1 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3 * config/tilegx/tilegx.c: Include "tree.h".
5 2013-10-23 Jakub Jelinek <jakub@redhat.com>
7 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
8 the info, not after it.
9 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
10 dump_ssaname_info call.
11 (pp_gimple_stmt_1): Adjust caller.
12 (dump_phi_nodes): Likewise. Don't print "# " here.
14 2013-10-22 Jan Hubicka <jh@suse.cz>
16 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New tuning flag.
17 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
18 * i386.c (expand_small_movmem_or_setmem): New function.
19 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New function
20 (alg_usable_p): Add support for value ranges; cleanup.
21 (ix86_expand_set_or_movmem): Add support for misaligned moves.
23 2013-10-22 Sterling Augustine <saugustine@google.com>
25 * doc/invoke.texi: Document -ggnu-pubnames.
26 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
28 * dwarf2out.c: Include gdb/gdb-index.h.
29 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
30 debug_generate_pub_sections.
31 (is_java, output_pubtables, output_pubname): New functions.
32 (include_pubname_in_output): Handle debug_generate_pub_sections at
34 (size_of_pubnames): Use new local space_for_flags based on
35 debug_generate_pub_sections.
36 (output_pubnames): Unify pubnames and pubtypes output logic.
37 Genericize comments. Call output_pubname.
38 (dwarf2out_finish): Move logic to output_pubnames and call it.
40 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
43 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
44 Remove CCCmode handling.
45 <case LTU>: Return 'c' suffix for CCCmode.
46 <case GEU>: Return 'nc' suffix for CCCmode.
47 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
48 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
49 (*sub<mode>3_cc_overflow): Ditto.
50 (*subsi3_zext_cc_overflow): Ditto.
52 2013-10-22 Steve Ellcey <sellcey@mips.com>
54 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
55 (AND (NOT OP1) (NOT OP2)).
57 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
59 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
60 meaning of merge-high and merge-low masks for little endian; avoid
61 use of vector-pack masks for little endian for mismatched modes.
63 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
65 * config/tilepro/tilepro.c: Include "tree.h".
67 2013-10-22 Andreas Schwab <schwab@suse.de>
69 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
72 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
74 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
75 [Ed Smith-Rowland]): New entries.
76 ([Stephen M. Webb]): Update.
78 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
80 * tree-ssa-ter.h: Remove duplicate copy of file contents.
82 2013-10-21 Marek Polacek <polacek@redhat.com>
85 * fold-const.c (fold_range_test): Return 0 if the type is not
88 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
90 * system.h: Move hwint.h include further down.
91 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
93 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
95 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
97 Fix volatile issues in optimize_bit_field_compare.
98 * fold-const.c (optimize_bit_field_compare): Bail out if
99 lvolatilep or rvolatilep.
101 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
103 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
104 and get_inner_reference returning different pmode for non-volatile
105 bit-field members dependent on flag_strict_volatile_bitfields.
106 * stor-layout.c (layout_decl): Remove special handling of
107 flag_strict_volatile_bitfields.
108 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
109 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
111 2013-10-21 Paulo Matos <pmatos@broadcom.com>
113 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
116 2013-10-21 Jeff Law <law@redhat.com>
118 * tree-ssa-threadedge.c (thread_through_normal_block): New
119 argument VISITED. Remove VISISTED as a local variable. When we
120 have a threadable jump, verify the destination of the jump has not
122 (thread_across_edge): Allocate VISITED bitmap once at function
123 scope and use it throughout. Make sure to set appropriate bits in
124 VISITED for E (start of jump thread path).
125 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
126 through a joiner if any edge on the path has a recorded jump thread.
128 2013-10-21 Ian Lance Taylor <iant@google.com>
130 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
131 don't imply that attributes can solve all problems.
132 (Directory Options): Fix typo.
134 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
136 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
137 extend and extend_add.
139 2013-10-21 Richard Biener <rguenther@suse.de>
141 PR tree-optimization/58794
142 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
143 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
145 2013-10-21 Richard Biener <rguenther@suse.de>
148 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
149 to (T) X for sign-changing conversions (or no conversion).
151 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
153 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
155 2013-10-20 Jan Hubicka <jh@suse.cz>
157 * config/i386/i386-tune.def: Add comment; organize into categories
159 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
161 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
162 argument. Update function comment.
163 (expand_set_or_movmem_via_rep): New function combining
164 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
165 (expand_movmem_via_rep_mov): Remove.
166 expand_setmem_via_rep_stos): Remove.
167 (expand_movmem_epilogue): Update calls correspondingly.
168 (expand_setmem_epilogue_via_loop): Likewise.
170 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
171 (expand_set_or_movmem_prologue): New function combining
172 expand_movmem_prologue and expand_setmem_prologue.
173 (expand_movmem_prologue): Remove.
174 (expand_setmem_prologue): Remove.
175 (expand_set_or_movmem_constant_prologue): New function combining
176 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
177 (expand_constant_movmem_prologue): Remove.
178 (expand_constant_setmem_prologue): Remove.
179 (promote_duplicated_reg): Allow vector-const0 value.
180 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
181 and ix86_expand_setmem.
182 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
183 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
185 2013-10-21 Diego Novillo <dnovillo@google.com>
187 * asan.c: Include tree.h
188 * bb-reorder.c: Likewise.
189 * cfgcleanup.c: Likewise.
190 * cfgloopmanip.c: Likewise.
191 * data-streamer-in.c: Likewise.
192 * data-streamer-out.c: Likewise.
193 * data-streamer.c: Likewise.
194 * dwarf2cfi.c: Likewise.
195 * graphite-blocking.c: Likewise.
196 * graphite-clast-to-gimple.c: Likewise.
197 * graphite-dependences.c: Likewise.
198 * graphite-interchange.c: Likewise.
199 * graphite-optimize-isl.c: Likewise.
200 * graphite-poly.c: Likewise.
201 * graphite-scop-detection.c: Likewise.
202 * graphite-sese-to-poly.c: Likewise.
203 * graphite.c: Likewise.
204 * ipa-devirt.c: Likewise.
205 * ipa-profile.c: Likewise.
208 * loop-init.c: Likewise.
209 * loop-unroll.c: Likewise.
210 * lower-subreg.c: Likewise.
211 * lto/lto-object.c: Likewise.
213 * reginfo.c: Likewise.
214 * tree-loop-distribution.c: Likewise.
215 * tree-parloops.c: Likewise.
216 * tree-ssa-strlen.c: Likewise.
217 * tree-streamer.c: Likewise.
218 * value-prof.c: Likewise.
219 * target-globals.c: Likewise.
220 * expr.h: Include tree-core.h instead of tree.h.
221 * gimple.h: Likewise.
222 * ipa-prop.h: Likewise.
223 * ipa-utils.h: Likewise.
224 * lto-streamer.h: Likewise.
225 * streamer-hooks.h: Likewise.
226 * ipa-reference.h: Include cgraph.h instead of tree.h.
227 * cgraph.h: Include basic-block.h instead of tree.h.
228 * tree-streamer.h: Do not include tree.h.
229 * genattrtab.c (write_header): Generate inclusion of
231 * genautomata.c (main): Likewise.
232 * genemit.c: Likewise.
233 * genopinit.c: Likewise.
234 * genoutput.c (output_prologue): Likewise.
235 * genpeep.c: Likewise.
237 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
239 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
241 (vec_unpacku_hi_v8hi): Likewise.
242 (vec_unpacku_lo_v16qi): Likewise.
243 (vec_unpacku_lo_v8hi): Likewise.
245 2013-10-20 Jan Hubicka <jh@suse.cz>
247 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
248 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
249 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
251 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
253 * config/mips/mips.h (ISA_HAS_WSBH): Define.
254 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
256 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
258 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
261 * system.h: Undef m_slot.
263 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
265 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
266 all elements for both endian flavors.
268 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
271 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
272 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
273 and SI_REG for 64bit SYSV ABI targets.
275 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
277 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
278 to multi_reg_return. Clarify that we are skipping USEs of multiple
279 return registers. Use bool type where appropriate.
281 2013-10-18 Jan Hubicka <jh@suse.cz>
283 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
285 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
286 (X86_TUNE_PUSH_MEMORY): Likewise.
287 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
288 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
289 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
291 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
292 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
294 (ix86_option_override_internal): Update to use tune features instead
297 2013-10-18 Cong Hou <congh@google.com>
299 PR tree-optimization/58508
300 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
301 statement that contains data refs with zero-step.
303 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
305 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
306 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
307 * gimple-low.c (gimple_check_call_arg,
308 gimple_check_call_matching_types): Move to cgraph.c.
309 * gimple-low.h: Remove prototype.
310 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
311 Relocate from gimple-low.c.
312 * cgraph.h: Add prototype. Don't include basic-block.h.
313 * gimplify.c: Add gimple-low to include list.
314 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
315 * tree-eh.c: Add gimple-low to include list.
316 * tree-nested.c: Likewise.
317 * cfgexpand.c: Add tree-ssa-address.h to include list.
319 * gimple-fold.c: Likewise.
320 * gimple-ssa-strength-reduction.c: Likewise.
321 * trans-mem.c: Likewise.
322 * tree-mudflap.c: Likewise.
323 * tree-ssa-loop-ivopts.c: Likewise.
324 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
325 (degenerate_phi_result): Move to tree-phinodes.c.
326 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
327 * tree-ssa-threadedge.c: Likewise.
328 * tree-vrp.c: Likewise.
329 * tree-phinodes.c (degenerate_phi_result): Relocate here.
330 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
331 * tree-phinodes.h (degenerate_phi_result): Add prototype.
332 * tree-ssa-copy.c: Include tree-ssa-dom.h.
333 * tree-ssa-forwprop.c: Likewise.
334 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
335 pass_data_cleanup_cfg_post_optimizing,
336 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
337 * tree-optimize.c: Delete File.
338 * graphite.c: Include tree-cfgcleanup.h.
339 * passes.c: Likewise.
340 * tree-cfg.c: Likewise.
341 * tree-profile.c: Likewise.
342 * tree-ssa-dse.c: Likewise.
343 * tree-ssa-loop-ivcanon.c: Likewise.
344 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
345 * tree-outof-ssa.c: Include sbitmap.h.
346 * tree-ssa-live.c: Likewise.
347 * tree-ssa-propagate.c: Likewise.
348 * tree-ssa-structalias.c: Likewise.
349 * tree-stdarg.c: Likewise.
350 * Makefile.in (OBJS): Delete tree-optimize.o.
351 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
352 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
353 * varasm.c: Include basic-block.h.
354 * cfgloop.h: Include function.h instead of basic-block.h
355 (bb_loop_depth): Move to cfgloop.c.
356 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
358 2013-10-18 Teresa Johnson <tejohnson@google.com>
360 * predict.c (probably_never_executed): Compare frequency-based
361 count to number of training runs.
362 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
364 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
366 * config/arm/arm.c (cortexa9_extra_costs): New table.
367 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
369 2013-10-18 Jeff Law <law@redhat.com>
371 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
373 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
374 (register_jump_thread): Add "registered_jump_thread" debug
376 * dbgcnt.def (registered_jump_thread): New debug counter.
378 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
380 * config/rs6000/rs6000.c: Include cgraph.h.
382 2013-10-18 Teresa Johnson <tejohnson@google.com>
384 * tree-ssa-tail-merge.c (replace_block_by): Update edge
385 weights during merging.
387 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
389 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
390 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
391 * tree-ssa-operands.h: Remove prototype.
392 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
393 * gimple.c (virtual_operand_p): Relocate from gimple.c.
394 * gimple.h: Add prototype.
395 * gimple-ssa.h: Include tree-ssa-operands.h.
396 * tree-dump.c: Add tree-cfg.h to include list.
397 * tree-ssa-alias.c: Add ipa-reference.h to include list.
398 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
399 * config/i386/i386.c: Don't include tree-flow.h.
400 * config/rs6000/rs6000.c: Likewise.
402 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
404 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
406 2013-10-18 Richard Biener <rguenther@suse.de>
408 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
409 or TYPE_UNSIGNED of integral types.
410 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
411 NULL_TREE for zero-precision integral types.
413 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
415 * config/aarch64/arm_neon.h
416 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
418 2013-10-17 Sriraman Tallam <tmsriram@google.com>
420 * opth-gen.awk: Define target_flags_explicit.
422 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
424 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
425 fields to the reg_addr array that describes the valid addressing
426 mode for any register, general purpose registers, floating point
427 registers, and Altivec registers.
428 (FIRST_RELOAD_REG_CLASS): Likewise.
