c (ix86_function_regparm): Use
538 opt_for_fn (decl, optimize) instead of optimize.
540 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
542 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
543 for SYMBOL_REF in large memory model.
545 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
547 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
549 (cortex-a57): Likewise.
550 (cortex-a57.cortex-a53): Likewise.
552 2014-02-06 Yury Gribov <y.gribov@samsung.com>
553 Kugan Vivekanandarajah <kuganv@linaro.org>
555 * config/arm/arm.c (arm_vector_alignment_reachable): Check
557 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
559 2014-02-06 Richard Biener <rguenther@suse.de>
561 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
562 set_loop_copy and initialize_original_copy_tables.
564 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
566 * config/aarch64/aarch64-simd.md
567 (aarch64_ashr_simddi): Change QI to SI.
569 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
570 Jakub Jelinek <jakub@redhat.com>
573 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
576 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
578 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
579 CODE_FOR_altivec_vpku[hw]um to
580 CODE_FOR_altivec_vpku[hw]um_direct.
581 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
582 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
583 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
584 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
586 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
588 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
589 generation for -maltivec=be.
590 (altivec_vsumsws): Simplify redundant test.
592 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
594 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
595 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
596 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
597 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
599 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
601 (altivec_vpks<VI_char>ss): Likewise.
602 (altivec_vpks<VI_char>us): Likewise.
603 (altivec_vpku<VI_char>us): Likewise.
604 (altivec_vpku<VI_char>um): Likewise.
605 (altivec_vpku<VI_char>um_direct): New (copy of
606 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
608 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
609 target is little endian and -maltivec=be is not specified.
610 (*altivec_vupkhs<VU_char>_direct): New (copy of
611 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal
613 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
614 target is little endian and -maltivec=be is not specified.
615 (*altivec_vupkls<VU_char>_direct): New (copy of
616 altivec_vupkls<VU_char> that always emits vupkls*, for internal
618 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
619 little endian and -maltivec=be is not specified.
620 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
621 little endian and -maltivec=be is not specified.
623 2014-02-05 Richard Henderson <rth@redhat.com>
626 * combine-stack-adj.c: Revert r206943.
627 * sched-int.h (struct deps_desc): Add last_args_size.
628 * sched-deps.c (init_deps): Initialize it.
629 (sched_analyze_insn): Add OUTPUT dependencies between insns that
630 contain REG_ARGS_SIZE notes.
632 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
634 * lto-cgraph.c (asm_nodes_output): Make global.
635 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
636 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA
638 (driver_handle_option): Handle OPT_fwpa.
640 2014-02-05 Jakub Jelinek <jakub@redhat.com>
643 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
644 a comment typo and formatting issue. If odr_hash hasn't been
645 created, return vNULL and set *completep to false.
648 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
649 bb with no successors.
651 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
654 * doc/invoke.texi (-march): Clarify documentation for ARM.
658 2014-02-05 Richard Biener <rguenther@suse.de>
660 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
661 when not vectorizing because of too many alias checks.
662 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
663 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
665 2014-02-05 Nick Clifton <nickc@redhat.com>
667 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
668 accept extended registers in any mode when compiling for the
671 2014-02-05 Yury Gribov <y.gribov@samsung.com>
673 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
674 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
675 sanitization attributes.
676 (can_inline_edge_p): Likewise.
677 (sanitize_attrs_match_for_inline_p): New function.
679 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
681 * ipa-prop.c (detect_type_change): Shor circuit testing of
682 type changes on THIS pointer.
684 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
687 * config/pa/pa.c (legitimize_tls_address): Return original address
688 if not passed a SYMBOL_REF rtx.
689 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
691 (pa_emit_move_sequence): Simplify TLS source operands.
692 (pa_legitimate_constant_p): Reject all TLS constants.
693 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
694 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
696 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
698 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
699 groups when we know they are controlled by LTO.
700 * varasm.c (default_binds_local_p_1): If object is in other partition,
701 it will be resolved locally.
703 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
705 * config/host-linux.c (linux_gt_pch_use_address): Don't
706 use SSIZE_MAX because it is not always defined.
708 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
711 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
712 threshold for pseudo splitting.
713 (update_ebb_live_info): Process call argument hard registers and
714 hard registers from insn definition too.
715 (max_small_class_regs_num): New constant.
716 (inherit_in_ebb): Update live hard regs through EBBs. Update
717 reloads_num only for small register classes. Don't split for
720 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
723 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
726 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
728 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default visibility is safe.
730 2014-02-04 Marek Polacek <polacek@redhat.com>
732 * gdbinit.in (pel): Define.
734 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
736 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
739 2014-02-04 Richard Biener <rguenther@suse.de>
742 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
743 in function context local.
744 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
745 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
746 similar to LTO_imported_decl_ref.
748 2014-02-04 Jakub Jelinek <jakub@redhat.com>
750 PR tree-optimization/60002
751 * cgraphclones.c (build_function_decl_skip_args): Clear
754 PR tree-optimization/60023
755 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
756 false to gsi_replace.
757 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
758 has been in some EH region and vec_stmt could throw, add
759 vec_stmt into the same EH region.
760 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
761 has no lhs, ignore it.
762 * internal-fn.c (expand_MASK_LOAD): Likewise.
765 * tree-inline.c (copy_forbidden): Fail for
766 __attribute__((optimize (0))) functions.
769 * omp-low.c (simd_clone_struct_copy): If from->inbranch
770 is set, copy one less argument.
771 (expand_simd_clones): Don't subtract clone_info->inbranch
772 from simd_clone_struct_alloc argument.
774 PR rtl-optimization/57915
775 * recog.c (simplify_while_replacing): If all unary/binary/relational
776 operation arguments are constant, attempt to simplify those.
779 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
780 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
782 2014-02-04 Richard Biener <rguenther@suse.de>
784 PR tree-optimization/60012
785 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
786 TBAA disambiguation to all DDRs.
788 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
791 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
792 (LINK_SPEC): Use it for -shared, -shared-libgcc.
794 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
797 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
799 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
801 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
802 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
804 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
807 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
808 to figure out targets of polymorphic calls with known decl.
809 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
810 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
811 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
812 (get_polymorphic_call_info): ... here.
813 (get_polymorphic_call_info_from_invariant): New function.
815 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
817 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
818 lookup via vtable pointer; check for type consistency
819 and turn inconsitent facts into UNREACHABLE.
820 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
821 * gimple-fold.c (gimple_get_virt_method_for_vtable):
822 Do not ICE on type inconsistent querries; return UNREACHABLE
825 2014-02-03 Richard Henderson <rth@twiddle.net>
828 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
829 already processed this node.
830 (normalize_one_pred_1): Pass along mark_set.
831 (normalize_one_pred): Create and destroy a pointer_set_t.
832 (normalize_one_pred_chain): Likewise.
834 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
836 PR gcov-profile/58602
837 * gcc/gcov-io.c (gcov_open): Open with truncation when mode<0
839 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
842 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on -fno-devirtualize;
843 Try to devirtualize by the knowledge of virtual table pointer given by
844 aggregate propagation.
845 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
846 ipa_print_node_jump_functions): Dump also offset that
847 is relevant for polymorphic calls.
848 (determine_known_aggregate_parts): Add arg_type parameter; use it
849 instead of determining the type from pointer type.
850 (ipa_compute_jump_functions_for_edge): Update call of
851 determine_known_aggregate_parts.
852 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
853 (gimple_get_virt_method_for_binfo): ... here; simplify using
854 vtable_pointer_value_to_vtable.
855 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
856 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
857 to unsigned HOST_WIDE_INT; Use vtable_pointer_value_to_vtable.
858 (vtable_pointer_value_to_vtable): Break out from ...; handle also
860 (vtable_pointer_value_to_binfo): ... here.
861 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
863 2014-02-03 Teresa Johnson <tejohnson@google.com>
865 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
866 redef of outer loop index variable.
868 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
872 * doc/extend.texi (Function Attributes): Typo.
874 2014-02-03 Cong Hou <congh@google.com>
876 PR tree-optimization/60000
877 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
878 if the vectorized statement is a store. A store statement can only
879 appear at the end of pattern statements.
881 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
883 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
884 (ix86_option_override_internal): Default long double to 64-bit for
885 32-bit Bionic and to 128-bit for 64-bit Bionic.
887 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
888 TARGET_LONG_DOUBLE_128 is true.
889 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
891 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
892 (mlong-double-64): Negate -mlong-double-128.
893 (mlong-double-128): New option.
895 * config/i386/i386-c.c (ix86_target_macros): Define
896 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
898 * doc/invoke.texi: Document -mlong-double-128.
900 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
902 PR rtl-optimization/60024
903 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
905 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
907 * doc/invoke.texi: (fprofile-reorder-functions): Fix typo.
909 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
911 PR rtl-optimization/57662
912 * sel-sched.c (code_motion_path_driver): Do not mark already not
913 existing blocks in the visiting bitmap.
915 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
917 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
918 on the insn being emitted.
920 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
921 Will Deacon <will.deacon@arm.com>
923 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
925 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
927 * config/arm/arm-tables.opt: Regenerate.
929 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
931 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
932 for vector types other than V16QImode.
933 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
934 define_expand, and call altivec_expand_vec_perm_le when producing
935 code with little endian element order.
936 (*altivec_vperm_<mode>_internal): New insn having previous
937 behavior of altivec_vperm_<mode>.
938 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
939 altivec_expand_vec_perm_le when producing code with little endian
941 (*altivec_vperm_<mode>_uns_internal): New insn having previous
942 behavior of altivec_vperm_<mode>_uns.
944 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
946 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
947 (altivec_vsumsws): Add handling for -maltivec=be with a little
949 (altivec_vsumsws_direct): New.
950 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
953 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
955 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
956 vtable_pointer_value_to_binfo): New functions.
957 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
958 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
960 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
962 * config/nios2/nios2.md (load_got_register): Initialize GOT
963 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
964 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
966 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
968 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
969 preserverd by passthrough, do not propagate the type.
971 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
973 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
974 (mips_atomic_assign_expand_fenv): New function.
975 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
977 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
979 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
980 (__builtin_mips_set_fcsr): Likewise.
981 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
983 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
984 (mips16_expand_set_fcsr): Likewise.
985 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
986 (mips16_set_fcsr_stub): Likewise.
987 (mips16_get_fcsr_one_only_stub): New class.
988 (mips16_set_fcsr_one_only_stub): Likewise.