429 (LAST_RELOAD_REG_CLASS): Likewise.
430 (struct reload_reg_map_type): Likewise.
431 (reload_reg_map_type): Likewise.
432 (RELOAD_REG_VALID): Likewise.
433 (RELOAD_REG_MULTIPLE): Likewise.
434 (RELOAD_REG_INDEXED): Likewise.
435 (RELOAD_REG_OFFSET): Likewise.
436 (RELOAD_REG_PRE_INCDEC): Likewise.
437 (RELOAD_REG_PRE_MODIFY): Likewise.
438 (reg_addr): Likewise.
439 (mode_supports_pre_incdec_p): New helper functions to say whether
440 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
441 (mode_supports_pre_modify_p): Likewise.
442 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
443 print the valid address mode bits for each mode.
444 (rs6000_debug_print_mode): Likewise.
445 (rs6000_debug_reg_global): Likewise.
446 (rs6000_setup_reg_addr_masks): New function to set up the address
447 mask bits for each type.
448 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
449 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
450 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
451 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
452 PRE_MODIFY are supported.
453 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
454 registers, instead of {src,dest}_av_p.
455 (rs6000_print_options_internal): Tweak the debug output slightly.
457 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
459 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
462 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
464 * tree-flow.h (struct omp_region): Move to omp-low.c.
465 Remove omp_ prototypes and variables.
466 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
467 (copy_var_decl): Relocate prototype from tree-flow.h.
468 * gimple.c (copy_var_decl): Relocate from omp-low.c.
469 * tree.h: Move prototype to omp-low.h.
470 * omp-low.h: New File. Relocate prototypes here.
471 * omp-low.c (struct omp_region): Make local here.
472 (root_omp_region): Make static.
473 (copy_var_decl) Move to gimple.c.
474 (new_omp_region): Make static.
475 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
476 * tree-cfg.c: Include omp-low.h.
477 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
478 * gimplify.c: Include omp-low.h.
479 * tree-parloops.c: Likewise.
481 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
483 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
484 operands are in memory, prefer to force non-matched operand 1 to
487 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
490 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
491 restrict TImode addresses to single indirect registers if both
492 -mquad-memory and -mvsx-timode are used.
493 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
494 we should emit load/store quad. Remove using %y for quad memory
497 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
498 constraints to allow load/store quad on machines where TImode is
499 not allowed in VSX registers. Use 'n' instead of 'F' constraint
500 for TImode to load integer constants.
502 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
504 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
506 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
508 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
509 handling of STACK_REG.
511 2013-10-17 Richard Biener <rguenther@suse.de>
513 PR tree-optimization/58143
514 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
516 (rewrite_to_defined_overflow): Likewise.
517 (move_computations_dom_walker::before_dom): Rewrite stmts
518 with undefined signed overflow that are not always executed
519 into unsigned arithmetic.
521 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
524 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
525 explicit isa flag to be an options variable, instead of using
526 global_options_set. Remove define from rs6000.h.
527 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
529 * config/rs6000/rs6000.c (rs6000_option_override_internal):
530 Initialize rs6000_isa_flags_explicit.
531 (rs6000_function_specific_save): Add gcc_options* parameter, so
532 that the powerpc builds after the 2013-10-15 changes.
533 (rs6000_function_specific_restore): Likewise.
535 2013-10-16 DJ Delorie <dj@redhat.com>
537 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
538 op is a REG before checking REGNO.
539 (rl78_alloc_physical_registers): Verify pattern is a SET before
542 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
544 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
546 (vec_unpacks_lo_v4sf): Likewise.
547 (vec_unpacks_float_hi_v4si): Likewise.
548 (vec_unpacks_float_lo_v4si): Likewise.
549 (vec_unpacku_float_hi_v4si): Likewise.
550 (vec_unpacku_float_lo_v4si): Likewise.
552 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
554 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
555 (vsx_concat_v2sf): Likewise.
557 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
559 * config/aarch64/aarch64.md
560 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
562 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
564 PR tree-optimization/58697
565 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
566 estimated_loop_iterations_int.
567 (max_stmt_executions_int): Call get_max_loop_iterations_int.
568 (get_max_loop_iterations_int): New. HWINT version of
569 get_max_loop_iterations.
570 * cfgloop.h: Add prototypes.
571 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
572 * loop-unroll.c (decide_peel_once_rolling): Call
573 get_estimated_loop_iterations_int.
574 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
575 * tree-ssa-loop-niter.h: Tweak prototypes.
577 2013-10-16 David Malcolm <dmalcolm@redhat.com>
579 * gengtype-parse.c (struct_field_seq): Ignore access-control
580 keywords ("public:" etc).
582 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
584 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
585 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
587 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
589 * config/arm/arm.opt (mlra): New option.
590 * config/arm/arm.c (arm_lra_p): New function.
591 (TARGET_LRA_P): Define.
593 2013-10-16 Paulo Matos <pmatos@broadcom.com>
595 * tree-core.h (tree_code_name): Remove.
596 * tree.h (get_tree_code_name): New prototype.
597 * tree.c (tree_code_name): Make static.
598 (get_tree_code_name): New function.
599 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
600 tree_class_check_failed, tree_range_check_failed,
601 tree_not_class_check_failed, omp_clause_check_failed,
602 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
603 wrapper get_tree_code_name instead of calling tree_code_name directly.
604 * tree-vrp.c (dump_asserts_for): Likewise.
605 * tree-dump.c (dequeue_and_dump): Likewise.
606 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
607 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
608 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
609 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
610 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
611 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
612 dump_gimple_omp_for): Likewise.
613 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
614 * tree-ssa-pre.c (print_pre_expr): Likewise.
615 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
616 * print-tree.c (print_node_brief, print_node): Likewise.
617 * gimple.c (gimple_check_failed): Likewise.
618 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
619 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
620 * config/mep/mep.c (mep_validate_vliw): Likewise.
621 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
622 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
624 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
626 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
629 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
631 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
632 (MULTILIB_MATCHES): Add multilib for -march=v8.
634 2013-10-15 Sriraman Tallam <tmsriram@google.com>
637 * optc-save-gen.awk: Add extra parameter to the save and restore
639 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
640 * tree.c (build_optimization_node): New parameter. Add extra parameter
641 to call to cl_optimization_save.
642 (build_target_option_node): New parameter. Add extra parameter
643 to call to cl_target_option_save.
644 * tree.h (build_optimization_node): New parameter.
645 (build_target_option_node): New parameter.
646 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
647 build_optimization_node and build_target_option_node.
648 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
649 (handle_pragma_push_options): Ditto.
650 * toplev.c (process_options): Ditto.
651 * opts.c (init_options_struct): Check for opts_set non-null.
652 * target.def (target_option.save): New parameter.
653 (target_option.restore): New parameter.
655 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
656 (ix86_pragma_target_parse): Ditto.
657 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
659 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
660 to build_optimization_node and build_target_option_node.
661 (rs6000_valid_attribute_p): Ditto.
662 (rs6000_pragma_target_parse): Ditto.
663 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
665 * config/i386/i386.h:
666 TARGET_64BIT_P: New Macro
667 TARGET_MMX_P: New Macro.
668 TARGET_3DNOW_P: New Macro.
669 TARGET_3DNOW_A_P: New Macro.
670 TARGET_SSE_P: New Macro.
671 TARGET_SSE2_P: New Macro.
672 TARGET_SSE3_P: New Macro.
673 TARGET_SSSE3_P: New Macro.
674 TARGET_SSE4_1_P: New Macro.
675 TARGET_SSE4_2_P: New Macro.
676 TARGET_AVX_P: New Macro.
677 TARGET_AVX2_P: New Macro.
678 TARGET_AVX512F_P: New Macro.
679 TARGET_AVX512PF_P: New Macro.
680 TARGET_AVX512ER_P: New Macro.
681 TARGET_AVX512CD_P: New Macro.
682 TARGET_FMA_P: New Macro.
683 TARGET_SSE4A_P: New Macro.
684 TARGET_FMA4_P: New Macro.
685 TARGET_XOP_P: New Macro.
686 TARGET_LWP_P: New Macro.
687 TARGET_ABM_P: New Macro.
688 TARGET_BMI_P: New Macro.
689 TARGET_BMI2_P: New Macro.
690 TARGET_LZCNT_P: New Macro.
691 TARGET_TBM_P: New Macro.
692 TARGET_POPCNT_P: New Macro.
693 TARGET_SAHF_P: New Macro.
694 TARGET_MOVBE_P: New Macro.
695 TARGET_CRC32_P: New Macro.
696 TARGET_AES_P: New Macro.
697 TARGET_PCLMUL_P: New Macro.
698 TARGET_CMPXCHG16B_P: New Macro.
699 TARGET_FSGSBASE_P: New Macro.
700 TARGET_RDRND_P: New Macro.
701 TARGET_F16C_P: New Macro.
702 TARGET_RTM_P: New Macro.
703 TARGET_HLE_P: New Macro.
704 TARGET_RDSEED_P: New Macro.
705 TARGET_PRFCHW_P: New Macro.
706 TARGET_ADX_P: New Macro.
707 TARGET_FXSR_P: New Macro.
708 TARGET_XSAVE_P: New Macro.
709 TARGET_XSAVEOPT_P: New Macro.
710 TARGET_LP64_P: New Macro.
711 TARGET_X32_P: New Macro.
712 TARGET_FPMATH_DEFAULT_P: New Macro.
713 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
714 * config/i386/i386.c (ix86_option_override_internal): New parameters.
716 Change ix86_tune_string to access opts->x_ix86_tune_string.
717 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
718 Change ix86_arch_string to access opts->x_ix86_arch_string.
719 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
720 Change ix86_pmode to access opts->x_ix86_pmode.
721 Change ix86_abi to access opts->x_ix86_abi.
722 Change ix86_cmodel to access opts->x_ix86_cmodel.
723 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
724 Change ix86_isa_flags_explicit to access
725 opts->x_ix86_isa_flags_explicit.
726 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
727 Change ix86_regparm to access opts->x_ix86_regparm.
728 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
729 Change ix86_preferred_stack_boundary_arg to access
730 opts->x_ix86_preferred_stack_boundary_arg.
731 Change ix86_force_align_arg_pointer to access
732 opts->x_ix86_force_align_arg_pointer.
733 Change ix86_incoming_stack_boundar_arg to access
734 opts->x_ix86_incoming_stack_boundar_arg.
735 Change ix86_fpmath to access opts->x_ix86_fpmath.
736 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
737 Change ix86_recip_name to access opts->x_ix86_recip_name.
738 Change ix86_stack_protector_guard to access
739 opts->x_ix86_stack_protector_guard.
740 Change ix86_tune_memcpy_strategy to access
741 opts->x_ix86_tune_memcpy_strategy.
742 Change ix86_tune_memset_strategy to access
743 opts->x_ix86_tune_memset_strategy.
744 Change global_options to access opts.
745 Change global_options_set to access opts_set.
746 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
747 Change TARGET_MMX to TARGET_MMX_P (opts->...).
748 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
749 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
750 Change TARGET_SSE to TARGET_SSE_P (opts->...).
751 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
752 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
753 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
754 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
755 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
756 Change TARGET_AVX to TARGET_AVX_P (opts->...).
757 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
758 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
759 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
760 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
761 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
762 Change TARGET_FMA to TARGET_FMA_P (opts->...).
763 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
764 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
765 Change TARGET_XOP to TARGET_XOP_P (opts->...).
766 Change TARGET_LWP to TARGET_LWP_P (opts->...).
767 Change TARGET_ABM to TARGET_ABM_P (opts->...).
768 Change TARGET_BMI to TARGET_BMI_P (opts->...).
769 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
770 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
771 Change TARGET_TBM to TARGET_TBM_P (opts->...).
772 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
773 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
774 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
775 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
776 Change TARGET_AES to TARGET_AES_P (opts->...).
777 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
778 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
779 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
780 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
781 Change TARGET_F16C to TARGET_F16C_P (opts->...).
782 Change TARGET_RTM to TARGET_RTM_P (opts->...).
783 Change TARGET_HLE to TARGET_HLE_P (opts->...).
784 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
785 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
786 Change TARGET_ADX to TARGET_ADX_P (opts->...).
787 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
788 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
789 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
790 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
791 Change TARGET_X32 to TARGET_X32_P (opts->...).
792 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
793 Change TARGET_FLOAT_RETURNS_IN_80387 to
794 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
795 (ix86_function_specific_save): New parameter. Use opts-> fields
796 to replace global fields.
797 (ix86_function_specific_restore): Ditto.
798 (ix86_valid_target_attribute_inner_p): New parameters.
800 Fix call to ix86_handle_option and set_option.
801 (ix86_valid_target_attribute_tree): New parameters.
802 Change global_options to access opts.
803 Change global_options_set to access opts_set.
804 Fix call to ix86_valid_target_attribute_inner_p.
805 Change ix86_tune_string to access opts->x_ix86_tune_string.
806 Change ix86_arch_string to access opts->x_ix86_arch_string.
807 Change ix86_fpmath to access opts->x_ix86_fpmath
808 Fix call to ix86_option_override_internal.
809 Fix call to ix86_add_new_builtins.
810 Fix calls to build_optimization_node and build_target_option_node.
811 (ix86_valid_target_attribute_p): Remove access to global_options.
812 Use new gcc_options structure func_options.
813 Fix call to ix86_valid_target_attribute_tree.
814 Fix call to build_optimization_node.
815 (get_builtin_code_for_version): Fix call to
816 ix86_valid_target_attribute_tree.
818 2013-10-15 David Malcolm <dmalcolm@redhat.com>
820 * Makefile.in (PICFLAG): New.
821 (enable_host_shared): New.
822 (INTERNAL_CFLAGS): Use PICFLAG.
823 (LIBIBERTY): Use pic build of libiberty.a if configured with
824 --enable-host-shared.
825 * configure.ac: Add --enable-host-shared, setting up new
827 * configure: Regenerate.
828 * doc/install.texi (--enable-shared): Add note contrasting it with ...
829 (--enable-host-shared): New option.
831 2013-10-15 Richard Biener <rguenther@suse.de>
833 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
834 for built-in functions.
836 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
838 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
839 (optimize_range_tests_1): New function,
840 extracted from optimize_range_tests.
841 (optimize_range_tests_xor): Similarly.
842 (optimize_range_tests_diff): New function.
843 (optimize_range_tests): Use optimize_range_tests_1.
845 2013-10-15 Cong Hou <congh@google.com>
847 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
848 requirement of the reduction pattern so that one operand of the
849 reduction operation can come from outside of the loop.
851 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
853 * config/arm/neon-schedgen.ml: Remove.
854 * config/arm/cortex-a9-neon.md: Remove comment regarding
857 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
859 * config/arm/types: Remove old neon types.
861 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
863 * config/arm/cortex-a7.md
864 (cortex_a7_neon_type): New.
865 (cortex_a7_neon_mul): Update for new types.
866 (cortex_a7_neon_mla): Likewise.
867 (cortex_a7_neon): Likewise.
869 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
871 * config/arm/cortex-a15-neon.md
872 (cortex_a15_neon_type): New,
874 (cortex_a15_neon_int_1): Remove.
875 (cortex_a15_neon_int_2): Likewise.
876 (cortex_a15_neon_int_3): Likewise.
877 (cortex_a15_neon_int_4): Likewise.
878 (cortex_a15_neon_int_5): Likewise.
879 (cortex_a15_neon_vqneg_vqabs): Likewise.
880 (cortex_a15_neon_vmov): Likewise.
881 (cortex_a15_neon_vaba): Likewise.
882 (cortex_a15_neon_vaba_qqq): Likewise.
883 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
884 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
885 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
887 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
888 (cortex_a15_neon_mla_qqq_8_16): Likewise.
889 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
890 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
891 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
892 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
893 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
894 (cortex_a15_neon_shift_1): Likewise.
895 (cortex_a15_neon_shift_2): Likewise.
896 (cortex_a15_neon_shift_3): Likewise.
897 (cortex_a15_neon_vshl_ddd): Likewise.
898 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
899 (cortex_a15_neon_vsra_vrsra): Likewise.
900 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
901 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
902 (cortex_a15_neon_bp_3cycle): Likewise.
903 (cortex_a15_neon_ldm_2): Likewise.
904 (cortex_a15_neon_stm_2): Likewise.
905 (cortex_a15_neon_mcr): Likewise.
906 (cortex_a15_neon_mrc): Likewise.
907 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
908 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
909 (cortex_a15_neon_fp_vmul_ddd): Likewise.
910 (cortex_a15_neon_fp_vmul_qqd): Likewise.
911 (cortex_a15_neon_fp_vmla_ddd): Likewise.
912 (cortex_a15_neon_fp_vmla_qqq): Likewise.
913 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
914 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
915 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
916 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
917 (cortex_a15_neon_bp_simple): Likewise.
918 (cortex_a15_neon_bp_2cycle): Likewise.
919 (cortex_a15_neon_bp_3cycle): Likewise.
920 (cortex_a15_neon_vld1_1_2_regs): Likewise.
921 (cortex_a15_neon_vld1_3_4_regs): Likewise.
922 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
923 (cortex_a15_neon_vld2_4_regs): Likewise.
924 (cortex_a15_neon_vld3_vld4): Likewise.
925 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
926 (cortex_a15_neon_vst1_3_4_regs): Likewise.
927 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
928 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
929 (cortex_a15_neon_vst3_vst4): Rename to...
930 (cortex_a15_neon_vst4): This, update for new attributes.
931 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
932 (cortex_a15_neon_vld3_vld4_lane): Likewise.
933 (cortex_a15_neon_vst1_vst2_lane): Likewise.
934 (cortex_a15_neon_vst3_vst4_lane): Likewise.
935 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
936 (cortex_a15_neon_ldm_2): Likewise.
937 (cortex_a15_neon_stm_2): Likewise.
938 (cortex_a15_neon_mcr): Likewise.
939 (cortex_a15_neon_mcr_2_mcrr): Likewise.
940 (cortex_a15_neon_mrc): Likewise.
941 (cortex_a15_neon_mrrc): Likewise.
943 (cortex_a15_neon_abd): New.
944 (cortex_a15_neon_abd_q): Likewise.
945 (cortex_a15_neon_aba): Likewise.
946 (cortex_a15_neon_aba_q): Likewise.
947 (cortex_a15_neon_acc): Likewise.
948 (cortex_a15_neon_acc_q): Likewise.
949 (cortex_a15_neon_arith_basic): Likewise.
950 (cortex_a15_neon_arith_complex): Likewise.
951 (cortex_a15_neon_multiply): Likewise.
952 (cortex_a15_neon_multiply_q): Likewise.
953 (cortex_a15_neon_mla): Likewise.
954 (cortex_a15_neon_mla_q): Likewise.
955 (cortex_a15_neon_sat_mla_long): Likewise.
956 (cortex_a15_neon_shift_acc): Likewise.
957 (cortex_a15_neon_shift_imm_basic): Likewise.
958 (cortex_a15_neon_shift_imm_complex): Likewise.
959 (cortex_a15_neon_shift_reg_basic): Likewise.
960 (cortex_a15_neon_shift_reg_basic_q): Likewise.
961 (cortex_a15_neon_shift_reg_complex): Likewise.
962 (cortex_a15_neon_shift_reg_complex_q): Likewise.
963 (cortex_a15_neon_fp_negabs): Likewise
964 (cortex_a15_neon_fp_arith): Likewise
965 (cortex_a15_neon_fp_arith_q): Likewise
966 (cortex_a15_neon_fp_cvt_int): Likewise
967 (cortex_a15_neon_fp_cvt_int_q): Likewise
968 (cortex_a15_neon_fp_cvt_16): Likewise
969 (cortex_a15_neon_fp_mul): Likewise
970 (cortex_a15_neon_fp_mul_q): Likewise
971 (cortex_a15_neon_fp_mla): Likewise
972 (cortex_a15_neon_fp_mla_q): Likewise
973 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
974 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
975 (cortex_a15_neon_bitops): Likewise.
976 (cortex_a15_neon_bitops_q): Likewise.
977 (cortex_a15_neon_from_gp): Likewise.
978 (cortex_a15_neon_from_gp_q): Likewise.
979 (cortex_a15_neon_tbl3_tbl4): Likewise.
980 (cortex_a15_neon_zip_q): Likewise.
981 (cortex_a15_neon_to_gp): Likewise.
982 (cortex_a15_neon_load_a): Likewise.
983 (cortex_a15_neon_load_b): Likewise.
984 (cortex_a15_neon_load_c): Likewise.
985 (cortex_a15_neon_load_d): Likewise.
986 (cortex_a15_neon_load_e): Likewise.
987 (cortex_a15_neon_load_f): Likewise.
988 (cortex_a15_neon_store_a): Likewise.
989 (cortex_a15_neon_store_b): Likewise.
990 (cortex_a15_neon_store_c): Likewise.
991 (cortex_a15_neon_store_d): Likewise.
992 (cortex_a15_neon_store_e): Likewise.
993 (cortex_a15_neon_store_f): Likewise.
994 (cortex_a15_neon_store_g): Likewise.
995 (cortex_a15_neon_store_h): Likewise.
996 (cortex_a15_vfp_to_from_gp): Likewise.
998 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1000 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
1002 (cortex_a9_neon_vshl_ddd): Remove.
1003 (cortex_a9_neon_vst3_vst4): Likewise.
1004 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
1006 (cortex_a9_neon_bit_ops_q): New.
1008 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
1009 (cortex_a9_neon_int_2): Likewise.
1010 (cortex_a9_neon_int_3): Likewise.
1011 (cortex_a9_neon_int_4): Likewise.
1012 (cortex_a9_neon_int_5): Likewise.
1013 (cortex_a9_neon_vqneg_vqabs): Likewise.
1014 (cortex_a9_neon_vmov): Likewise.
1015 (cortex_a9_neon_vaba): Likewise.
1016 (cortex_a9_neon_vaba_qqq): Likewise.
1017 (cortex_a9_neon_shift_1): Likewise.
1018 (cortex_a9_neon_shift_2): Likewise.
1019 (cortex_a9_neon_shift_3): Likewise.
1020 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1021 (cortex_a9_neon_vsra_vrsra): Likewise.
1022 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1023 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1024 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1026 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1027 (cortex_a9_neon_mla_qqq_8_16): Likewise.
1028 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
1030 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1031 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1032 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
1033 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1034 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
1035 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
1036 (cortex_a9_neon_fp_vsum): Likewise.
1037 (cortex_a9_neon_fp_vmul_ddd): Likewise.
1038 (cortex_a9_neon_fp_vmul_qqd): Likewise.
1039 (cortex_a9_neon_fp_vmla_ddd): Likewise.
1040 (cortex_a9_neon_fp_vmla_qqq): Likewise.
1041 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
1042 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
1043 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1044 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1045 (cortex_a9_neon_bp_simple): Likewise.
1046 (cortex_a9_neon_bp_2cycle): Likewise.
1047 (cortex_a9_neon_bp_3cycle): Likewise.
1048 (cortex_a9_neon_ldr): Likewise.
1049 (cortex_a9_neon_str): Likewise.
1050 (cortex_a9_neon_vld1_1_2_regs): Likewise.
1051 (cortex_a9_neon_vld1_3_4_regs): Likewise.
1052 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1053 (cortex_a9_neon_vld2_4_regs): Likewise.
1054 (cortex_a9_neon_vld3_vld4): Likewise.
1055 (cortex_a9_neon_vld1_vld2_lane): Likewise.
1056 (cortex_a9_neon_vld3_vld4_lane): Likewise.
1057 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
1058 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1059 (cortex_a9_neon_vst1_3_4_regs): Likewise.
1060 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
1061 (cortex_a9_neon_vst1_vst2_lane): Likewise.
1062 (cortex_a9_neon_vst3_vst4_lane): Likewise.
1063 (cortex_a9_neon_mcr): Likewise.
1064 (cortex_a9_neon_mcr_2_mcrr): Likewise.
1065 (cortex_a9_neon_mrc): Likewise.
1066 (cortex_a9_neon_mrrc): Likewise.
1068 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1070 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
1072 (cortex_a8_neon_vshl_ddd): Remove.
1073 (cortex_a8_neon_vst3_vst4): Likewise.
1074 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
1076 (cortex_a8_neon_bit_ops_q): New.
1078 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
1079 (cortex_a8_neon_int_2): Likewise..
1080 (cortex_a8_neon_int_3): Likewise.
1081 (cortex_a8_neon_int_5): Likewise.
1082 (cortex_a8_neon_vqneg_vqabs): Likewise.
1083 (cortex_a8_neon_int_4): Likewise.
1084 (cortex_a8_neon_vaba): Likewise.
1085 (cortex_a8_neon_vaba_qqq): Likewise.
1086 (cortex_a8_neon_shift_1): Likewise.
1087 (cortex_a8_neon_shift_2): Likewise.
1088 (cortex_a8_neon_shift_3): Likewise.