989 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
990 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
991 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
992 (hard_float): New availability predicate.
993 (mips_builtins): Add get_fcsr and set_fcsr.
994 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
995 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
996 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
997 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
998 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
1001 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
1003 * config/mips/mips.c (mips_one_only_stub): New class.
1004 (mips_need_mips16_rdhwr_p): Replace with...
1005 (mips16_rdhwr_stub): ...this new variable.
1006 (mips16_stub_call_address): New function.
1007 (mips16_rdhwr_one_only_stub): New class.
1008 (mips_expand_thread_pointer): Use mips16_stub_call_address.
1009 (mips_output_mips16_rdhwr): Delete.
1010 (mips_finish_stub): New function.
1011 (mips_code_end): Use it to handle rdhwr stubs.
1013 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
1016 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
1017 when calculating size of integer atomic types.
1019 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
1021 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
1023 2014-02-01 Jakub Jelinek <jakub@redhat.com>
1025 PR tree-optimization/60003
1026 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
1027 * profile.c (branch_prob): Use gimple_call_builtin_p
1028 to check for BUILT_IN_SETJMP_RECEIVER.
1029 * tree-inline.c (copy_bb): Call notice_special_calls.
1031 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
1034 * lra-constraints.c (process_alt_operands): Update reload_sum only
1037 2014-01-31 Richard Henderson <rth@redhat.com>
1040 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
1041 until after else_eh is processed.
1043 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
1045 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
1046 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
1047 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
1048 in smmintrin.h, remove them.
1049 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
1050 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
1051 * config/i386/i386.md (ROUND_SAE): Fix value.
1052 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
1053 (const48_operand): New.
1054 * config/i386/subst.md (round), (round_expand): Use
1055 const_4_or_8_to_11_operand.
1056 (round_saeonly), (round_saeonly_expand): Use const48_operand.
1058 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
1060 * config/i386/constraints.md (Yk): Swap meaning with k.
1061 * config/i386/i386.md (movhi_internal): Change Yk to k.
1062 (movqi_internal): Ditto.
1063 (*k<logic><mode>): Ditto.
1066 (kandn<mode>): Ditto.
1067 (*<code>hi_1): Ditto.
1068 (*<code>qi_1): Ditto.
1069 (kxnor<mode>): Ditto.
1070 (kortestzhi): Ditto.
1071 (kortestchi): Ditto.
1073 (*one_cmplhi2_1): Ditto.
1074 (*one_cmplqi2_1): Ditto.
1075 * config/i386/sse.md (): Change k to Yk.
1076 (avx512f_load<mode>_mask): Ditto.
1077 (avx512f_blendm<mode>): Ditto.
1078 (avx512f_store<mode>_mask): Ditto.
1079 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
1080 (avx512f_storedqu<mode>_mask): Ditto.
1081 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
1083 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
1084 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
1085 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
1086 (avx512f_maskcmp<mode>3): Ditto.
1087 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
1088 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
1089 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
1090 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
1091 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
1092 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
1093 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
1094 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
1095 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
1096 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
1097 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
1098 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
1099 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
1100 (vec_extract_lo_<mode>_maskm): Ditto.
1101 (vec_extract_hi_<mode>_maskm): Ditto.
1102 (avx512f_vternlog<mode>_mask): Ditto.
1103 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
1104 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
1105 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
1106 (avx512f_<code>v8div16qi2_mask): Ditto.
1107 (avx512f_<code>v8div16qi2_mask_store): Ditto.
1108 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
1109 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
1110 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
1111 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
1112 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
1113 (*avx512pf_gatherpf<mode>df_mask): Ditto.
1114 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
1115 (*avx512pf_scatterpf<mode>df_mask): Ditto.
1116 (avx512cd_maskb_vec_dupv8di): Ditto.
1117 (avx512cd_maskw_vec_dupv16si): Ditto.
1118 (avx512f_vpermi2var<mode>3_maskz): Ditto.
1119 (avx512f_vpermi2var<mode>3_mask): Ditto.
1120 (avx512f_vpermi2var<mode>3_mask): Ditto.
1121 (avx512f_vpermt2var<mode>3_maskz): Ditto.
1122 (*avx512f_gathersi<mode>): Ditto.
1123 (*avx512f_gathersi<mode>_2): Ditto.
1124 (*avx512f_gatherdi<mode>): Ditto.
1125 (*avx512f_gatherdi<mode>_2): Ditto.
1126 (*avx512f_scattersi<mode>): Ditto.
1127 (*avx512f_scatterdi<mode>): Ditto.
1128 (avx512f_compress<mode>_mask): Ditto.
1129 (avx512f_compressstore<mode>_mask): Ditto.
1130 (avx512f_expand<mode>_mask): Ditto.
1131 * config/i386/subst.md (mask): Change k to Yk.
1132 (mask_scalar_merge): Ditto.
1135 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
1137 * doc/extend.texi (Vector Extensions): Document ?: in C++.
1139 2014-01-31 Richard Biener <rguenther@suse.de>
1142 * builtins.c (fold_builtin_memory_op): Make sure to not
1143 use a floating-point mode or a boolean or enumeral type for
1146 2014-01-30 DJ Delorie <dj@redhat.com>
1148 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
1149 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
1150 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
1151 whenever main() has an epilogue.
1153 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1155 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
1156 unused variable "field".
1157 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
1158 (vsx_mergeh_<mode>): Likewise.
1159 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
1160 (altivec_vmrghh): Likewise.
1161 (altivec_vmrghw): Likewise.
1162 (altivec_vmrglb): Likewise.
1163 (altivec_vmrglh): Likewise.
1164 (altivec_vmrglw): Likewise.
1165 (altivec_vspltb): Add missing uses.
1166 (altivec_vsplth): Likewise.
1167 (altivec_vspltw): Likewise.
1168 (altivec_vspltsf): Likewise.
1170 2014-01-30 Jakub Jelinek <jakub@redhat.com>
1173 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
1174 frame related instructions.
1176 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
1178 PR rtl-optimization/59959
1179 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
1180 any reload of register whose subreg is invalid.
1182 2014-01-30 Jakub Jelinek <jakub@redhat.com>
1184 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
1185 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
1186 Add missing return type - void.
1188 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1190 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
1191 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
1192 remove element index adjustment for endian (now handled in vsx.md
1194 (altivec_expand_vec_perm_const): Use
1195 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
1196 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
1197 (vsx_xxspltw_<mode>): Adjust element index for little endian.
1198 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
1199 define_expand and a new define_insn *altivec_vspltb_internal;
1200 adjust for -maltivec=be on a little endian target.
1201 (altivec_vspltb_direct): New.
1202 (altivec_vsplth): Divide into a define_expand and a new
1203 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
1204 little endian target.
1205 (altivec_vsplth_direct): New.
1206 (altivec_vspltw): Divide into a define_expand and a new
1207 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
1208 little endian target.
1209 (altivec_vspltw_direct): New.
1210 (altivec_vspltsf): Divide into a define_expand and a new
1211 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
1212 a little endian target.
1214 2014-01-30 Richard Biener <rguenther@suse.de>
1216 PR tree-optimization/59993
1217 * tree-ssa-forwprop.c (associate_pointerplus): Check we
1218 can propagate form the earlier stmt and avoid the transform
1219 when the intermediate result is needed.
1221 2014-01-30 Alangi Derick <alangiderick@gmail.com>
1223 * README.Portability: Fix typo.
1225 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
1227 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
1228 comparison_operator with ordered_comparison_operator.
1230 2014-01-30 Nick Clifton <nickc@redhat.com>
1232 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
1233 Rename to mn10300_store_multiple_regs.
1234 * config/mn10300/mn10300.c: Likewise.
1235 * config/mn10300/mn10300.md (store_movm): Fix typo: call
1236 store_multiple_regs.
1237 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
1238 Call mn10300_store_multiple_regs.
1240 2014-01-30 Nick Clifton <nickc@redhat.com>
1241 DJ Delorie <dj@redhat.com>
1243 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
1244 %fp 2 to keep registers after it properly word-aligned.
1245 (rl78_alloc_physical_registers_umul): Handle the case where both
1246 input operands are the same.
1248 2014-01-30 Richard Biener <rguenther@suse.de>
1250 PR tree-optimization/59903
1251 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
1254 2014-01-30 Jason Merrill <jason@redhat.com>
1257 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
1260 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
1262 2014-01-30 Richard Biener <rguenther@suse.de>
1264 PR tree-optimization/59951
1265 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
1267 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
1270 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
1271 SFmode to DFmode case.
1273 2014-01-29 DJ Delorie <dj@redhat.com>
1275 * config/msp430/msp430.opt (-minrt): New.
1276 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
1278 (ENDFILE_SPEC): Likewise.
1280 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
1282 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
1283 (estimate_function_body_sizes): Use it.
1285 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
1288 * dwarf2out.c (is_cxx_auto): New.
1289 (is_base_type): Use it.
1290 (gen_type_die_with_usage): Likewise.
1292 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1294 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
1295 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
1296 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
1297 -maltivec=be with LE targets.
1298 (vsx_mergeh_<mode>): Likewise.
1299 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
1300 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
1301 (altivec_vmrghb): Replace with define_expand and new
1302 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
1303 (altivec_vmrghb_direct): New define_insn.
1304 (altivec_vmrghh): Replace with define_expand and new
1305 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
1306 (altivec_vmrghh_direct): New define_insn.
1307 (altivec_vmrghw): Replace with define_expand and new
1308 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
1309 (altivec_vmrghw_direct): New define_insn.
1310 (*altivec_vmrghsf): Adjust for endianness.
1311 (altivec_vmrglb): Replace with define_expand and new
1312 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
1313 (altivec_vmrglb_direct): New define_insn.
1314 (altivec_vmrglh): Replace with define_expand and new
1315 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
1316 (altivec_vmrglh_direct): New define_insn.
1317 (altivec_vmrglw): Replace with define_expand and new
1318 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
1319 (altivec_vmrglw_direct): New define_insn.
1320 (*altivec_vmrglsf): Adjust for endianness.
1321 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
1322 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
1323 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
1324 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
1325 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
1326 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
1327 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
1328 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
1330 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
1332 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
1333 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
1336 2014-01-29 Richard Biener <rguenther@suse.de>
1338 PR tree-optimization/58742
1339 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
1340 associate_pointerplus_align.
1341 (associate_pointerplus_diff): New function.
1342 (associate_pointerplus): Likewise. Call associate_pointerplus_align
1343 and associate_pointerplus_diff.