1089 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1090 (cortex_a8_neon_vsra_vrsra): Likewise.
1091 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1092 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1093 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1095 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1096 (cortex_a8_neon_mla_qqq_8_16): Likewise.
1097 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
1099 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1100 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1101 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
1102 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1103 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
1104 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
1105 (cortex_a8_neon_fp_vsum): Likewise.
1106 (cortex_a8_neon_fp_vmul_ddd): Likewise.
1107 (cortex_a8_neon_fp_vmul_qqd): Likewise.
1108 (cortex_a8_neon_fp_vmla_ddd): Likewise.
1109 (cortex_a8_neon_fp_vmla_qqq): Likewise.
1110 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
1111 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
1112 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1113 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1114 (cortex_a8_neon_bp_simple): Likewise.
1115 (cortex_a8_neon_bp_2cycle): Likewise.
1116 (cortex_a8_neon_bp_3cycle): Likewise.
1117 (cortex_a8_neon_ldr): Likewise.
1118 (cortex_a8_neon_str): Likewise.
1119 (cortex_a8_neon_vld1_1_2_regs): Likewise.
1120 (cortex_a8_neon_vld1_3_4_regs): Likewise.
1121 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1122 (cortex_a8_neon_vld2_4_regs): Likewise.
1123 (cortex_a8_neon_vld3_vld4): Likewise.
1124 (cortex_a8_neon_vld1_vld2_lane): Likewise.
1125 (cortex_a8_neon_vld3_vld4_lane): Likewise.
1126 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1127 (cortex_a8_neon_vst1_3_4_regs): Likewise.
1128 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
1129 (cortex_a8_neon_vst1_vst2_lane): Likewise.
1130 (cortex_a8_neon_vst3_vst4_lane): Likewise.
1131 (cortex_a8_neon_mcr): Likewise.
1132 (cortex_a8_neon_mcr_2_mcrr): Likewise.
1133 (cortex_a8_neon_mrc): Likewise.
1134 (cortex_a8_neon_mrrc): Likewise.
1136 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1138 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
1140 * config/aarch64/aarch64-simd.md (neon_type): Remove.
1141 (aarch64_simd_dup<mode>): Add "type" attribute.
1142 (aarch64_dup_lane<mode>): Likewise.
1143 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
1144 (*aarch64_simd_mov<mode>): Likewise.
1145 (aarch64_simd_mov_from_<mode>low): Likewise.
1146 (aarch64_simd_mov_from_<mode>high): Likewise.
1147 (orn<mode>3): Likewise.
1148 (bic<mode>3): Likewise.
1149 (add<mode>3): Likewise.
1150 (sub<mode>3): Likewise.
1151 (mul<mode>3): Likewise.
1152 (*aarch64_mul3_elt<mode>): Likewise.
1153 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
1154 (*aarch64_mul3_elt_to_128df): Likewise.
1155 (*aarch64_mul3_elt_to_64v2df): Likewise.
1156 (neg<mode>2): Likewise.
1157 (abs<mode>2): Likewise.
1158 (abd<mode>_3): Likewise.
1159 (aba<mode>_3): Likewise.
1160 (fabd<mode>_3): Likewise.
1161 (*fabd_scalar<mode>3): Likewise.
1162 (and<mode>3): Likewise.
1163 (ior<mode>3): Likewise.
1164 (xor<mode>3): Likewise.
1165 (one_cmpl<mode>2): Likewise.
1166 (aarch64_simd_vec_set<mode>): Likewise.
1167 (aarch64_simd_lshr<mode>): Likewise.
1168 (aarch64_simd_ashr<mode>): Likewise.
1169 (aarch64_simd_imm_shl<mode>): Likewise.
1170 (aarch64_simd_reg_sshl<mode): Likewise.
1171 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
1172 (aarch64_simd_reg_shl<mode>_signed): Likewise.
1173 (aarch64_simd_vec_setv2di): Likewise.
1174 (aarch64_simd_vec_set<mode>): Likewise.
1175 (aarch64_mla<mode>): Likewise.
1176 (*aarch64_mla_elt<mode>): Likewise.
1177 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
1178 (aarch64_mls<mode>): Likewise.
1179 (*aarch64_mls_elt<mode>): Likewise.
1180 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
1181 (<su><maxmin><mode>3): Likewise.
1182 (move_lo_quad_<mode>): Likewise.
1183 (aarch64_simd_move_hi_quad_<mode>): Likewise.
1184 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
1185 (vec_pack_trunc_<mode>): Likewise.
1186 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
1187 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
1188 (*aarch64_<su>mlal_lo<mode>): Likewise.
1189 (*aarch64_<su>mlal_hi<mode>): Likewise.
1190 (*aarch64_<su>mlsl_lo<mode>): Likewise.
1191 (*aarch64_<su>mlsl_hi<mode>): Likewise.
1192 (*aarch64_<su>mlal<mode>): Likewise.
1193 (*aarch64_<su>mlsl<mode>): Likewise.
1194 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
1195 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
1196 (add<mode>3): Likewise.
1197 (sub<mode>3): Likewise.
1198 (mul<mode>3): Likewise.
1199 (div<mode>3): Likewise.
1200 (neg<mode>2): Likewise.
1201 (abs<mode>2): Likewise.
1202 (fma<mode>4): Likewise.
1203 (*aarch64_fma4_elt<mode>): Likewise.
1204 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
1205 (*aarch64_fma4_elt_to_128df): Likewise.
1206 (*aarch64_fma4_elt_to_64v2df): Likewise.
1207 (fnma<mode>4): Likewise.
1208 (*aarch64_fnma4_elt<mode>): Likewise.
1209 (*aarch64_fnma4_elt_<vswap_width_name><mode>
1210 (*aarch64_fnma4_elt_to_128df): Likewise.
1211 (*aarch64_fnma4_elt_to_64v2df): Likewise.
1212 (<frint_pattern><mode>2): Likewise.
1213 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
1214 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
1215 (vec_unpacks_lo_v4sf): Likewise.
1216 (aarch64_float_extend_lo_v2df): Likewise.
1217 (vec_unpacks_hi_v4sf): Likewise.
1218 (aarch64_float_truncate_lo_v2sf): Likewise.
1219 (aarch64_float_truncate_hi_v4sf): Likewise.
1220 (aarch64_vmls<mode>): Likewise.
1221 (<su><maxmin><mode>3): Likewise.
1222 (<maxmin_uns><mode>3): Likewise.
1223 (reduc_<sur>plus_<mode>): Likewise.
1224 (reduc_<sur>plus_v2di): Likewise.
1225 (reduc_<sur>plus_v2si): Likewise.
1226 (reduc_<sur>plus_<mode>): Likewise.
1227 (aarch64_addpv4sf): Likewise.
1228 (clz<mode>2): Likewise.
1229 (reduc_<maxmin_uns>_<mode>): Likewise.
1230 (reduc_<maxmin_uns>_v2di): Likewise.
1231 (reduc_<maxmin_uns>_v2si): Likewise.
1232 (reduc_<maxmin_uns>_<mode>): Likewise.
1233 (reduc_<maxmin_uns>_v4sf): Likewise.
1234 (aarch64_simd_bsl<mode>_internal): Likewise.
1235 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
1236 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
1237 (aarch64_get_lane<mode>): Likewise.
1238 (*aarch64_combinez<mode>): Likewise.
1239 (aarch64_combine<mode>): Likewise.
1240 (aarch64_simd_combine<mode>): Likewise.
1241 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
1242 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
1243 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
1244 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
1245 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
1246 (aarch64_<sur>h<addsub><mode>): Likewise.
1247 (aarch64_<sur><addsub>hn<mode>): Likewise.
1248 (aarch64_<sur><addsub>hn2<mode>): Likewise.
1249 (aarch64_pmul<mode>): Likewise.
1250 (aarch64_<su_optab><optab><mode>): Likewise.
1251 (aarch64_<sur>qadd<mode>): Likewise.
1252 (aarch64_sqmovun<mode>): Likewise.
1253 (aarch64_<sur>qmovn<mode>): Likewise.
1254 (aarch64_s<optab><mode>): Likewise.
1255 (aarch64_sq<r>dmulh<mode>): Likewise.
1256 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1257 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
1258 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1259 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
1260 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1261 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1262 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
1263 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
1264 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
1265 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
1266 (aarch64_sqdmull<mode>): Likewise.
1267 (aarch64_sqdmull_lane<mode>_internal): Likewise.
1268 (aarch64_sqdmull_n<mode>): Likewise.
1269 (aarch64_sqdmull2<mode>_internal): Likewise.
1270 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
1271 (aarch64_sqdmull2_n<mode>_internal): Likewise.
1272 (aarch64_<sur>shl<mode>): Likewise.
1273 (aarch64_<sur>q<r>shl<mode>
1274 (aarch64_<sur>shll_n<mode>): Likewise.
1275 (aarch64_<sur>shll2_n<mode>): Likewise.
1276 (aarch64_<sur>shr_n<mode>): Likewise.
1277 (aarch64_<sur>sra_n<mode>): Likewise.
1278 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
1279 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
1280 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
1281 (aarch64_cm<optab><mode>): Likewise.
1282 (aarch64_cm<optab>di): Likewise.
1283 (aarch64_cm<optab><mode>): Likewise.
1284 (aarch64_cm<optab>di): Likewise.
1285 (aarch64_cmtst<mode>): Likewise.
1286 (aarch64_cmtstdi): Likewise.
1287 (aarch64_cm<optab><mode>): Likewise.
1288 (*aarch64_fac<optab><mode>): Likewise.
1289 (aarch64_addp<mode>): Likewise.
1290 (aarch64_addpdi): Likewise.
1291 (sqrt<mode>2): Likewise.
1292 (vec_load_lanesoi<mode>): Likewise.
1293 (vec_store_lanesoi<mode>): Likewise.
1294 (vec_load_lanesci<mode>): Likewise.
1295 (vec_store_lanesci<mode>): Likewise.
1296 (vec_load_lanesxi<mode>): Likewise.
1297 (vec_store_lanesxi<mode>): Likewise.
1298 (*aarch64_mov<mode>): Likewise.
1299 (aarch64_ld2<mode>_dreg): Likewise.
1300 (aarch64_ld2<mode>_dreg): Likewise.
1301 (aarch64_ld3<mode>_dreg): Likewise.
1302 (aarch64_ld3<mode>_dreg): Likewise.
1303 (aarch64_ld4<mode>_dreg): Likewise.
1304 (aarch64_ld4<mode>_dreg): Likewise.
1305 (aarch64_tbl1<mode>): Likewise.
1306 (aarch64_tbl2v16qi): Likewise.
1307 (aarch64_combinev16qi): Likewise.
1308 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
1309 (aarch64_st2<mode>_dreg): Likewise.
1310 (aarch64_st2<mode>_dreg): Likewise.
1311 (aarch64_st3<mode>_dreg): Likewise.
1312 (aarch64_st3<mode>_dreg): Likewise.
1313 (aarch64_st4<mode>_dreg): Likewise.
1314 (aarch64_st4<mode>_dreg): Likewise.
1315 (*aarch64_simd_ld1r<mode>): Likewise.
1316 (aarch64_frecpe<mode>): Likewise.
1317 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
1318 (aarch64_frecps<mode>): Likewise.
1320 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1322 * config/arm/iterators.md (V_elem_ch): New.
1324 (VQH_type): Likewise.
1325 * config/arm/arm.md (is_neon_type): New.
1326 (conds): Use is_neon_type.
1327 (anddi3_insn): Update type attribute.
1328 (xordi3_insn): Likewise.
1329 (one_cmpldi2): Likewise.
1330 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
1331 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
1332 (*movmisalign<mode>_neon_store): Likewise.
1333 (*movmisalign<mode>_neon_load): Likewise.
1334 (vec_set<mode>_internal): Likewise.
1335 (vec_set<mode>_internal): Likewise.
1336 (vec_setv2di_internal): Likewise.
1337 (vec_extract<mode>): Likewise.
1338 (vec_extract<mode>): Likewise.
1339 (vec_extractv2di): Likewise.
1340 (*add<mode>3_neon): Likewise.
1341 (adddi3_neon): Likewise.
1342 (*sub<mode>3_neon): Likewise.
1343 (subdi3_neon): Likewise.
1344 (fma<VCVTF:mode>4): Likewise.
1345 (fma<VCVTF:mode>4_intrinsic): Likewise.
1346 (*fmsub<VCVTF:mode>4): Likewise.
1347 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
1348 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
1349 (ior<mode>3): Likewise.
1350 (and<mode>3): Likewise.
1351 (orn<mode>3_neon): Likewise.
1352 (orndi3_neon): Likewise.
1353 (bic<mode>3_neon): Likewise.
1354 (bicdi3_neon): Likewise.
1355 (xor<mode>3): Likewise.
1356 (one_cmpl<mode>2): Likewise.
1357 (abs<mode>2): Likewise.
1358 (neg<mode>2): Likewise.
1359 (negdi2_neon): Likewise.
1360 (*umin<mode>3_neon): Likewise.
1361 (*umax<mode>3_neon): Likewise.
1362 (*smin<mode>3_neon): Likewise.
1363 (*smax<mode>3_neon): Likewise.
1364 (vashl<mode>3): Likewise.
1365 (vashr<mode>3_imm): Likewise.
1366 (vlshr<mode>3_imm): Likewise.
1367 (ashl<mode>3_signed): Likewise.
1368 (ashl<mode>3_unsigned): Likewise.
1369 (neon_load_count): Likewise.
1370 (ashldi3_neon_noclobber): Likewise.
1371 (ashldi3_neon): Likewise.
1372 (signed_shift_di3_neon): Likewise.
1373 (unsigned_shift_di3_neon): Likewise.
1374 (ashrdi3_neon_imm_noclobber): Likewise.
1375 (lshrdi3_neon_imm_noclobber): Likewise.
1376 (<shift>di3_neon): Likewise.
1377 (widen_ssum<mode>3): Likewise.
1378 (widen_usum<mode>3): Likewise.
1379 (quad_halves_<code>v4si): Likewise.
1380 (quad_halves_<code>v4sf): Likewise.
1381 (quad_halves_<code>v8hi): Likewise.
1382 (quad_halves_<code>v16qi): Likewise.
1383 (reduc_splus_v2di): Likewise.
1384 (neon_vpadd_internal<mode>): Likewise.
1385 (neon_vpsmin<mode>): Likewise.
1386 (neon_vpsmax<mode>): Likewise.
1387 (neon_vpumin<mode>): Likewise.
1388 (neon_vpumax<mode>): Likewise.
1389 (*ss_add<mode>_neon): Likewise.
1390 (*us_add<mode>_neon): Likewise.
1391 (*ss_sub<mode>_neon): Likewise.
1392 (*us_sub<mode>_neon): Likewise.
1393 (neon_vadd<mode>_unspec): Likewise.
1394 (neon_vaddl<mode>): Likewise.
1395 (neon_vaddw<mode>): Likewise.
1396 (neon_vhadd<mode>): Likewise.
1397 (neon_vqadd<mode>): Likewise.
1398 (neon_vaddhn<mode>): Likewise.
1399 (neon_vmul<mode>): Likewise.
1400 (neon_vfms<VCVTF:mode>): Likewise.
1401 (neon_vmlal<mode>): Likewise.
1402 (neon_vmls<mode>): Likewise.
1403 (neon_vmlsl<mode>): Likewise.
1404 (neon_vqdmulh<mode>): Likewise.
1405 (neon_vqdmlal<mode>): Likewise.
1406 (neon_vqdmlsl<mode>): Likewise.
1407 (neon_vmull<mode>): Likewise.
1408 (neon_vqdmull<mode>): Likewise.
1409 (neon_vsub<mode>_unspec): Likewise.
1410 (neon_vsubl<mode>): Likewise.
1411 (neon_vsubw<mode>): Likewise.
1412 (neon_vqsub<mode>): Likewise.
1413 (neon_vhsub<mode>): Likewise.
1414 (neon_vsubhn<mode>): Likewise.
1415 (neon_vceq<mode>): Likewise.
1416 (neon_vcge<mode>): Likewise.
1417 (neon_vcgeu<mode>): Likewise.
1418 (neon_vcgt<mode>): Likewise.
1419 (neon_vcgtu<mode>): Likewise.
1420 (neon_vcle<mode>): Likewise.
1421 (neon_vclt<mode>): Likewise.
1422 (neon_vcage<mode>): Likewise.
1423 (neon_vcagt<mode>): Likewise.
1424 (neon_vtst<mode>): Likewise.
1425 (neon_vabd<mode>): Likewise.
1426 (neon_vabdl<mode>): Likewise.
1427 (neon_vaba<mode>): Likewise.
1428 (neon_vabal<mode>): Likewise.
1429 (neon_vmax<mode>): Likewise.
1430 (neon_vmin<mode>): Likewise.
1431 (neon_vpaddl<mode>): Likewise.
1432 (neon_vpadal<mode>): Likewise.
1433 (neon_vpmax<mode>): Likewise.
1434 (neon_vpmin<mode>): Likewise.
1435 (neon_vrecps<mode>): Likewise.
1436 (neon_vrsqrts<mode>): Likewise.
1437 (neon_vqabs<mode>): Likewise.
1438 (neon_vqneg<mode>): Likewise.
1439 (neon_vcls<mode>): Likewise.
1440 (clz<mode>2): Likewise.
1441 (popcount<mode>2): Likewise.
1442 (neon_vrecpe<mode>): Likewise.
1443 (neon_vrsqrte<mode>): Likewise.
1444 (neon_vget_lane<mode>_sext_internal): Likewise.
1445 (neon_vget_lane<mode>_zext_internal): Likewise.
1446 (neon_vdup_n<mode>): Likewise.
1447 (neon_vdup_n<mode>): Likewise.
1448 (neon_vdup_nv2di): Likewise.
1449 (neon_vdup_lane<mode>_interal): Likewise.
1450 (*neon_vswp<mode>): Likewise.
1451 (neon_vcombine<mode>): Likewise.
1452 (float<mode><V_cvtto>2): Likewise.
1453 (floatuns<mode><V_cvtto>2): Likewise.
1454 (fix_trunc<mode><V_cvtto>2): Likewise.
1455 (fixuns_trunc<mode><V_cvtto>2
1456 (neon_vcvt<mode>): Likewise.
1457 (neon_vcvt<mode>): Likewise.
1458 (neon_vcvtv4sfv4hf): Likewise.
1459 (neon_vcvtv4hfv4sf): Likewise.
1460 (neon_vcvt_n<mode>): Likewise.
1461 (neon_vcvt_n<mode>): Likewise.
1462 (neon_vmovn<mode>): Likewise.
1463 (neon_vqmovn<mode>): Likewise.
1464 (neon_vqmovun<mode>): Likewise.
1465 (neon_vmovl<mode>): Likewise.
1466 (neon_vmul_lane<mode>): Likewise.
1467 (neon_vmul_lane<mode>): Likewise.
1468 (neon_vmull_lane<mode>): Likewise.
1469 (neon_vqdmull_lane<mode>): Likewise.
1470 (neon_vqdmulh_lane<mode>): Likewise.
1471 (neon_vqdmulh_lane<mode>): Likewise.
1472 (neon_vmla_lane<mode>): Likewise.
1473 (neon_vmla_lane<mode>): Likewise.
1474 (neon_vmlal_lane<mode>): Likewise.
1475 (neon_vqdmlal_lane<mode>): Likewise.
1476 (neon_vmls_lane<mode>): Likewise.
1477 (neon_vmls_lane<mode>): Likewise.
1478 (neon_vmlsl_lane<mode>): Likewise.
1479 (neon_vqdmlsl_lane<mode>): Likewise.
1480 (neon_vext<mode>): Likewise.
1481 (neon_vrev64<mode>): Likewise.
1482 (neon_vrev32<mode>): Likewise.
1483 (neon_vrev16<mode>): Likewise.
1484 (neon_vbsl<mode>_internal): Likewise.
1485 (neon_vshl<mode>): Likewise.
1486 (neon_vqshl<mode>): Likewise.
1487 (neon_vshr_n<mode>): Likewise.
1488 (neon_vshrn_n<mode>): Likewise.
1489 (neon_vqshrn_n<mode>): Likewise.
1490 (neon_vqshrun_n<mode>): Likewise.
1491 (neon_vshl_n<mode>): Likewise.
1492 (neon_vqshl_n<mode>): Likewise.
1493 (neon_vqshlu_n<mode>): Likewise.
1494 (neon_vshll_n<mode>): Likewise.
1495 (neon_vsra_n<mode>): Likewise.
1496 (neon_vsri_n<mode>): Likewise.
1497 (neon_vsli_n<mode>): Likewise.
1498 (neon_vtbl1v8qi): Likewise.
1499 (neon_vtbl2v8qi): Likewise.
1500 (neon_vtbl3v8qi): Likewise.
1501 (neon_vtbl4v8qi): Likewise.
1502 (neon_vtbl1v16qi): Likewise.
1503 (neon_vtbl2v16qi): Likewise.
1504 (neon_vcombinev16qi): Likewise.
1505 (neon_vtbx1v8qi): Likewise.
1506 (neon_vtbx2v8qi): Likewise.
1507 (neon_vtbx3v8qi): Likewise.
1508 (neon_vtbx4v8qi): Likewise.
1509 (*neon_vtrn<mode>_insn): Likewise.
1510 (*neon_vzip<mode>_insn): Likewise.
1511 (*neon_vuzp<mode>_insn): Likewise.
1512 (neon_vld1<mode>): Likewise.
1513 (neon_vld1_lane<mode>): Likewise.
1514 (neon_vld1_lane<mode>): Likewise.
1515 (neon_vld1_dup<mode>): Likewise.
1516 (neon_vld1_dup<mode>): Likewise.
1517 (neon_vld1_dupv2di): Likewise.
1518 (neon_vst1<mode>): Likewise.
1519 (neon_vst1_lane<mode>): Likewise.
1520 (neon_vst1_lane<mode>): Likewise.
1521 (neon_vld2<mode>): Likewise.
1522 (neon_vld2<mode>): Likewise.
1523 (neon_vld2_lane<mode>): Likewise.
1524 (neon_vld2_lane<mode>): Likewise.
1525 (neon_vld2_dup<mode>): Likewise.
1526 (neon_vst2<mode>): Likewise.
1527 (neon_vst2<mode>): Likewise.
1528 (neon_vst2_lane<mode>): Likewise.
1529 (neon_vst2_lane<mode>): Likewise.
1530 (neon_vld3<mode>): Likewise.
1531 (neon_vld3qa<mode>): Likewise.
1532 (neon_vld3qb<mode>): Likewise.
1533 (neon_vld3_lane<mode>): Likewise.
1534 (neon_vld3_lane<mode>): Likewise.
1535 (neon_vld3_dup<mode>): Likewise.
1536 (neon_vst3<mode>): Likewise.
1537 (neon_vst3qa<mode>): Likewise.
1538 (neon_vst3qb<mode>): Likewise.
1539 (neon_vst3_lane<mode>): Likewise.
1540 (neon_vst3_lane<mode>): Likewise.
1541 (neon_vld4<mode>): Likewise.
1542 (neon_vld4qa<mode>): Likewise.
1543 (neon_vld4qb<mode>): Likewise.
1544 (neon_vld4_lane<mode>): Likewise.
1545 (neon_vld4_lane<mode>): Likewise.
1546 (neon_vld4_dup<mode>): Likewise.
1547 (neon_vst4<mode>): Likewise.
1548 (neon_vst4qa<mode>): Likewise.
1549 (neon_vst4qb<mode>): Likewise.
1550 (neon_vst4_lane<mode>): Likewise.
1551 (neon_vst4_lane<mode>): Likewise.
1552 (neon_vec_unpack<US>_lo_<mode>): Likewise.
1553 (neon_vec_unpack<US>_hi_<mode>): Likewise.
1554 (neon_vec_<US>mult_lo_<mode>): Likewise.
1555 (neon_vec_<US>mult_hi_<mode>): Likewise.
1556 (neon_vec_<US>shiftl_<mode>): Likewise.
1557 (neon_unpack<US>_<mode>): Likewise.
1558 (neon_vec_<US>mult_<mode>): Likewise.
1559 (vec_pack_trunc_<mode>): Likewise.
1560 (neon_vec_pack_trunc_<mode>): Likewise.
1561 (neon_vabd<mode>_2): Likewise.
1562 (neon_vabd<mode>_3): Likewise.
1564 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1566 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
1567 (load_pair): Update type attribute.
1568 (store_pair): Update type attribute.
1569 * config/aarch64/iterators.md (q): New.
1571 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1573 * config/arm/types.md: Add new types for Neon insns.