1345 2014-01-29 Richard Biener <rguenther@suse.de>
1347 * lto-streamer.h (LTO_major_version): Bump to 3.
1348 (LTO_minor_version): Reset to 0.
1350 2014-01-29 Renlin Li <Renlin.Li@arm.com>
1352 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
1353 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
1354 (arm_file_start): Generate correct asm header for armv7ve.
1355 * config/arm/bpabi.h: Add multilib support for armv7ve.
1356 * config/arm/driver-arm.c: Change the architectures of cortex-a7
1357 and cortex-a15 to armv7ve.
1358 * config/arm/t-aprofile: Add multilib support for armv7ve.
1359 * doc/invoke.texi: Document -march=armv7ve.
1361 2014-01-29 Richard Biener <rguenther@suse.de>
1363 PR tree-optimization/58742
1364 * tree-ssa-forwprop.c (associate_plusminus): Return true
1365 if we changed sth, defer EH cleanup to ...
1366 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
1367 (simplify_mult): New function.
1369 2014-01-29 Jakub Jelinek <jakub@redhat.com>
1372 PR tree-optimization/59920
1373 * tree.c (build_common_builtin_nodes): Remove
1374 __builtin_setjmp_dispatcher initialization.
1375 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
1376 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
1377 instead of gsi_after_labels + manually skipping debug stmts.
1378 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
1379 ignore bbs with IFN_ABNORMAL_DISPATCHER.
1380 * tree-inline.c (copy_edges_for_bb): Remove
1381 can_make_abnormal_goto argument, instead add abnormal_goto_dest
1382 argument. Ignore computed_goto_p stmts. Don't call
1383 make_abnormal_goto_edges. If a call might need abnormal edges
1384 for non-local gotos, see if it already has an edge to
1385 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
1386 with true argument, don't do anything then, otherwise add
1387 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
1388 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
1390 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
1391 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
1392 (lower_stmt): Don't set data->calls_builtin_setjmp.
1393 (lower_builtin_setjmp): Adjust comment.
1394 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
1395 * tree-cfg.c (found_computed_goto): Remove.
1396 (factor_computed_gotos): Remove.
1397 (make_goto_expr_edges): Return bool, true for computed gotos.
1398 Don't call make_abnormal_goto_edges.
1399 (build_gimple_cfg): Don't set found_computed_goto, don't call
1400 factor_computed_gotos.
1401 (computed_goto_p): No longer static.
1402 (make_blocks): Don't set found_computed_goto.
1403 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
1404 (make_edges): If make_goto_expr_edges returns true, push bb
1405 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
1406 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
1407 vector. Record mapping between bbs and OpenMP regions if there
1408 are any, adjust make_gimple_omp_edges caller. Call
1409 handle_abnormal_edges.
1410 (make_abnormal_goto_edges): Remove.
1411 * tree-cfg.h (make_abnormal_goto_edges): Remove.
1412 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
1413 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
1414 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
1415 * internal-fn.def (ABNORMAL_DISPATCHER): New.
1416 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
1417 filling *region also set *region_idx to (*region)->entry->index.
1420 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
1421 For REGs set ORIGINAL_REGNO.
1423 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
1425 * doc/md.texi: Mention that a target shouldn't implement
1426 vec_widen_(s|u)mul_even/odd pair if it is less efficient
1429 2014-01-29 Jakub Jelinek <jakub@redhat.com>
1431 PR tree-optimization/59594
1432 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
1433 a copy of the datarefs vector rather than the vector itself.
1435 2014-01-28 Jason Merrill <jason@redhat.com>
1438 * dwarf2out.c (auto_die): New static.
1439 (gen_type_die_with_usage): Handle C++1y 'auto'.
1440 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
1443 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
1446 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
1447 (SPEC_X32): Likewise.
1448 (SPEC_64): Likewise.
1449 * config/i386/i386.c (ix86_option_override_internal): Turn off
1450 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
1452 (x86_file_start): Output .code16gcc for TARGET_16BIT.
1453 * config/i386/i386.h (TARGET_16BIT): New macro.
1454 (TARGET_16BIT_P): Likewise.
1455 * config/i386/i386.opt: Add m16.
1456 * doc/invoke.texi: Document -m16.
1458 2014-01-28 Jakub Jelinek <jakub@redhat.com>
1460 PR preprocessor/59935
1461 * input.c (location_get_source_line): Bail out on when line number
1462 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
1464 2014-01-28 Richard Biener <rguenther@suse.de>
1466 PR tree-optimization/58742
1467 * tree-ssa-forwprop.c (associate_plusminus): Handle
1468 pointer subtraction of the form (T)(P + A) - (T)P.
1470 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1472 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
1475 2014-01-28 Richard Biener <rguenther@suse.de>
1478 2014-01-28 Richard Biener <rguenther@suse.de>
1480 PR rtl-optimization/45364
1481 PR rtl-optimization/59890
1482 * var-tracking.c (local_get_addr_clear_given_value): Handle
1483 already cleared slot.
1484 (val_reset): Handle not allocated local_get_addr_cache.
1485 (vt_find_locations): Use post-order on the inverted CFG.
1487 2014-01-28 Richard Biener <rguenther@suse.de>
1489 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
1491 2014-01-28 Richard Biener <rguenther@suse.de>
1493 PR rtl-optimization/45364
1494 PR rtl-optimization/59890
1495 * var-tracking.c (local_get_addr_clear_given_value): Handle
1496 already cleared slot.
1497 (val_reset): Handle not allocated local_get_addr_cache.
1498 (vt_find_locations): Use post-order on the inverted CFG.
1500 2014-01-28 Alan Modra <amodra@gmail.com>
1502 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
1503 * configure.ac <recursive call for build != host>: Define
1504 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
1505 and LD_FOR_BUILD too.
1506 * configure: Regenerate.
1508 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
1510 * config/i386/i386.c (get_builtin_code_for_version): Separate
1511 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
1512 Broadwell from Haswell.
1514 2014-01-27 Steve Ellcey <sellcey@mips.com>
1516 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
1517 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
1518 * config/mips/mips.c (mips_option_override): Change setting
1520 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
1521 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
1522 Change from Mask to Var.
1524 2014-01-27 Jeff Law <law@redhat.com>
1526 * ipa-inline.c (inline_small_functions): Fix typo.
1528 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
1530 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
1531 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
1532 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
1533 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
1534 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
1535 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
1536 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
1537 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
1538 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
1539 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
1540 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
1541 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
1542 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
1543 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
1544 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
1545 (_mm512_storeu_epi64): Ditto.
1546 (_mm512_cmpge_epi32_mask): Ditto.
1547 (_mm512_cmpge_epu32_mask): Ditto.
1548 (_mm512_cmpge_epi64_mask): Ditto.
1549 (_mm512_cmpge_epu64_mask): Ditto.
1550 (_mm512_cmple_epi32_mask): Ditto.
1551 (_mm512_cmple_epu32_mask): Ditto.
1552 (_mm512_cmple_epi64_mask): Ditto.
1553 (_mm512_cmple_epu64_mask): Ditto.
1554 (_mm512_cmplt_epi32_mask): Ditto.
1555 (_mm512_cmplt_epu32_mask): Ditto.
1556 (_mm512_cmplt_epi64_mask): Ditto.
1557 (_mm512_cmplt_epu64_mask): Ditto.
1558 (_mm512_cmpneq_epi32_mask): Ditto.
1559 (_mm512_cmpneq_epu32_mask): Ditto.
1560 (_mm512_cmpneq_epi64_mask): Ditto.
1561 (_mm512_cmpneq_epu64_mask): Ditto.
1562 (_mm512_expand_pd): Ditto.
1563 (_mm512_expand_ps): Ditto.
1564 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
1565 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
1566 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
1567 * config/i386/i386.c (ix86_builtins): Add
1568 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
1569 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
1570 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
1571 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
1572 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
1573 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
1574 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
1575 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
1576 IX86_BUILTIN_PMOVUSQW512_MEM.
1577 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
1578 __builtin_ia32_pmovsqd512mem_mask,
1579 __builtin_ia32_pmovqd512mem_mask,
1580 __builtin_ia32_pmovusqw512mem_mask,
1581 __builtin_ia32_pmovsqw512mem_mask,
1582 __builtin_ia32_pmovqw512mem_mask,
1583 __builtin_ia32_pmovusdw512mem_mask,
1584 __builtin_ia32_pmovsdw512mem_mask,
1585 __builtin_ia32_pmovdw512mem_mask,
1586 __builtin_ia32_pmovqb512mem_mask,
1587 __builtin_ia32_pmovusqb512mem_mask,
1588 __builtin_ia32_pmovsqb512mem_mask,
1589 __builtin_ia32_pmovusdb512mem_mask,
1590 __builtin_ia32_pmovsdb512mem_mask,
1591 __builtin_ia32_pmovdb512mem_mask.
1592 (bdesc_args): Add __builtin_ia32_expanddf512,
1593 __builtin_ia32_expandsf512.
1594 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
1595 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
1596 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
1597 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
1598 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
1599 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
1600 (avx512f_<code>v8div16qi2_mask_store): This.
1601 (avx512f_expand<mode>): New.
1603 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
1605 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
1607 (_mm512_mask_prefetch_i64gather_pd): Ditto.
1608 (_mm512_prefetch_i32scatter_pd): Ditto.
1609 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
1610 (_mm512_prefetch_i64scatter_pd): Ditto.
1611 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
1612 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
1613 (_mm512_mask_prefetch_i64gather_ps): Ditto.
1614 (_mm512_prefetch_i32scatter_ps): Ditto.
1615 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
1616 (_mm512_prefetch_i64scatter_ps): Ditto.
1617 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
1618 * config/i386/i386-builtin-types.def: Define
1619 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
1620 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
1621 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
1622 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
1623 IX86_BUILTIN_SCATTERPFQPD.
1624 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
1625 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
1626 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
1627 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
1628 __builtin_ia32_scatterpfqps.
1629 (ix86_expand_builtin): Expand new built-ins.
1630 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
1631 fix memory access data type.
1632 (*avx512pf_gatherpf<mode>_mask): Ditto.
1633 (*avx512pf_gatherpf<mode>): Ditto.
1634 (avx512pf_scatterpf<mode>): Ditto.
1635 (*avx512pf_scatterpf<mode>_mask): Ditto.
1636 (*avx512pf_scatterpf<mode>): Ditto.
1637 (GATHER_SCATTER_SF_MEM_MODE): New.