1575 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1576 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1577 Sergey Lega <sergey.s.lega@intel.com>
1578 Anna Tikhonova <anna.tikhonova@intel.com>
1579 Ilya Tocar <ilya.tocar@intel.com>
1580 Andrey Turetskiy <andrey.turetskiy@intel.com>
1581 Ilya Verbin <ilya.verbin@intel.com>
1582 Kirill Yukhin <kirill.yukhin@intel.com>
1583 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1585 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
1586 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
1587 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
1589 (srcp14<mode>): Ditto.
1590 (rsqrt14<mode>): Ditto.
1591 (rsqrt14<mode>): Ditto.
1592 (avx512f_vmscalef<mode>): Ditto.
1593 (avx512f_scalef<mode>): Ditto.
1594 (avx512f_getexp<mode>): Ditto.
1595 (avx512f_sgetexp<mode>): Ditto.
1596 (avx512f_fixupimm<mode>): Ditto.
1597 (avx512f_sfixupimm<mode>): Ditto.
1598 (avx512f_rndscale<mode>): Ditto.
1599 (*avx512er_exp2<mode>): Ditto.
1600 (*avx512er_rcp28<mode>): Ditto.
1601 (avx512er_rsqrt28<mode>): Ditto.
1602 (avx512f_getmant<mode>): Ditto.
1603 (avx512f_getmant<mode>): Ditto.
1604 (avx512f_rndscale<mode>): Fix formatting.
1606 2013-10-15 Martin Jambor <mjambor@suse.cz>
1608 * ipa-utils.h (ipa_edge_within_scc): Declare.
1609 * ipa-cp.c (edge_within_scc): Moved...
1610 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
1612 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1613 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1614 Sergey Lega <sergey.s.lega@intel.com>
1615 Anna Tikhonova <anna.tikhonova@intel.com>
1616 Ilya Tocar <ilya.tocar@intel.com>
1617 Andrey Turetskiy <andrey.turetskiy@intel.com>
1618 Ilya Verbin <ilya.verbin@intel.com>
1619 Kirill Yukhin <kirill.yukhin@intel.com>
1620 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1622 * config/i386/predicates.md (const_8_to_15_operand): New.
1623 (const_16_to_31_operand): Ditto.
1624 * config/i386/sse.md (V8FI): New.
1626 (reduc_splus_v8df): Ditto.
1627 (reduc_splus_v16sf): Ditto.
1628 (avx512f_vextract<shuffletype>32x4_1): Ditto.
1629 (vec_extract_hi_<mode>): Ditto.
1630 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
1631 (vec_set_lo_<mode>): Ditto.
1632 (vec_set_hi_<mode>): Ditto.
1633 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
1634 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
1635 (avx512f_pshufd_1): Ditto.
1636 (avx512f_broadcast<mode>): Ditto.
1637 (avx512f_broadcast<mode>): Ditto.
1638 (define_split): Split vec_extract_lo into move.
1639 (ssequartermode): Ditto.
1640 (ssedoublemode): Extened with wider modes.
1641 (vec_extract_lo_<mode>): Ditto.
1643 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1644 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1645 Sergey Lega <sergey.s.lega@intel.com>
1646 Anna Tikhonova <anna.tikhonova@intel.com>
1647 Ilya Tocar <ilya.tocar@intel.com>
1648 Andrey Turetskiy <andrey.turetskiy@intel.com>
1649 Ilya Verbin <ilya.verbin@intel.com>
1650 Kirill Yukhin <kirill.yukhin@intel.com>
1651 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1653 * config/i386/predicates.md (register_or_constm1_operand): New.
1654 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
1655 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
1656 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
1657 UNSPEC_SCATTER_PREFETCH
1659 (avx512f_ucmp<mode>3): Ditto.
1660 (avx512f_vternlog<mode>): Ditto.
1661 (avx512f_align<mode>): Ditto.
1662 (<shift_insn><mode>3): Ditto.
1663 (avx512f_<rotate>v<mode>): Ditto.
1664 (avx512f_<rotate><mode>): Ditto.
1665 (avx512f_eq<mode>3): Ditto.
1666 (avx512f_eq<mode>3_1): Ditto.
1667 (avx512f_gt<mode>3): Ditto.
1668 (avx512f_testm<mode>3): Ditto.
1669 (avx512f_testnm<mode>3): Ditto.
1670 (avx512pf_gatherpf<mode>): Ditto.
1671 (*avx512pf_gatherpf<mode>_mask): Ditto.
1672 (*avx512pf_gatherpf<mode>): Ditto.
1673 (avx512pf_scatterpf<mode>): Ditto.
1674 (*avx512pf_scatterpf<mode>_mask): Ditto.
1675 (*avx512pf_scatterpf<mode>): Ditto.
1676 (avx512f_vec_dup_gpr<mode>): Ditto.
1677 (clz<mode>2): Ditto.
1678 (conflict<mode>): Ditto.
1679 (REDUC_SMINMAX_MODE): Extened with wider modes.
1680 (reduc_<code>_<mode>): Ditto.
1681 (vlshr<mode>3): Ditto.
1682 (vashl<mode>3): Ditto.
1684 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1685 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1686 Sergey Lega <sergey.s.lega@intel.com>
1687 Anna Tikhonova <anna.tikhonova@intel.com>
1688 Ilya Tocar <ilya.tocar@intel.com>
1689 Andrey Turetskiy <andrey.turetskiy@intel.com>
1690 Ilya Verbin <ilya.verbin@intel.com>
1691 Kirill Yukhin <kirill.yukhin@intel.com>
1692 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1694 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
1697 (avx512fmaskmode): Ditto.
1698 (bcstscalarsuff): Ditto.
1699 (avx512f_blendm<mode>): Ditto.
1700 (cmp_imm_predicate): Ditto.
1701 (avx512f_cmp<mode>3): Ditto.
1702 (avx512f_vec_dup<mode>): Ditto.
1703 (avx512f_vec_dup_mem<mode>): Ditto.
1704 (avx512f_vpermi2var<mode>3): Ditto.
1705 (avx512f_vpermt2var<mode>3): Ditto.
1706 (vec_init<mode>): Ditto.
1707 (avx512f_gathersi<mode>): Ditto.
1708 (*avx512f_gathersi<mode>): Ditto.
1709 (*avx512f_gathersi<mode>_2): Ditto.
1710 (avx512f_gatherdi<mode>): Ditto.
1711 (*avx512f_gatherdi<mode>): Ditto.
1712 (*avx512f_gatherdi<mode>_2): Ditto.
1713 (avx512f_scattersi<mode>): Ditto.
1714 (*avx512f_scattersi<mode>): Ditto.
1715 (avx512f_scatterdi<mode>): Ditto.
1716 (*avx512f_scatterdi<mode>): Ditto.
1717 (sseintprefix): Extened with wider modes.
1718 (VEC_GATHER_IDXSI): Ditto.
1719 (VEC_GATHER_IDXDI): Ditto.
1720 (VEC_GATHER_SRCDI): Ditto.
1722 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
1723 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1725 * config/arm/t-aprofile: New file.
1726 * config.gcc: Handle --with-multilib-list option.
1728 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
1730 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
1733 2013-10-15 Richard Biener <rguenther@suse.de>
1735 * tree-loop-distribution.c (build_empty_rdg): Inline into
1737 (rdg_flag_vertex): Inline into single user.
1738 (rdg_flag_vertex_and_dependent): Likewise.
1739 (build_rdg_partition_for_vertex): Remove processed bitmap.
1740 (rdg_build_partitions): Simplify.
1742 2013-10-15 Richard Biener <rguenther@suse.de>
1744 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
1745 Restructure forwarding through conversions and copies to
1746 avoid performing copy-propagation the wrong way. Adjust
1747 recursion invocations.
1748 (forward_propagate_addr_expr): Add argument stating if we
1749 are recursing from a single-use.
1750 (ssa_forward_propagate_and_combine): Adjust.
1752 2013-10-14 David Malcolm <dmalcolm@redhat.com>
1754 * dumpfile.h (gcc::dump_manager): New class, to hold state
1755 relating to dumpfile management.
1756 (get_dump_file_name): Remove in favor of method of dump_manager.
1757 (dump_initialized_p): Likewise.
1758 (dump_start): Likewise.
1759 (dump_finish): Likewise.
1760 (dump_switch_p): Likewise.
1761 (dump_register): Likewise.
1762 (get_dump_file_info): Likewise.
1763 * context.c (gcc::context::context): Construct the dump_manager
1765 * context.h (gcc::context::get_dumps): New.
1766 (gcc::context::m_dumps): New.
1767 * coverage.c (coverage_init): Port to dump_manager API.
1768 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
1769 (extra_dump_files_in_use): Likewise.
1770 (extra_dump_files_alloced): Likewise.
1771 (gcc::dump_manager::dump_manager): New.
1772 (dump_register): Convert to...
1773 (gcc::dump_manager::dump_register): ...method, replacing
1774 function-static next_dump with m_next_dump field.
1775 (get_dump_file_info): Convert to...
1776 (gcc::dump_manager::get_dump_file_info): ...method.
1777 (get_dump_file_name): Convert to...
1778 (gcc::dump_manager::get_dump_file_name): ...method.
1779 (dump_start): Convert to...
1780 (gcc::dump_manager::dump_start): ...method.
1781 (dump_finish): Convert to...
1782 (gcc::dump_manager::dump_finish): ...method.
1783 (dump_begin): Replace body with...
1784 (gcc::dump_manager::dump_begin): ...new method.
1785 (dump_phase_enabled_p): Convert to...
1786 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1787 (dump_phase_enabled_p): Convert to...
1788 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1789 (dump_initialized_p): Convert to...
1790 (gcc::dump_manager::dump_initialized_p): ...method.
1791 (dump_flag_name): Replace body with...
1792 (gcc::dump_manager::dump_flag_name): ...new method.
1793 (dump_enable_all): Convert to...
1794 (gcc::dump_manager::dump_enable_all): ...new method.
1795 (opt_info_enable_passes): Convert to...
1796 (gcc::dump_manager::opt_info_enable_passes): ...new method.
1797 (dump_switch_p_1): Convert to...
1798 (gcc::dump_manager::dump_switch_p_1): ...new method.
1799 (dump_switch_p): Convert to...
1800 (gcc::dump_manager::dump_switch_p): ...new method.
1801 (opt_info_switch_p): Port to dump_manager API.
1802 (enable_rtl_dump_file): Likewise.
1803 * opts-global.c (handle_common_deferred_options): Port to new
1805 * passes.c (pass_manager::finish_optimization_passes): Likewise.
1806 (pass_manager::register_one_dump_file): Likewise.
1807 (pass_manager::register_pass): Likewise.
1808 (pass_init_dump_file): Likewise.
1809 (pass_fini_dump_file): Likewise.
1810 * statistics.c (statistics_early_init): Likewise.
1812 2013-10-14 Richard Biener <rguenther@suse.de>
1814 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
1815 iterative_hash_canonical_type, gimple_canonical_type_hash,
1816 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
1817 gimple_register_canonical_type, print_gimple_types_stats,
1818 free_gimple_type_tables): Move to lto/lto.c
1819 (gt-gimple.h): Do not include.
1820 * gimple.h (gimple_register_canonical_type,
1821 print_gimple_types_stats, free_gimple_type_tables): Remove.
1822 * Makefile.in (GTFILES): Remove gimple.c.
1824 2013-10-14 Travis Snoozy <quandary@remstate.com>
1827 * config/msp430/msp430.c (msp430_option_override): Correct thinko
1828 scanning for msp430x targets.
1830 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
1833 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
1834 (registers_ok_for_ldd_peep): Move around.
1835 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
1836 * config/sparc/sparc.md (widening peepholes): Use it.
1838 2013-10-14 Richard Biener <rguenther@suse.de>
1842 * gimple.c (iterative_hash_canonical_type): Make sure to
1843 record the hash into the correct hashtable slot.
1845 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
1847 PR rtl-optimization/58662
1848 * combine.c (try_combine): Take into account death nodes on I2 when
1849 splitting a PARALLEL of two independent SETs. Fix dump message.
1851 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
1854 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
1855 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
1856 * config/sh/t-sh (sh_treg_combine.o): New entry.
1857 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
1858 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
1859 (register_sh_passes): New function. Register sh_treg_combine pass.
1860 (sh_option_override): Invoke it.
1861 (sh_canonicalize_comparison): Handle op0_preserve_value.
1862 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
1863 opportunities. Canonicalize branch condition.
1864 (nott): Allow only if pseudos can be created for non-SH2A.
1866 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
1869 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
1870 (ix86_expand_movmem): Replace copy_addr_to_reg with
1871 ix86_copy_addr_to_reg.
1872 (ix86_expand_setmem): Likewise.