1638 (avx512pf_gatherpf<mode>df): Ditto.
1639 (*avx512pf_gatherpf<mode>df_mask): Ditto.
1640 (*avx512pf_scatterpf<mode>df): Ditto.
1642 2014-01-27 Jakub Jelinek <jakub@redhat.com>
1645 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
1646 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
1649 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
1651 * common/config/arm/arm-common.c
1652 (arm_rewrite_mcpu): Handle multiple names.
1654 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
1656 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
1658 * gimple-builder.h (create_gimple_tmp): Delete.
1660 2014-01-27 Christian Bruel <christian.bruel@st.com>
1662 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
1665 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
1667 * config/pa/pa.md (call): Generate indirect long calls to non-local
1668 functions when outputing 32-bit code.
1669 (call_value): Likewise except for special call to buggy powf function.
1671 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
1672 portable runtime and PIC indirect calls.
1673 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
1674 and PIC call sequences. Use ldo instead of blr to set return register
1675 in PIC call sequence.
1677 2014-01-25 Walter Lee <walt@tilera.com>
1679 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
1680 avoid clobbering a live register.
1682 2014-01-25 Walter Lee <walt@tilera.com>
1684 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
1685 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
1686 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
1687 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
1689 2014-01-25 Walter Lee <walt@tilera.com>
1691 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
1692 arguments on even registers.
1693 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
1695 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
1696 (BIGGEST_ALIGNMENT): Ditto.
1697 (BIGGEST_FIELD_ALIGNMENT): Ditto.
1699 2014-01-25 Walter Lee <walt@tilera.com>
1701 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
1702 insns before bundling.
1703 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
1705 2014-01-25 Walter Lee <walt@tilera.com>
1707 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
1708 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
1709 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
1711 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
1713 * config/mips/constraints.md (kl): Delete.
1714 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
1715 define expands, using...
1716 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
1717 instructions for MIPS16.
1718 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
1719 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
1721 2014-01-25 Walter Lee <walt@tilera.com>
1723 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
1727 2014-01-25 Walter Lee <walt@tilera.com>
1729 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
1730 (TARGET_EXPAND_TO_RTL_HOOK): Define.
1732 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
1734 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
1737 2014-01-25 Jakub Jelinek <jakub@redhat.com>
1739 * print-rtl.c (in_call_function_usage): New var.
1740 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
1741 EXPR_LIST mode as mode and not as reg note name.
1744 * cfgloopmanip.c (copy_loop_info): If
1745 loop->warned_aggressive_loop_optimizations, make sure
1746 the flag is set in target loop too.
1748 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
1750 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
1752 * builtins.def: Likewise.
1753 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
1754 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
1755 * ira.c (ira_setup_eliminable_regset): Likewise.
1756 * omp-low.c (gate_expand_omp): Likewise.
1757 (execute_lower_omp): Likewise.
1758 (diagnose_sb_0): Likewise.
1759 (gate_diagnose_omp_blocks): Likewise.
1760 (simd_clone_clauses_extract): Likewise.
1763 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1765 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
1766 correction for little endian...
1767 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
1770 2014-01-24 Jeff Law <law@redhat.com>
1772 PR tree-optimization/59919
1773 * tree-vrp.c (find_assert_locations_1): Do not register asserts
1774 for non-returning calls.
1776 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
1778 * common/config/aarch64/aarch64-common.c
1779 (aarch64_rewrite_mcpu): Handle multiple names.
1780 * config/aarch64/aarch64.h
1781 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
1783 2014-01-24 Dodji Seketeli <dodji@redhat.com>
1785 * input.c (add_file_to_cache_tab): Handle the case where fopen
1788 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
1791 * config/i386/i386.md (pushsf splitter): Get stack adjustment
1792 from push operand if code of push isn't PRE_DEC.
1794 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
1797 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
1798 -mquad-memory-atomic. Update -mquad-memory documentation to say
1799 it is only used for non-atomic loads/stores.
1801 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
1802 -mquad-memory or -mquad-memory-atomic switches.
1804 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
1805 -mquad-memory-atomic to ISA 2.07 support.
1807 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
1808 to separate support of normal quad word memory operations (ldq, stq)
1809 from the atomic quad word memory operations.
1811 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
1812 support to separate non-atomic quad word operations from atomic
1813 quad word operations. Disable non-atomic quad word operations in
1814 little endian mode so that we don't have to swap words after the
1815 load and before the store.
1816 (quad_load_store_p): Add comment about atomic quad word support.
1817 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
1818 options printed with -mdebug=reg.
1820 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
1821 -mquad-memory-atomic as the test for whether we have quad word
1822 atomic instructions.
1823 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
1824 or -mp8-vector are used, allow byte/half-word atomic operations.
1826 * config/rs6000/sync.md (load_lockedti): Insure that the address
1827 is a proper indexed or indirect address for the lqarx instruction.
1828 On little endian systems, swap the hi/lo registers after the lqarx
1830 (load_lockedpti): Use indexed_or_indirect_operand predicate to
1831 insure the address is valid for the lqarx instruction.
1832 (store_conditionalti): Insure that the address is a proper indexed
1833 or indirect address for the stqcrx. instruction. On little endian
1834 systems, swap the hi/lo registers before doing the stqcrx.
1836 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
1837 insure the address is valid for the stqcrx. instruction.
1839 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
1840 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
1841 type of quad memory support is available.
1843 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
1846 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
1847 there is a danger of looping.
1849 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
1851 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
1852 force flag_ira_loop_pressure if set via command line.
1854 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
1856 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
1857 (ashr_simd): New builtin handling DI mode.
1858 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
1859 (aarch64_sshr_simddi): New match pattern.
1860 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
1861 (vshrd_n_s64): Likewise.
1862 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
1864 2014-01-23 Nick Clifton <nickc@redhat.com>
1866 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
1867 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
1868 favour of mcu specific scripts.
1869 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
1872 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
1873 Alex Velenko <Alex.Velenko@arm.com>
1875 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
1876 (vaddv_s16): Likewise.
1877 (vaddv_s32): Likewise.
1878 (vaddv_u8): Likewise.
1879 (vaddv_u16): Likewise.
1880 (vaddv_u32): Likewise.
1881 (vaddvq_s8): Likewise.
1882 (vaddvq_s16): Likewise.
1883 (vaddvq_s32): Likewise.
1884 (vaddvq_s64): Likewise.
1885 (vaddvq_u8): Likewise.
1886 (vaddvq_u16): Likewise.
1887 (vaddvq_u32): Likewise.
1888 (vaddvq_u64): Likewise.
1889 (vaddv_f32): Likewise.
1890 (vaddvq_f32): Likewise.
1891 (vaddvq_f64): Likewise.
1892 (vmaxv_f32): Likewise.
1893 (vmaxv_s8): Likewise.
1894 (vmaxv_s16): Likewise.
1895 (vmaxv_s32): Likewise.
1896 (vmaxv_u8): Likewise.
1897 (vmaxv_u16): Likewise.
1898 (vmaxv_u32): Likewise.
1899 (vmaxvq_f32): Likewise.
1900 (vmaxvq_f64): Likewise.
1901 (vmaxvq_s8): Likewise.
1902 (vmaxvq_s16): Likewise.
1903 (vmaxvq_s32): Likewise.
1904 (vmaxvq_u8): Likewise.
1905 (vmaxvq_u16): Likewise.
1906 (vmaxvq_u32): Likewise.
1907 (vmaxnmv_f32): Likewise.
1908 (vmaxnmvq_f32): Likewise.
1909 (vmaxnmvq_f64): Likewise.
1910 (vminv_f32): Likewise.
1911 (vminv_s8): Likewise.
1912 (vminv_s16): Likewise.
1913 (vminv_s32): Likewise.
1914 (vminv_u8): Likewise.
1915 (vminv_u16): Likewise.
1916 (vminv_u32): Likewise.
1917 (vminvq_f32): Likewise.
1918 (vminvq_f64): Likewise.
1919 (vminvq_s8): Likewise.
1920 (vminvq_s16): Likewise.
1921 (vminvq_s32): Likewise.
1922 (vminvq_u8): Likewise.
1923 (vminvq_u16): Likewise.
1924 (vminvq_u32): Likewise.
1925 (vminnmv_f32): Likewise.
1926 (vminnmvq_f32): Likewise.
1927 (vminnmvq_f64): Likewise.
1929 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
1931 * config/aarch64/aarch64-simd.md
1932 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
1933 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
1934 (*aarch64_mul3_elt<mode>): Likewise.
1935 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
1936 (*aarch64_mul3_elt_to_64v2df): Likewise.
1937 (*aarch64_mla_elt<mode>): Likewise.
1938 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
1939 (*aarch64_mls_elt<mode>): Likewise.
1940 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
1941 (*aarch64_fma4_elt<mode>): Likewise.
1942 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
1943 (*aarch64_fma4_elt_to_64v2df): Likewise.
1944 (*aarch64_fnma4_elt<mode>): Likewise.
1945 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
1946 (*aarch64_fnma4_elt_to_64v2df): Likewise.
1947 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1948 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
1949 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1950 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1951 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
1952 (aarch64_sqdmull_lane<mode>_internal): Likewise.
1953 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
1955 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
1957 * config/aarch64/aarch64-simd.md
1958 (aarch64_be_checked_get_lane<mode>): New define_expand.
1959 * config/aarch64/aarch64-simd-builtins.def
1960 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
1961 New builtin definition.
1962 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
1963 Use new safe be builtin.
1965 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
1967 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
1969 (aarch64_be_st1<mode>): Likewise.
1970 (aarch_ld1<VALL:mode>): Define_expand modified.
1971 (aarch_st1<VALL:mode>): Likewise.
1972 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
1973 (UNSPEC_ST1): Likewise.
1975 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
1977 * config/microblaze/microblaze.md: Add trap insn and attribute
1979 2014-01-23 Dodji Seketeli <dodji@redhat.com>
1981 PR preprocessor/58580
1982 * input.h (location_get_source_line): Take an additional line_size
1984 (void diagnostics_file_cache_fini): Declare new function.
1985 * input.c (struct fcache): New type.
1986 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
1987 New static constants.
1988 (diagnostic_file_cache_init, total_lines_num)
1989 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
1990 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
1991 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
1992 (get_next_line, read_next_line, goto_next_line, read_line_num):
1993 New static function definitions.
1994 (diagnostic_file_cache_fini): New function.
1995 (location_get_source_line): Take an additional output line_len
1996 parameter. Re-write using lookup_or_add_file_to_cache_tab and
1998 * diagnostic.c (diagnostic_finish): Call
1999 diagnostic_file_cache_fini.
2000 (adjust_line): Take an additional input parameter for the length
2001 of the line, rather than calculating it with strlen.
2002 (diagnostic_show_locus): Adjust the use of
2003 location_get_source_line and adjust_line with respect to their new
2004 signature. While displaying a line now, do not stop at the first
2005 null byte. Rather, display the zero byte as a space and keep
2006 going until we reach the size of the line.
2007 * Makefile.in: Add vec.o to OBJS-libcommon
2009 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
2010 Ilya Tocar <ilya.tocar@intel.com>
2012 * config/i386/avx512fintrin.h (_mm512_kmov): New.
2013 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
2014 (__builtin_ia32_kmov16): Ditto.
2015 * config/i386/i386.md (UNSPEC_KMOV): New.
2018 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
2020 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
2021 (_mm512_storeu_si512): Ditto.
2023 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
2026 * rtl.h (get_referenced_operands): Declare.
2027 * recog.c (get_referenced_operands): New function.
2028 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
2029 operands have been referenced when recording LO_SUM references.
2031 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
2033 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
2035 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
2037 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
2038 Enable for generic and recent AMD targets.
2040 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
2042 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
2043 ARG_SIZE note when adjustment was eliminated.
2045 2014-01-22 Jeff Law <law@redhat.com>
2047 PR tree-optimization/59597
2048 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
2049 in file. Accept new argument REGISTERING and use it to modify
2050 dump output appropriately.
2051 (register_jump_thread): Corresponding changes.
2052 (mark_threaded_blocks): Reinstate code to cancel unprofitable
2053 thread paths involving joiner blocks. Add code to dump cancelled
2054 jump threading paths.
2056 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
2058 PR rtl-optimization/59477
2059 * lra-constraints.c (inherit_in_ebb): Process call for living hard
2060 regs. Update reloads_num and potential_reload_hard_regs for all insns.
2062 2014-01-22 Tom Tromey <tromey@redhat.com>
2064 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
2066 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
2068 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
2070 PR rtl-optimization/59896
2071 * lra-constraints.c (process_alt_operands): Check unused note for
2072 matched operands of insn with no output reloads.
2074 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
2076 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
2077 (mips_move_from_gpr_cost): Likewise.
2079 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
2081 PR rtl-optimization/59858
2082 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
2083 ira_class_hard_regs_num.
2084 (process_alt_operands): Increase reject for dying matched operand.
2086 2014-01-21 Jakub Jelinek <jakub@redhat.com>
2089 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
2090 smaller than size, perform several stores or loads and stores
2091 at dst + count - size to store or copy all of size bytes, rather
2092 than just last modesize bytes.
2094 2014-01-20 DJ Delorie <dj@redhat.com>
2096 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
2097 that CLOBBERs are REGs before propogating their values.
2099 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
2102 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
2103 (cgraph_inline_failed_type): New function.
2104 * cgraph.h (DEFCIFCODE): Add type.
2105 (cgraph_inline_failed_type_t): New enum.
2106 (cgraph_inline_failed_type): New prototype.
2107 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
2108 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
2109 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
2110 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
2111 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
2112 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
2113 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
2114 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
2115 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
2116 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
2117 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
2118 OPTIMIZATION_MISMATCH.
2119 * tree-inline.c (expand_call_inline): Emit errors during
2120 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
2122 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
2125 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
2126 mode attribute in insn output.
2128 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
2130 * output.h (output_constant): Delete.
2131 * varasm.c (output_constant): Make private.
2133 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
2135 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
2137 2014-01-20 Jakub Jelinek <jakub@redhat.com>
2140 * tree.h (fold_builtin_strcat): New prototype.
2141 * builtins.c (fold_builtin_strcat): No longer static. Add len
2142 argument, if non-NULL, don't call c_strlen. Optimize
2143 directly into __builtin_memcpy instead of __builtin_strcpy.
2144 (fold_builtin_2): Adjust fold_builtin_strcat caller.
2145 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
2147 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
2149 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
2150 for SImode_address_operand operands, having only a REG argument.
2152 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
2154 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
2155 loader name using mbig-endian.
2156 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
2158 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
2160 * doc/invoke.texi (-march): Clarify documentation for AArch64.
2164 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
2166 * config/aarch64/aarch64-protos.h
2167 (aarch64_cannot_change_mode_class_ptr): Declare.
2168 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
2169 aarch64_cannot_change_mode_class_ptr): New.
2170 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
2171 backend hook aarch64_cannot_change_mode_class.
2173 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
2175 * common/config/aarch64/aarch64-common.c
2176 (aarch64_handle_option): Don't handle any option order logic here.
2177 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
2178 selected_cpu, warn on architecture version mismatch.
2179 (aarch64_override_options): Fix parsing order for option strings.
2181 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2182 Iain Sandoe <iain@codesourcery.com>
2185 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
2186 warning. Amend comment to reflect current functionality.
2188 2014-01-20 Richard Biener <rguenther@suse.de>
2191 * builtins.c (fold_builtin_strcat): Remove case better handled
2192 by tree-ssa-strlen.c.
2194 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
2196 * config/aarch64/aarch64.opt
2197 (mcpu, march, mtune): Make case-insensitive.
2199 2014-01-20 Jakub Jelinek <jakub@redhat.com>
2202 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
2203 if operands[1] is a REG or ZERO_EXTEND of a REG.
2205 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
2207 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
2209 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
2211 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
2212 long non-pic millicode calls.
2214 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2216 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
2218 2014-01-19 Kito Cheng <kito@0xlab.org>
2220 * builtins.c (expand_movstr): Check movstr expand done or fail.
2222 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
2223 H.J. Lu <hongjiu.lu@intel.com>
2226 * config/i386/i386.md (*lea<mode>): Zero-extend return register
2227 to DImode for zero-extended addresses.
2229 2014-01-19 Jakub Jelinek <jakub@redhat.com>
2231 PR rtl-optimization/57763
2232 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
2233 on the new indirect jump_insn and increment LABEL_NUSES (label).
2235 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
2239 * config.gcc (x86_archs): New variable.
2240 (x86_64_archs): Likewise.
2241 (x86_cpus): Likewise.
2242 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
2243 --with-arch/--with-cpu= options.
2244 Support --with-arch=/--with-cpu={nehalem,westmere,
2245 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
2247 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
2249 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
2250 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
2252 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
2254 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
2256 2014-01-18 Jakub Jelinek <jakub@redhat.com>
2259 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
2260 clear cpp_get_options (parse_in)->warn_unused_macros for
2261 ix86_target_macros_internal with cpp_define.
2263 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
2265 * jump.c (delete_related_insns): Keep (use (insn))s.
2266 * reorg.c (redundant_insn): Check for barriers too.
2268 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
2270 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
2272 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
2274 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
2275 call to $$dyncall when TARGET_LONG_CALLS is true.
2277 2014-01-17 Jeff Law <law@redhat.com>
2279 * ree.c (combine_set_extension): Temporarily disable test for
2280 changing number of hard registers.
2282 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
2285 * ipa-inline-analysis.c (inline_free_summary):
2286 Do not free summary of aliases.
2288 2014-01-17 Jakub Jelinek <jakub@redhat.com>
2291 * gimplify.c (gimplify_expr): Use create_tmp_var
2292 instead of create_tmp_var_raw. If cond doesn't have
2293 integral type, don't add the IFN_ANNOTATE builtin at all.
2295 2014-01-17 Martin Jambor <mjambor@suse.cz>
2298 * ipa-cp.c (prev_edge_clone): New variable.
2299 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
2300 Also resize prev_edge_clone vector.
2301 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
2302 (ipcp_edge_removal_hook): New function.
2303 (ipcp_driver): Register ipcp_edge_removal_hook.
2305 2014-01-17 Andrew Pinski <apinski@cavium.com>
2306 Steve Ellcey <sellcey@mips.com>
2309 * config/mips/mips.c (mips_print_operand): Check operand mode instead
2312 2014-01-17 Jeff Law <law@redhat.com>
2315 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
2316 so that pass_ccp runs first.
2318 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
2320 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
2321 (ix86_adjust_cost): Use !TARGET_XXX.
2322 (do_reorder_for_imul): Likewise.
2323 (swap_top_of_ready_list): Likewise.
2324 (ix86_sched_reorder): Likewise.
2326 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
2328 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
2329 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
2330 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
2331 (intel_memset): New. Duplicate slm_memset.
2332 (intel_cost): New. Duplicate slm_cost.
2333 (m_INTEL): New macro.
2334 (processor_target_table): Add "intel".
2335 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
2336 with PROCESSOR_INTEL for "intel".
2337 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
2338 PROCESSOR_SILVERMONT.
2339 (ix86_issue_rate): Likewise.
2340 (ix86_adjust_cost): Likewise.
2341 (ia32_multipass_dfa_lookahead): Likewise.
2342 (swap_top_of_ready_list): Likewise.
2343 (ix86_sched_reorder): Likewise.
2344 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
2345 instead of TARGET_OPT_AGU.
2346 * config/i386/i386.h (TARGET_INTEL): New.
2347 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
2348 (processor_type): Add PROCESSOR_INTEL.
2349 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
2350 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
2352 2014-01-17 Marek Polacek <polacek@redhat.com>
2355 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
2358 2014-01-17 Richard Biener <rguenther@suse.de>
2360 PR tree-optimization/46590
2361 * opts.c (default_options_table): Add entries for
2362 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
2363 all enabled at -O1 but not for -Og.
2364 * common.opt (fbranch-count-reg): Remove Init(1).
2365 (fmove-loop-invariants): Likewise.
2366 (ftree-pta): Likewise.
2368 2014-01-17 Jakub Jelinek <jakub@redhat.com>
2370 * config/i386/i386.c (ix86_data_alignment): For compatibility with
2371 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
2372 decls to at least the GCC 4.8 used alignments.
2375 * tree-nested.c (convert_nonlocal_reference_stmt,
2376 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
2377 of GIMPLE_BIND stmts, adjust associated decls.
2379 2014-01-17 Richard Biener <rguenther@suse.de>
2381 PR tree-optimization/46590
2382 * vec.h (vec<>::bseach): New member function implementing
2383 binary search according to C89 bsearch.
2384 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
2385 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
2386 bitmap pointer again. Make accesses_in_loop a flat array.