1874 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
1876 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
1877 provided by ix86_fp_compare_mode instead of CCFPUmode.
1879 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
1881 * config/aarch64/arm_neon.h
1882 (vtbx<1,3>_<psu>8): Fix register constriants.
1884 2013-10-11 Jeff Law <law@redhat.com>
1886 PR tree-optimization/58640
1887 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
1888 threading paths that cross over two loop entry points.
1890 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1892 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
1893 handle vector float as well.
1894 (*vsx_le_perm_load_v4si): Likewise.
1895 (*vsx_le_perm_store_v2di): Likewise.
1896 (*vsx_le_perm_store_v4si): Likewise.
1898 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1900 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
1901 directly to circumvent subtract from splat{31} workaround.
1902 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
1904 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
1905 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
1906 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
1907 (altivec_vperm_<mode>): Convert to define_insn_and_split to
1908 separate big and little endian logic.
1909 (*altivec_vperm_<mode>_internal): New define_insn.
1910 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
1911 separate big and little endian logic.
1912 (*altivec_vperm_<mode>_uns_internal): New define_insn.
1913 (vec_permv16qi): Add little endian logic.
1915 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
1917 * doc/extend.texi (returns_nonnull): Remove arguments.
1919 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1920 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1921 Sergey Lega <sergey.s.lega@intel.com>
1922 Anna Tikhonova <anna.tikhonova@intel.com>
1923 Ilya Tocar <ilya.tocar@intel.com>
1924 Andrey Turetskiy <andrey.turetskiy@intel.com>
1925 Ilya Verbin <ilya.verbin@intel.com>
1926 Kirill Yukhin <kirill.yukhin@intel.com>
1927 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1929 * config/i386/sse.md (VI48F_256_512): New.
1930 (avx2_permvar<mode>): Change to ...
1931 (<avx2_avx512f>_permvar<mode>): This.
1933 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1934 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1935 Sergey Lega <sergey.s.lega@intel.com>
1936 Anna Tikhonova <anna.tikhonova@intel.com>
1937 Ilya Tocar <ilya.tocar@intel.com>
1938 Andrey Turetskiy <andrey.turetskiy@intel.com>
1939 Ilya Verbin <ilya.verbin@intel.com>
1940 Kirill Yukhin <kirill.yukhin@intel.com>
1941 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1943 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
1944 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
1945 * config/i386/sse.md (VI4_AVX): New.
1946 (sf2simodelower): Ditto.
1947 (sse2_cvtps2dq): Change to ...
1948 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
1950 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1951 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1952 Sergey Lega <sergey.s.lega@intel.com>
1953 Anna Tikhonova <anna.tikhonova@intel.com>
1954 Ilya Tocar <ilya.tocar@intel.com>
1955 Andrey Turetskiy <andrey.turetskiy@intel.com>
1956 Ilya Verbin <ilya.verbin@intel.com>
1957 Kirill Yukhin <kirill.yukhin@intel.com>
1958 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1960 * config/i386/sse.md (V_512): New.
1962 (vcond<V_512:mode><VF_512:mode>): Ditto.
1963 (vcond<V_512:mode><VI_512:mode>): Ditto.
1964 (vcondu<V_512:mode><VI_512:mode>): Ditto.
1966 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1967 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1968 Sergey Lega <sergey.s.lega@intel.com>
1969 Anna Tikhonova <anna.tikhonova@intel.com>
1970 Ilya Tocar <ilya.tocar@intel.com>
1971 Andrey Turetskiy <andrey.turetskiy@intel.com>
1972 Ilya Verbin <ilya.verbin@intel.com>
1973 Kirill Yukhin <kirill.yukhin@intel.com>
1974 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1976 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
1977 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
1979 (fma<mode>4): Removed condition.
1980 (fms<mode>4): Ditto.
1981 (fnma<mode>4): Ditto.
1982 (fnms<mode>4): Ditto.
1983 (fma4i_fmadd_<mode>): Ditto.
1984 (*fma_fmadd_<mode>): Ditto.
1985 (*fma_fmsub_<mode>): Ditto.
1986 (*fma_fnmadd_<mode>): Ditto.
1987 (*fma_fnmsub_<mode>): Ditto.
1988 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
1989 (*fma_fmaddsub_<mode>): Ditto.
1990 (*fma_fmsubadd_<mode>): Ditto.
1991 (*fmai_fmadd_<mode>): Ditto.
1992 (*fmai_fmsub_<mode>): Ditto.
1993 (*fmai_fnmadd_<mode>): Ditto.
1994 (*fmai_fnmsub_<mode>): Ditto.
1996 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1997 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1998 Sergey Lega <sergey.s.lega@intel.com>
1999 Anna Tikhonova <anna.tikhonova@intel.com>
2000 Ilya Tocar <ilya.tocar@intel.com>
2001 Andrey Turetskiy <andrey.turetskiy@intel.com>
2002 Ilya Verbin <ilya.verbin@intel.com>
2003 Kirill Yukhin <kirill.yukhin@intel.com>
2004 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2006 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
2007 (VI124_256): Changed to ...
2008 (VI124_256_48_512): This.
2009 (ssepackmode): Extended with wider modes.
2010 (<code><mode>3): Changed iterator.
2011 (*avx2_<code><mode>3): Ditto.
2012 (vec_pack_trunc_<mode>): Ditto.
2014 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2015 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2016 Sergey Lega <sergey.s.lega@intel.com>
2017 Anna Tikhonova <anna.tikhonova@intel.com>
2018 Ilya Tocar <ilya.tocar@intel.com>
2019 Andrey Turetskiy <andrey.turetskiy@intel.com>
2020 Ilya Verbin <ilya.verbin@intel.com>
2021 Kirill Yukhin <kirill.yukhin@intel.com>
2022 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2024 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
2025 (VI8F_256_512): Ditto.
2026 (abs<mode>2): Changed iterator.
2027 (avx2_perm<mode>): Changed to ...
2028 (<avx2_avx512f>_perm<mode>): This.
2029 (avx2_perm<mode>_1): Changed to ...
2030 (<avx2_avx512f>_perm<mode>_1): This.
2032 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2033 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2034 Sergey Lega <sergey.s.lega@intel.com>
2035 Anna Tikhonova <anna.tikhonova@intel.com>
2036 Ilya Tocar <ilya.tocar@intel.com>
2037 Andrey Turetskiy <andrey.turetskiy@intel.com>
2038 Ilya Verbin <ilya.verbin@intel.com>
2039 Kirill Yukhin <kirill.yukhin@intel.com>
2040 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2042 * config/i386/sse.md (VI48_AVX512F): New.
2043 (VI48_AVX2): Changed to ...
2044 (VI48_AVX2_48_AVX512F): This.
2045 (avx2_ashrv<mode>): Changed to ...
2046 (<avx2_avx512f>_ashrv<mode>): This.
2047 (avx2_<shift_insn>v<mode>): Changed to ...
2048 (<avx2_avx512f>_<shift_insn>v<mode>): This.
2050 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2051 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2052 Sergey Lega <sergey.s.lega@intel.com>
2053 Anna Tikhonova <anna.tikhonova@intel.com>
2054 Ilya Tocar <ilya.tocar@intel.com>
2055 Andrey Turetskiy <andrey.turetskiy@intel.com>
2056 Ilya Verbin <ilya.verbin@intel.com>
2057 Kirill Yukhin <kirill.yukhin@intel.com>
2058 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2060 * config/i386/sse.md (VI4_AVX512F): New.
2061 (VI8_AVX2_AVX512F): Ditto.
2062 (mul<mode>3): Extended with wider modes.
2063 (*<sse4_1_avx2>_mul<mode>3): Ditto.
2064 (mul<mode>3): Ditto.
2065 (vec_widen_<s>mult_odd_<mode>): Ditto.
2067 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2068 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2069 Sergey Lega <sergey.s.lega@intel.com>
2070 Anna Tikhonova <anna.tikhonova@intel.com>
2071 Ilya Tocar <ilya.tocar@intel.com>
2072 Andrey Turetskiy <andrey.turetskiy@intel.com>
2073 Ilya Verbin <ilya.verbin@intel.com>
2074 Kirill Yukhin <kirill.yukhin@intel.com>
2075 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2077 * config/i386/sse.md (VI2_AVX512F): New.
2078 (VI124_AVX512F): Ditto.
2079 (sseunpackmode): Extended with wider modes.
2080 (sseunpackfltmode): Ditto.
2081 (vec_unpacks_float_hi_<mode>): Ditto.
2082 (vec_unpacks_float_lo_<mode>): Ditto.
2083 (vec_unpacku_float_hi_<mode>): Ditto.
2084 (vec_unpacku_float_lo_<mode>): Ditto.
2085 (vec_unpacks_lo_<mode>): Ditto.
2086 (vec_unpacks_hi_<mode>): Ditto.
2087 (vec_unpacku_lo_<mode>): Ditto.
2088 (vec_unpacku_hi_<mode>): Ditto.
2090 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2091 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2092 Sergey Lega <sergey.s.lega@intel.com>
2093 Anna Tikhonova <anna.tikhonova@intel.com>
2094 Ilya Tocar <ilya.tocar@intel.com>
2095 Andrey Turetskiy <andrey.turetskiy@intel.com>
2096 Ilya Verbin <ilya.verbin@intel.com>
2097 Kirill Yukhin <kirill.yukhin@intel.com>
2098 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2100 * config/i386/i386.md (multdiv): New.
2101 (multdiv_mnemonic): Ditto.
2102 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
2103 (<sse>_vm<multdiv_mnemonic><mode>3): This.
2104 (<sse>_vmdiv<mode>3): Removed.
2106 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2107 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2108 Sergey Lega <sergey.s.lega@intel.com>
2109 Anna Tikhonova <anna.tikhonova@intel.com>
2110 Ilya Tocar <ilya.tocar@intel.com>
2111 Andrey Turetskiy <andrey.turetskiy@intel.com>
2112 Ilya Verbin <ilya.verbin@intel.com>
2113 Kirill Yukhin <kirill.yukhin@intel.com>
2114 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2116 * config/i386/sse.md (V): Extended with wider modes.
2118 (ssehalfvecmode): Ditto.
2120 (ssepackfltmode): Ditto.
2121 (avx_vec_concat<mode>): Ditto.
2122 (V_256_512): New iterator.
2123 (VF2_512_256): Ditto.
2124 (si2dfmode): New attribute.
2125 (si2dfmodelower): Ditto.
2127 (concat_tg_mode): Ditto.
2128 (floatv4siv4df2): Changed to ...
2129 (float<si2dfmodelower><mode>2): This.
2130 (avx_cvtps2pd256): Changed to ...
2131 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
2132 (vec_pack_trunc_v4df): Changed to ...
2133 (vec_pack_trunc_<mode>): This.
2134 (avx_vpermil<mode>): Changed to ...
2135 (<sse2_avx_avx512f>_vpermil<mode>): This.
2136 (<fixsuffix>fix_truncv8dfv8si2): New.
2137 (vec_pack_sfix_trunc_v8df): Ditto.
2138 (avx512f_rndscale<mode>): Ditto.
2139 (avx512f_roundpd512): Ditto.
2140 (vec_pack_ufix_trunc_<mode>): Updated iterator.
2142 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2143 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2144 Sergey Lega <sergey.s.lega@intel.com>
2145 Anna Tikhonova <anna.tikhonova@intel.com>
2146 Ilya Tocar <ilya.tocar@intel.com>
2147 Andrey Turetskiy <andrey.turetskiy@intel.com>
2148 Ilya Verbin <ilya.verbin@intel.com>
2149 Kirill Yukhin <kirill.yukhin@intel.com>
2150 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2152 * config/i386/i386.md (any_fix): New iterator.
2153 (fixsuffix): New attribute.
2154 * config/i386/sse.md (VF1): Extened with wider modes.
2158 (sseintvecmodelower): Ditto.
2159 (ssescalarmode): Ditto.
2160 (ssescalarnum): Ditto.
2162 (ssexmmmode): Ditto.
2163 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
2164 (<sse>_rcp<mode>2): Change iterator.
2165 (rsqrt<mode>2): Ditto.
2166 (<sse>_rsqrt<mode>2): Ditto.
2167 (avx2_vec_dup<mode>): Ditto.
2168 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
2169 (round<mode>2_sfix): Ditto.
2170 (avx2_pbroadcast<mode>): Ditto.
2171 (*andnot<mode>3): Handle XI mode.
2172 (*<code><mode>3): Ditto.
2173 (AVXTOSSEMODE): Removed.
2174 (avx_vpermil<mode>): Changed to ...