2387 (mem_ref_obstack): New global.
2388 (outermost_indep_loop): Adjust for mem_ref->stored changes.
2389 (mark_ref_stored): Likewise.
2390 (ref_indep_loop_p_2): Likewise.
2391 (set_ref_stored_in_loop): New helper function.
2392 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
2393 (memref_free): Adjust.
2394 (record_mem_ref_loc): Simplify.
2395 (gather_mem_refs_stmt): Adjust.
2396 (sort_locs_in_loop_postorder_cmp): New function.
2397 (analyze_memory_references): Sort accesses_in_loop after
2398 loop postorder number.
2399 (find_ref_loc_in_loop_cmp): New function.
2400 (for_all_locs_in_loop): Find relevant cluster of locs in
2401 accesses_in_loop and iterate without recursion.
2402 (execute_sm): Avoid uninit warning.
2403 (struct ref_always_accessed): Simplify.
2404 (ref_always_accessed::operator ()): Likewise.
2405 (ref_always_accessed_p): Likewise.
2406 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
2407 loop postorder numbers here.
2408 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
2411 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
2414 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
2415 on decls for which assemble_alias has been called.
2417 2014-01-17 Nick Clifton <nickc@redhat.com>
2419 * config/msp430/msp430.opt: (mcpu): New option.
2420 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
2421 (msp430_option_override): Parse target_cpu. If the MCU name
2422 matches a generic string, clear target_mcu.
2423 (msp430_attr): Allow numeric interrupt values up to 63.
2424 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
2425 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
2427 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
2429 * config/msp430/msp430.md (popm): Use %J rather than %I.
2430 (addsi3): Use msp430_nonimmediate_operand for operand 2.
2431 (addhi_cy_i): Use immediate_operand for operand 2.
2432 * doc/invoke.texi: Document -mcpu option.
2434 2014-01-17 Richard Biener <rguenther@suse.de>
2436 PR rtl-optimization/38518
2437 * df.h (df_analyze_loop): Declare.
2438 * df-core.c: Include cfgloop.h.
2439 (df_analyze_1): Split out main part of df_analyze.
2440 (df_analyze): Adjust.
2441 (loop_inverted_post_order_compute): New function.
2442 (loop_post_order_compute): Likewise.
2443 (df_analyze_loop): New function avoiding whole-function
2445 * loop-invariant.c (find_defs): Use df_analyze_loop.
2446 (find_invariants): Adjust.
2447 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
2449 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
2451 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
2452 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
2454 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
2456 * ipa-ref.c (ipa_remove_stmt_references): Fix references
2457 traversal when removing references.
2459 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
2462 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
2464 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
2467 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
2468 pushing a reload for an autoinc when we had previously reloaded an
2469 inner part of the address.
2471 2014-01-16 Jakub Jelinek <jakub@redhat.com>
2473 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
2475 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
2476 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
2477 when not giving up or versioning for alias only because of
2479 (vect_analyze_data_ref_dependences): Set to true.
2480 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
2482 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
2483 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
2487 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
2490 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
2491 operand 0 predicate for gathers, use a new pseudo as subtarget.
2493 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
2496 * lra-constraints.c (process_alt_operands): Add printing debug info.
2497 Check absence of input/output reloads for matched operands too.
2499 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
2501 PR rtl-optimization/59835
2502 * ira.c (ira_init_register_move_cost): Increase cost for
2505 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
2507 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
2509 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
2512 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
2513 non-register objects. Use gen_(high/low)part more consistently.
2516 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
2519 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
2520 endian support, remove tests for WORDS_BIG_ENDIAN.
2521 (p8_mfvsrd_3_<mode>): Likewise.
2522 (reload_gpr_from_vsx<mode>): Likewise.
2523 (reload_gpr_from_vsxsf): Likewise.
2524 (p8_mfvsrd_4_disf): Likewise.
2526 2014-01-16 Richard Biener <rguenther@suse.de>
2528 PR rtl-optimization/46590
2529 * lcm.c (compute_antinout_edge): Use postorder iteration.
2530 (compute_laterin): Use inverted postorder iteration.
2532 2014-01-16 Nick Clifton <nickc@redhat.com>
2535 * varasm.c (output_constant): Return the number of bytes actually
2537 (output_constructor_array_range): Update the field size with the
2538 number of bytes emitted by output_constant.
2539 (output_constructor_regular_field): Likewise. Also do not
2540 complain if the total number of bytes emitted is now greater
2541 than the expected fieldpos.
2542 * output.h (output_constant): Update prototype and descriptive comment.
2544 2014-01-16 Marek Polacek <polacek@redhat.com>
2547 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
2548 it is error_mark_node.
2550 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
2552 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
2553 VALID_AVX256_REG_OR_OI_MODE.
2555 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
2557 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
2558 current procedure should be profiled.
2560 2014-01-15 Andrew Pinski <apinski@cavium.com>
2562 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
2563 of moving from/to the STACK_REG register class.
2565 2014-01-15 Richard Henderson <rth@redhat.com>
2568 * reginfo.c (global_regs_decl): Globalize.
2569 * rtl.h (global_regs_decl): Declare.
2570 * ira.c (do_reload): Diagnose frame_pointer_needed and it
2571 reserved via global_regs.
2573 2014-01-15 Teresa Johnson <tejohnson@google.com>
2575 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
2577 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2579 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
2580 and vmulosh rather than call gen_vec_widen_smult_*.
2581 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
2582 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
2583 (vec_widen_smult_even_v16qi): Likewise.
2584 (vec_widen_umult_even_v8hi): Likewise.
2585 (vec_widen_smult_even_v8hi): Likewise.
2586 (vec_widen_umult_odd_v16qi): Likewise.
2587 (vec_widen_smult_odd_v16qi): Likewise.
2588 (vec_widen_umult_odd_v8hi): Likewise.
2589 (vec_widen_smult_odd_v8hi): Likewise.
2590 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
2591 vmuloub rather than call gen_vec_widen_umult_*.
2592 (vec_widen_umult_lo_v16qi): Likewise.
2593 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
2594 vmulosb rather than call gen_vec_widen_smult_*.
2595 (vec_widen_smult_lo_v16qi): Likewise.
2596 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
2597 rather than call gen_vec_widen_umult_*.
2598 (vec_widen_umult_lo_v8hi): Likewise.
2599 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
2600 rather than call gen_vec_widen_smult_*.
2601 (vec_widen_smult_lo_v8hi): Likewise.
2603 2014-01-15 Jeff Law <law@redhat.com>
2605 PR tree-optimization/59747
2606 * ree.c (find_and_remove_re): Properly handle case where a second
2607 eliminated extension requires widening a copy created for elimination
2608 of a prior extension.
2609 (combine_set_extension): Ensure that the number of hard regs needed
2610 for a destination register does not change when we widen it.
2612 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
2614 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
2615 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
2616 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
2617 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
2618 (avr-*-rtems*): Likewise.
2619 (bfin*-rtems*): Likewise.
2620 (moxie-*-rtems*): Likewise.
2621 (h8300-*-rtems*): Likewise.
2622 (i[34567]86-*-rtems*): Likewise.
2623 (lm32-*-rtems*): Likewise.
2624 (m32r-*-rtems*): Likewise.
2625 (m68k-*-rtems*): Likewise.
2626 (microblaze*-*-rtems*): Likewise.
2627 (mips*-*-rtems*): Likewise.
2628 (powerpc-*-rtems*): Likewise.
2629 (sh-*-rtems*): Likewise.
2630 (sparc-*-rtems*): Likewise.
2631 (sparc64-*-rtems*): Likewise.
2632 (v850-*-rtems*): Likewise.
2633 (m32c-*-rtems*): Likewise.
2635 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
2637 PR rtl-optimization/59511
2638 * ira.c (ira_init_register_move_cost): Use memory costs for some
2639 cases of register move cost calculations.
2640 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
2641 instead of BB frequency.
2642 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
2643 * lra-assigns.c (find_hard_regno_for): Ditto.
2645 2014-01-15 Richard Biener <rguenther@suse.de>
2647 PR tree-optimization/59822
2648 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
2649 (vectorizable_load): Use it to hoist defs of uses of invariant
2650 loads out of the loop.
2652 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
2653 Kugan Vivekanandarajah <kuganv@linaro.org>
2656 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
2659 2014-01-15 Richard Biener <rguenther@suse.de>
2661 PR rtl-optimization/59802
2662 * lcm.c (compute_available): Use inverted postorder to seed
2663 the initial worklist.
2665 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2668 * config/s390/s390.c (s390_preferred_reload_class): Don't return
2669 ADDR_REGS for invalid symrefs in non-PIC code.
2671 2014-01-15 Jakub Jelinek <jakub@redhat.com>
2674 * builtins.c (determine_block_size): Initialize *probable_max_size
2675 even if len_rtx is CONST_INT.
2677 2014-01-14 Andrew Pinski <apinski@cavium.com>
2679 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
2680 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
2681 (cortexa53_tunings): Likewise.
2682 (aarch64_sched_issue_rate): New function.
2683 (TARGET_SCHED_ISSUE_RATE): Define.
2685 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
2687 * ira-costs.c (find_costs_and_classes): Add missed
2688 ira_init_register_move_cost_if_necessary.
2690 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
2693 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
2695 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
2698 * config/i386/i386.c (type_natural_mode): Add a bool parameter
2699 to indicate if type is used for function return value. Warn ABI
2700 change if the vector mode isn't available for function return value.
2701 (ix86_function_arg_advance): Pass false to type_natural_mode.
2702 (ix86_function_arg): Likewise.
2703 (ix86_gimplify_va_arg): Likewise.
2704 (function_arg_32): Don't warn ABI change.
2705 (ix86_function_value): Pass true to type_natural_mode.
2706 (ix86_return_in_memory): Likewise.
2707 (ix86_struct_value_rtx): Removed.
2708 (TARGET_STRUCT_VALUE_RTX): Likewise.
2710 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
2712 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
2713 converting a conditional jump into a conditional return.
2715 2014-01-14 Richard Biener <rguenther@suse.de>
2717 PR tree-optimization/58921
2718 PR tree-optimization/59006
2719 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
2720 hoisting invariant stmts.
2721 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
2722 invariant loads on the preheader edge if possible.
2724 2014-01-14 Joey Ye <joey.ye@arm.com>
2726 * doc/plugin.texi (Building GCC plugins): Update to C++.
2728 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
2730 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
2731 (_mm_rcp28_round_ss): Ditto.
2732 (_mm_rsqrt28_round_sd): Ditto.
2733 (_mm_rsqrt28_round_ss): Ditto.
2734 (_mm_rcp28_sd): Ditto.
2735 (_mm_rcp28_ss): Ditto.
2736 (_mm_rsqrt28_sd): Ditto.
2737 (_mm_rsqrt28_ss): Ditto.
2738 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
2739 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
2740 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
2741 (IX86_BUILTIN_RCP28SD): Ditto.
2742 (IX86_BUILTIN_RCP28SS): Ditto.
2743 (IX86_BUILTIN_RSQRT28SD): Ditto.
2744 (IX86_BUILTIN_RSQRT28SS): Ditto.
2745 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
2746 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
2747 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
2748 (ix86_expand_special_args_builtin): Expand new FTYPE.
2749 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
2750 (srcp14<mode>): Make insn unary.
2751 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
2752 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
2753 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
2754 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
2755 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
2756 Fix rounding: make it SAE only.
2757 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
2759 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
2761 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
2762 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
2763 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
2764 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
2765 (round_saeonly_mask_scalar_operand4): Ditto.
2766 (round_saeonly_mask_scalar_op3): Ditto.
2767 (round_saeonly_mask_scalar_op4): Ditto.
2769 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2771 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
2772 Implement -maltivec=be for vec_insert and vec_extract.
2774 2014-01-10 DJ Delorie <dj@redhat.com>
2776 * config/msp430/msp430.md (call_internal): Don't allow memory
2777 references with SP as the base register.
2778 (call_value_internal): Likewise.
2779 * config/msp430/constraints.md (Yc): New. For memory references
2780 that don't use SP as a base register.
2782 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
2783 "an integer without a # prefix"
2784 * config/msp430/msp430.md (epilogue_helper): Use it.
2786 2014-01-13 Jakub Jelinek <jakub@redhat.com>
2789 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
2790 AVX512F gather builtins.
2791 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
2792 on gather decls with INTEGER_TYPE masktype.
2793 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
2794 directly into the builtin rather than hoisting it before loop.
2796 PR tree-optimization/59387
2797 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
2798 (scev_const_prop): If folded_casts and type has undefined overflow,
2799 use force_gimple_operand instead of force_gimple_operand_gsi and
2800 for each added stmt if it is assign with
2801 arith_code_with_undefined_signed_overflow, call
2802 rewrite_to_defined_overflow.
2803 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
2804 gimple-fold.h instead.
2805 (arith_code_with_undefined_signed_overflow,
2806 rewrite_to_defined_overflow): Moved to ...
2807 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
2808 rewrite_to_defined_overflow): ... here. No longer static.
2809 Include gimplify-me.h.
2810 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
2811 rewrite_to_defined_overflow): New prototypes.
2813 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2815 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
2817 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
2819 * builtins.c (get_object_alignment_2): Minor tweak.
2820 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
2822 2014-01-13 Christian Bruel <christian.bruel@st.com>
2824 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
2825 optimized non constant lengths.
2827 2014-01-13 Jakub Jelinek <jakub@redhat.com>
2830 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
2831 load as __atomic_load_N if possible.
2833 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
2835 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
2837 (rs6000_expand_builtin): Adjust call.
2839 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
2842 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
2843 * config/rs6000/rs6000.c: Include target-globals.h.
2844 (rs6000_set_current_function): Instead of doing target_reinit
2845 unconditionally, use save_target_globals_default_opts and
2846 restore_target_globals.
2848 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
2850 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
2851 (rs6000_expand_builtin): Handle mffs and mtfsf.
2852 (rs6000_init_builtins): Define mffs and mtfsf.
2853 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
2854 (rs6000_mffs): New pattern.
2855 (rs6000_mtfsf): New pattern.
2857 2014-01-11 Bin Cheng <bin.cheng@arm.com>
2859 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
2860 Start narrowing with START. Apply candidate-use pair
2861 and check overall cost in narrowing.
2862 (iv_ca_prune): Pass new argument.
2864 2014-01-10 Jeff Law <law@redhat.com>
2867 * ree.c (combine_reaching_defs): Ensure the defining statement
2868 occurs before the extension when optimizing extensions with
2869 different source and destination hard registers.
2871 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
2874 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
2875 vtables into the type inheritance graph.
2877 2014-01-10 Jakub Jelinek <jakub@redhat.com>
2879 PR rtl-optimization/59754
2880 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
2881 modes in the REGNO != REGNO case.
2883 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2885 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
2887 2014-01-10 Jakub Jelinek <jakub@redhat.com>
2889 PR tree-optimization/59745
2890 * tree-predcom.c (tree_predictive_commoning_loop): Call
2891 free_affine_expand_cache if giving up because components is NULL.
2893 * target-globals.c (save_target_globals): Allocate < 4KB structs using
2894 GC in payload of target_globals struct instead of allocating them on
2895 the heap and the larger structs separately using GC.
2896 * target-globals.h (struct target_globals): Make regs, hard_regs,
2897 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
2898 of GTY((skip)) and change type to void *.
2899 (reset_target_globals): Cast loads from those fields to corresponding
2902 2014-01-10 Steve Ellcey <sellcey@mips.com>
2905 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
2906 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
2907 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
2909 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
2912 * aarch64-modes.def (CC_Zmode): New flags mode.
2913 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
2914 represents an equality.
2915 (aarch64_get_condition_code): Handle CC_Zmode.
2916 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
2918 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2920 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
2921 extraction in good case.
2923 2014-01-10 Richard Biener <rguenther@suse.de>
2925 PR tree-optimization/59374
2926 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
2927 checking after SLP discovery. Mark stmts not participating
2928 in any SLP instance properly.
2930 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2932 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
2933 when handling a SET rtx.
2935 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2937 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
2938 (cortex-a57): Likewise.
2939 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
2941 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2943 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
2944 non-iwmmxt builtins.
2946 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
2950 * ipa-devirt.c record_target_from_binfo): Take as argument
2951 stack of binfos and lookup matching one for virtual inheritance.
2952 (possible_polymorphic_call_targets_1): Update.
2954 2014-01-10 Huacai Chen <chenhc@lemote.com>
2956 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
2957 kernel strings for Loongson-2E/2F/3A.
2959 2014-01-10 Jakub Jelinek <jakub@redhat.com>
2962 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
2963 is_gimple_call before calling gimple_call_internal_p.
2965 2014-01-09 Steve Ellcey <sellcey@mips.com>
2967 * Makefile.in (TREE_FLOW_H): Remove.
2968 (TREE_SSA_H): Add file names from tree-flow.h.
2969 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
2970 * tree.h: Remove tree-flow.h reference.
2971 * hash-table.h: Remove tree-flow.h reference.
2972 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
2973 reference with tree-ssa-loop.h.
2975 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2977 * doc/invoke.texi: Add -maltivec={be,le} options, and document
2978 default element-order behavior for -maltivec.
2979 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
2980 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
2981 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
2982 when targeting big endian, at least for now.
2983 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
2985 2014-01-09 Jakub Jelinek <jakub@redhat.com>
2988 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
2989 var satisfies use_register_for_decl, just take into account type
2990 alignment, rather than decl alignment.
2992 PR tree-optimization/59622
2993 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
2994 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
2995 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
2996 Don't devirtualize for inplace at all. For targets.length () == 1,
2997 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
2999 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
3001 * config/i386/i386.md (cpu): Remove the unused btver1.
3003 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
3005 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
3007 2014-01-09 Jakub Jelinek <jakub@redhat.com>
3010 * tree-core.h (struct target_globals): New forward declaration.
3011 (struct tree_target_option): Add globals field.
3012 * tree.h (TREE_TARGET_GLOBALS): Define.
3013 (prepare_target_option_nodes_for_pch): New prototype.
3014 * target-globals.h (struct target_globals): Define even if
3016 * tree.c (prepare_target_option_node_for_pch,
3017 prepare_target_option_nodes_for_pch): New functions.
3018 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
3019 * config/i386/i386.c: Include target-globals.h.
3020 (ix86_set_current_function): Instead of doing target_reinit
3021 unconditionally, use save_target_globals_default_opts and
3022 restore_target_globals.
3024 2014-01-09 Richard Biener <rguenther@suse.de>
3026 PR tree-optimization/59715
3027 * tree-cfg.h (split_critical_edges): Declare.
3028 * tree-cfg.c (split_critical_edges): Export.
3029 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
3031 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
3033 * cfgexpand.c (expand_stack_vars): Optionally disable
3034 asan stack protection.
3035 (expand_used_vars): Likewise.
3036 (partition_stack_vars): Likewise.
3037 * asan.c (asan_emit_stack_protection): Optionally disable
3038 after return stack usage.
3039 (instrument_derefs): Optionally disable memory access instrumentation.
3040 (instrument_builtin_call): Likewise.
3041 (instrument_strlen_call): Likewise.
3042 (asan_protect_global): Optionally disable global variables protection.
3043 * doc/invoke.texi: Added doc for new options.
3044 * params.def: Added new options.
3045 * params.h: Likewise.
3047 2014-01-09 Jakub Jelinek <jakub@redhat.com>
3049 PR rtl-optimization/59724
3050 * ifcvt.c (cond_exec_process_if_block): Don't call
3051 flow_find_head_matching_sequence with 0 longest_match.
3052 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
3053 non-active insns if !stop_after.
3054 (try_head_merge_bb): Revert 2014-01-07 changes.
3056 2014-01-08 Jeff Law <law@redhat.com>
3058 * ree.c (get_sub_rtx): New function, extracted from...
3059 (merge_def_and_ext): Here.
3060 (combine_reaching_defs): Use get_sub_rtx.
3062 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
3064 * cgraph.h (varpool_variable_node): Do not choke on null node.
3066 2014-01-08 Catherine Moore <clm@codesourcery.com>
3068 * config/mips/mips.md (simple_return): Attempt to use JRC
3070 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
3072 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
3074 PR rtl-optimization/59137
3075 * reorg.c (steal_delay_list_from_target): Call update_block for
3077 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
3079 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3081 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
3082 two duplicate entries.
3084 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
3087 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
3089 * config/mips/mips.c (mips_truncated_op_cost): New function.
3090 (mips_rtx_costs): Adjust test for BADDU.
3091 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
3093 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
3095 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
3096 (*baddu_si): ...this new pattern.
3098 2014-01-08 Jakub Jelinek <jakub@redhat.com>
3101 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
3103 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
3106 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
3108 (expand_expr, expand_normal): Adjust.
3109 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
3110 inner_reference_p. Use inner_reference_p to expand inner references.
3111 (store_expr): Adjust.
3112 * cfgexpand.c (expand_call_stmt): Adjust.
3114 2014-01-08 Rong Xu <xur@google.com>
3116 * gcov-io.c (gcov_var): Move from gcov-io.h.
3117 (gcov_position): Ditto.
3118 (gcov_is_error): Ditto.
3119 (gcov_rewrite): Ditto.
3120 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
3121 only part to libgcc/libgcov.h.
3123 2014-01-08 Marek Polacek <polacek@redhat.com>
3126 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
3128 2014-01-08 Marek Polacek <polacek@redhat.com>
3131 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
3133 2014-01-08 Jakub Jelinek <jakub@redhat.com>
3135 PR rtl-optimization/59649
3136 * stor-layout.c (get_mode_bounds): For BImode return
3137 0 and STORE_FLAG_VALUE.
3139 2014-01-08 Richard Biener <rguenther@suse.de>
3142 * gimple.h (is_gimple_builtin_call): Remove.
3143 (gimple_builtin_call_types_compatible_p): New.
3144 (gimple_call_builtin_p): New overload.
3145 * gimple.c (is_gimple_builtin_call): Remove.
3146 (validate_call): Rename to ...
3147 (gimple_builtin_call_types_compatible_p): ... this and export. Also
3149 (validate_type): New static function.
3150 (gimple_call_builtin_p): New overload and adjust.
3151 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
3152 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
3153 (gimple_fold_stmt_to_constant_1): Likewise.
3154 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
3156 2014-01-08 Richard Biener <rguenther@suse.de>
3159 * gimplify.c (gimplify_expr): Gimplify register-register type
3160 VIEW_CONVERT_EXPRs to separate stmts.
3162 2014-01-07 Jeff Law <law@redhat.com>
3165 * ree.c (combine_set_extension): Handle case where source
3166 and destination registers in an extension insn are different.
3167 (combine_reaching_defs): Allow source and destination registers
3168 in extension to be different under limited circumstances.
3169 (add_removable_extension): Remove restriction that the
3170 source and destination registers in the extension are the same.
3171 (find_and_remove_re): Emit a copy from the extension's
3172 destination to its source after the defining insn if
3173 the source and destination registers are different.
3176 * ifcvt.c (merge_if_block): If we are merging a block with more than
3177 one successor with a block with no successors, remove any BARRIER
3178 after the second block.
3180 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
3182 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
3184 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
3187 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
3188 for 14-bit register offsets when INT14_OK_STRICT is false.
3190 2014-01-07 Roland Stigge <stigge@antcom.de>
3191 Michael Meissner <meissner@linux.vnet.ibm.com>
3194 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
3195 Only check TFmode for SPE constants. Don't check TImode or TDmode.
3197 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
3199 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
3202 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
3204 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
3205 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
3206 rtx is const0_rtx or not.
3208 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
3211 * target-globals.c (save_target_globals): Remove this_fn_optab
3213 * toplev.c: Include optabs.h.
3214 (target_reinit): Temporarily restore the global options if another
3215 set of options are in force.
3217 2014-01-07 Jakub Jelinek <jakub@redhat.com>
3219 PR rtl-optimization/58668
3220 * cfgcleanup.c (flow_find_cross_jump): Don't count
3221 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
3222 to determine what is counted.
3223 (flow_find_head_matching_sequence): Use active_insn_p to determine
3225 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
3227 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
3228 determine what is counted.
3230 PR tree-optimization/59643
3231 * tree-predcom.c (split_data_refs_to_components): If one dr is
3232 read and one write, determine_offset fails and the write isn't
3233 in the bad component, just put the read into the bad component.
3235 2014-01-07 Mike Stump <mikestump@comcast.net>
3236 Jakub Jelinek <jakub@redhat.com>
3239 * tree-core.h (struct tree_optimization_option): Change optabs
3240 type from unsigned char * to void *.
3241 * optabs.c (init_tree_optimization_optabs): Adjust
3242 TREE_OPTIMIZATION_OPTABS initialization.
3244 2014-01-06 Jakub Jelinek <jakub@redhat.com>
3247 * config/i386/i386.h (struct machine_function): Add
3248 no_drap_save_restore field.
3249 * config/i386/i386.c (ix86_save_reg): Use
3250 !cfun->machine->no_drap_save_restore instead of
3251 crtl->stack_realign_needed.
3252 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
3253 this function clears frame_pointer_needed. Set
3254 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
3255 and DRAP reg is needed.
3257 2014-01-06 Marek Polacek <polacek@redhat.com>
3260 * doc/implement-c.texi: Mention that other integer types are
3261 permitted as bit-field types in strictly conforming mode.
3263 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
3265 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
3268 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
3270 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
3272 2014-01-06 Martin Jambor <mjambor@suse.cz>
3275 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
3277 * ipa-prop.c (ipa_print_node_params): Fix indentation.
3279 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
3283 * var-tracking.c (add_stores): Preserve the value of the source even if
3284 we don't record the store.
3286 2014-01-06 Terry Guo <terry.guo@arm.com>
3288 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
3290 2014-01-05 Iain Sandoe <iain@codesourcery.com>
3293 * config/darwin.c (darwin_function_section): Adjust return values to
3294 correspond to optimisation changes made in r206070.
3296 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
3298 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
3299 from prefetch_block tune setting.
3300 (nocona_cost): Correct size of prefetch block to 64.
3302 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
3304 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
3305 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
3306 used to save the static chain register in the computation of the offset
3307 from which the FP registers need to be restored.
3309 2014-01-04 Jakub Jelinek <jakub@redhat.com>
3311 PR tree-optimization/59519
3312 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
3313 ICE if get_current_def (current_new_name) is already non-NULL, as long
3314 as it is a phi result of some other phi in *new_exit_bb that has
3317 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
3318 or vmovdqu* for misaligned_operand.
3319 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
3320 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
3321 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
3322 aligned_mem for AVX512F masked aligned load and store builtins and for
3325 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
3327 PR tree-optimization/59651
3328 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
3329 Address range for negative step should be added by TYPE_SIZE_UNIT.
3331 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
3333 * config/m68k/m68k.c (handle_move_double): Handle pushes with
3334 overlapping registers also for registers other than the stack pointer.
3336 2014-01-03 Marek Polacek <polacek@redhat.com>
3339 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
3342 2014-01-03 Jakub Jelinek <jakub@redhat.com>
3345 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
3348 * config/i386/i386.md (MODE_SIZE): New mode attribute.
3349 (push splitter): Use <P:MODE_SIZE> instead of
3350 GET_MODE_SIZE (<P:MODE>mode).
3351 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
3352 (mov -1, reg peephole2): Likewise.
3353 * config/i386/sse.md (*mov<mode>_internal,
3354 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
3355 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
3356 *<code><mode>3, *andnot<mode>3<mask_name>,
3357 <mask_codefor><code><mode>3<mask_name>): Likewise.
3358 * config/i386/subst.md (mask_mode512bit_condition,
3359 sd_mask_mode512bit_condition): Likewise.
3361 2014-01-02 Xinliang David Li <davidxl@google.com>
3363 PR tree-optimization/59303
3364 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
3365 (dump_predicates): Better output format.
3366 (pred_equal_p): New function.
3367 (is_neq_relop_p): Ditto.
3368 (is_neq_zero_form_p): Ditto.
3369 (pred_expr_equal_p): Ditto.
3370 (pred_neg_p): Ditto.
3371 (simplify_pred): Ditto.
3372 (simplify_preds_2): Ditto.
3373 (simplify_preds_3): Ditto.
3374 (simplify_preds_4): Ditto.
3375 (simplify_preds): Ditto.
3377 (push_to_worklist): Ditto.
3378 (get_pred_info_from_cmp): Ditto.
3379 (is_degenerated_phi): Ditto.
3380 (normalize_one_pred_1): Ditto.
3381 (normalize_one_pred): Ditto.
3382 (normalize_one_pred_chain): Ditto.
3383 (normalize_preds): Ditto.
3384 (normalize_cond_1): Remove function.
3385 (normalize_cond): Ditto.
3386 (is_gcond_subset_of): Ditto.
3387 (is_subset_of_any): Ditto.
3388 (is_or_set_subset_of): Ditto.
3389 (is_and_set_subset_of): Ditto.
3390 (is_norm_cond_subset_of): Ditto.
3391 (pred_chain_length_cmp): Ditto.
3392 (convert_control_dep_chain_into_preds): Type change.
3393 (find_predicates): Ditto.
3394 (find_def_preds): Ditto.
3395 (destroy_predicates_vecs): Ditto.
3396 (find_matching_predicates_in_rest_chains): Ditto.
3397 (use_pred_not_overlap_with_undef_path_pred): Ditto.
3398 (is_pred_expr_subset): Ditto.
3399 (is_pred_chain_subset_of): Ditto.
3400 (is_included_in): Ditto.
3401 (is_superset_of): Ditto.
3403 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
3405 Update copyright years.
3407 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
3409 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
3410 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
3411 config/arc/arc.md, config/arc/arc.opt,
3412 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
3413 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
3414 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
3415 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
3416 config/linux-protos.h, config/linux.c, config/winnt-c.c,
3417 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
3418 vtable-verify.c, vtable-verify.h: Use the standard form for the
3421 2014-01-02 Tobias Burnus <burnus@net-b.de>
3423 * gcc.c (process_command): Update copyright notice dates.
3424 * gcov-dump.c: Ditto.
3426 * doc/cpp.texi: Bump @copying's copyright year.
3427 * doc/cppinternals.texi: Ditto.
3428 * doc/gcc.texi: Ditto.
3429 * doc/gccint.texi: Ditto.
3430 * doc/gcov.texi: Ditto.
3431 * doc/install.texi: Ditto.
3432 * doc/invoke.texi: Ditto.
3434 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3436 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
3438 2014-01-01 Jakub Jelinek <jakub@redhat.com>
3440 * config/i386/sse.md (*mov<mode>_internal): Guard
3441 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
3443 PR rtl-optimization/59647
3444 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
3445 new_rtx into UNSIGNED_FLOAT rtxes.
3447 Copyright (C) 2014 Free Software Foundation, Inc.
3449 Copying and distribution of this file, with or without modification,
3450 are permitted in any medium without royalty provided the copyright
3451 notice and this notice are preserved.