2175 (<sse2_avx_avx512f>_vpermil<mode>): This.
2177 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2178 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2179 Sergey Lega <sergey.s.lega@intel.com>
2180 Anna Tikhonova <anna.tikhonova@intel.com>
2181 Ilya Tocar <ilya.tocar@intel.com>
2182 Andrey Turetskiy <andrey.turetskiy@intel.com>
2183 Ilya Verbin <ilya.verbin@intel.com>
2184 Kirill Yukhin <kirill.yukhin@intel.com>
2185 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2187 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
2188 (<sse>_comi): Ditto.
2189 (<sse>_ucomi): Ditto.
2190 (sse_cvtss2siq_2): Ditto.
2191 (sse2_cvtsd2si): Ditto.
2192 (sse2_cvtsd2siq): Ditto.
2193 (sse2_cvttsd2si): Ditto.
2194 (sse2_cvttsd2siq): Ditto.
2195 (<shift_insn><mode>3): Ditto.
2196 (sse2_cvtsi2sdq): Update constraint and prefix.
2197 (sse_cvtsi2ss): Update prefix.
2198 (sse_cvtsi2ssq): Ditto.
2200 2013-10-11 Jakub Jelinek <jakub@redhat.com>
2202 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
2203 ignore internal calls.
2205 2013-10-11 Richard Biener <rguenther@suse.de>
2207 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
2208 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
2209 TREE_OVERFLOW is set.
2211 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
2213 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
2215 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
2216 * gimple.h: Likewise.
2218 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
2220 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
2223 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
2225 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
2226 * configure: Regenerate.
2228 2013-10-11 Jakub Jelinek <jakub@redhat.com>
2230 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
2231 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
2232 formatting fixes, use pp_colon instead of pp_character (..., ':'),
2233 similarly pp_right_paren.
2234 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
2235 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
2236 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
2237 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
2239 (omp_declare_simd_clauses_equal,
2240 omp_remove_redundant_declare_simd_attrs): New functions.
2241 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
2242 (walk_tree_1): Handle new OpenMP 4.0 clauses.
2243 * tree.h (OMP_LOOP_CHECK): Define.
2244 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
2245 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
2246 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
2247 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
2248 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
2249 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
2250 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
2251 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
2252 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
2253 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
2254 OMP_CLAUSE_SIMDLEN_EXPR): Define.
2255 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
2256 (omp_remove_redundant_declare_simd_attrs): New prototype.
2257 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
2258 GIMPLE_OMP_TEAMS): New codes.
2259 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
2260 * omp-low.c (struct omp_context): Add cancel_label and cancellable
2262 (target_nesting_level): New variable.
2263 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
2264 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
2265 for collapse > 1 static schedule unless ordered.
2266 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
2267 (determine_parallel_type): Adjust get_ws_args_for caller.
2268 (install_var_field): Handle mask & 4 for double indirection.
2269 (scan_sharing_clauses): Ignore shared clause on teams construct.
2270 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
2271 (create_omp_child_function): If inside target or declare target
2272 constructs, set "omp declare target" attribute on the child function.
2273 (find_combined_for): New function.
2274 (scan_omp_parallel): Handle combined loops.
2275 (scan_omp_target, scan_omp_teams): New functions.
2276 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
2277 restrictions and set ctx->cancellable for cancellable constructs.
2278 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
2279 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
2280 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
2281 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
2282 (omp_clause_aligned_alignment): New function.
2283 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
2284 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
2285 on teams constructs. Handle user defined reductions and new
2287 (lower_reduction_clauses): Don't set placeholder to address of ref
2288 if it has already the right type.
2289 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
2290 (expand_parallel_call): Use the new non-_start suffixed builtins,
2291 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
2292 and GOMP_parallel_end after the call.
2293 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
2294 (expand_omp_for_init_counts): Handle combined loops.
2295 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
2297 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
2298 end of cancellable loops.
2299 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
2300 Likewise. Handle collapse > 1 loops.
2301 (expand_omp_simd): Handle combined loops.
2302 (expand_omp_for): Add inner_stmt argument, adjust callers of
2303 expand_omp_for* functions, use expand_omp_for_static*chunk even
2304 for collapse > 1 unless ordered.
2305 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
2306 of cancellable sections.
2307 (expand_omp_single): Remove need_barrier variable, just rely on
2308 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
2309 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
2310 (expand_omp_atomic_load, expand_omp_atomic_store,
2311 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
2312 (expand_omp_target): New function.
2313 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
2314 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
2315 (build_omp_regions_1): Immediately close region for
2316 GF_OMP_TARGET_KIND_UPDATE.
2317 (maybe_add_implicit_barrier_cancel): New function.
2318 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
2320 (lower_omp_single): Likewise. Add clobber after the barrier.
2321 (lower_omp_taskgroup): New function.
2322 (lower_omp_for): Handle combined loops. Adjust
2323 lower_rec_input_clauses caller. Handle cancellation.
2324 (lower_depend_clauses): New function.
2325 (lower_omp_taskreg): Lower depend clauses. Adjust
2326 lower_rec_input_clauses caller. Add clobber after the call. Handle
2328 (lower_omp_target, lower_omp_teams): New functions.
2329 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
2330 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
2331 and GOMP_cancellation_point calls.
2332 (lower_omp): Fold stmts inside of target region.
2333 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
2334 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2335 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
2336 (BT_FN_VOID_OMPFN_PTR_UINT,
2337 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
2338 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
2339 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
2340 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
2341 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
2342 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
2343 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
2344 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
2345 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
2346 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
2347 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
2348 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
2349 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
2350 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
2351 BUILT_IN_GOMP_PARALLEL_END.
2352 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
2353 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2354 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
2355 GF_OMP_FOR_KIND_DISTRIBUTE.
2356 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
2357 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
2358 (dump_gimple_omp_return): Print lhs if it has any.
2359 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
2360 gimple_omp_atomic_seq_cst_p.
2361 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
2362 and GIMPLE_OMP_TEAMS.
2363 * langhooks.c (lhd_omp_mappable_type): New function.
2364 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
2365 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
2367 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
2368 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
2369 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
2370 (struct gimplify_omp_ctx): Add combined_loop field.
2371 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
2372 on stmts inside of target region.
2373 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
2374 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
2375 ORT_TARGET and ORT_TARGET_DATA.
2376 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
2378 (omp_notice_threadprivate_variable): Complain about threadprivate
2379 variables in target region.
2380 (omp_notice_variable): Complain about vars with non-mappable type
2381 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
2382 (omp_check_private): Ignore ORT_TARGET* regions.
2383 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
2384 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
2385 (find_combined_omp_for): New function.
2386 (gimplify_omp_for): Handle gimplification of combined loops.
2387 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
2389 (gimplify_omp_target_update): New function.
2390 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
2391 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
2392 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
2393 (gimplify_body): If fndecl has "omp declare target" attribute, add
2394 implicit ORT_TARGET context around it.
2395 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
2396 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
2397 * tree-nested.c (convert_nonlocal_reference_stmt,
2398 convert_local_reference_stmt, convert_gimple_call): Handle
2399 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2400 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
2401 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
2402 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
2403 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
2404 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
2405 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
2406 BUILT_IN_GOMP_LOOP_END_CANCEL,
2407 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
2408 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
2409 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
2410 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
2411 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
2412 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
2413 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
2414 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
2415 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
2416 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
2417 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
2418 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
2419 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
2420 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
2421 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
2422 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2423 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
2424 gimple_build_omp_teams): New functions.
2425 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
2426 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
2427 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
2428 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2429 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
2430 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
2431 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
2432 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
2433 GF_OMP_ATOMIC_SEQ_CST): New.
2434 (gimple_build_omp_taskgroup, gimple_build_omp_target,
2435 gimple_build_omp_teams): New prototypes.
2436 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
2437 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2438 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
2439 GIMPLE_OMP_SINGLE as end of range.
2440 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
2441 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
2442 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
2443 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
2444 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
2445 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
2446 gimple_omp_target_kind, gimple_omp_target_set_kind,
2447 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
2448 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
2449 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
2450 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
2451 gimple_omp_teams_set_clauses): New inlines.
2452 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
2453 and GIMPLE_OMP_TASKGROUP.
2454 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
2455 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
2456 enum omp_clause_proc_bind_kind): New.
2457 (union omp_clause_subcode): Add depend_kind, map_kind and
2458 proc_bind_kind fields.
2459 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
2460 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2461 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
2462 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
2463 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
2465 2013-10-10 Teresa Johnson <tejohnson@google.com>
2467 * predict.c (tree_estimate_probability): Add new parameter
2468 for estimate_bb_frequencies.
2469 (estimate_bb_frequencies): Add new parameter to force estimation.
2470 (rebuild_frequencies): When max frequency in function is small,
2471 recompute counts from frequencies.
2472 * predict.h (estimate_bb_frequencies): New parameter.
2474 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2476 * ipa-inline.c (ipa_inline): Fix leak of "order" when
2477 optimizations are disabled.
2479 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2481 * coverage.c (coverage_finish): Fix leak of da_file_name.
2483 2013-10-10 Jan Hubicka <jh@suse.cz>
2485 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
2486 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
2489 2013-10-10 Jakub Jelinek <jakub@redhat.com>
2492 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
2493 if any labels are in FALLTHRU_BB, use a special label emitted
2494 immediately after the asm goto insn rather than label_rtx
2496 (expand_asm_stmt): Adjust caller.
2497 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
2498 edge if the last insn in predecessor is a jump with single successor,
2499 but it isn't simplejump_p.
2501 2013-10-10 Richard Biener <rguenther@suse.de>
2503 PR tree-optimization/58656
2504 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
2506 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2508 * gimplify.c: Include expr.h and tm_p.h for targets with special
2509 va-arg padding requirements.
2511 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2513 * tree-flow.h: Move some prototypes to gimple.h.
2514 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
2515 * gimple.h: Relocate some prototypes from tree-flow.h
2516 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
2518 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
2519 (build_va_arg_indirect_ref): Relocate and make static.
2520 (std_gimplify_va_arg_expr): Relocate here.
2521 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
2522 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
2524 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2526 * doc/md.texi: Document the mnemonic attribute.
2528 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2531 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
2532 (set_attr_alternative x ...) when searching for user defined
2535 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2537 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
2539 2013-10-09 Easwaran Raman <eraman@google.com>
2541 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
2542 * cfgexpand.c (defer_stack_allocation): ...use here
2543 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
2545 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
2547 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
2548 (operand_equal_for_value_replacement): New function, extracted from
2549 value_replacement and enhanced to catch more cases.
2550 (value_replacement): Use operand_equal_for_value_replacement.
2552 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2554 * loop-doloop.c (doloop_modify, doloop_optimize): Use
2555 get_max_loop_iterations.
2557 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2559 * config/arm/aarch-common.c (arm_early_load_addr_dep):
2560 Place comment above function.
2562 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2564 * tree-flow.h: Remove all remaining prototypes, enums and structs that
2565 are not related to tree-cfg.c.
2566 * tree-ssa-address.h: New file. Relocate prototypes.
2567 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
2568 (addr_for_mem_ref): New. Combine call to get_address_description and
2569 return addr_for_mem_ref.
2570 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
2571 * tree-ssa-live.h: Adjust prototypes.
2572 * passes.c: Include tree-ssa-live.h.
2573 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
2574 * graphite.c (graphite_transform_loops): Make static.
2575 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2576 make_pass_graphite, pass_data_graphite_transforms,
2577 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
2578 * ipa-pure-const.c (warn_function_noreturn): Make static.
2579 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2580 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2581 Relocate from tree-cfg.c
2582 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
2584 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2585 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2586 Move to ipa-pure-const.c.
2587 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
2588 Relocate from tree-optimize.c.
2589 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
2590 make_pass_fixup_cfg): Move to tree-cfg.c.
2591 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
2592 Relocate some prototypes.
2593 * tree-data-ref.h (tree_check_data_deps) Add prototype.
2594 * tree-dump.c (dump_function_to_file): Remove prototype.
2595 Add tree-flow.h to the include file.
2596 * tree-dump.h: Remove prototype.
2597 * tree-parloops.h: New File. Add prototypes.
2598 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
2599 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
2600 from tree-ssa-loop.c.
2601 * tree-predcom.c (run_tree_predictive_commoning,
2602 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
2603 Relocate here from tree-ssa-loop.c.
2604 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
2605 ssa_name_values.release ().
2606 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
2607 (ssa_name_values): Relocate from tree-flow.h.
2608 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
2609 * tree-ssa-loop.c (run_tree_predictive_commoning,
2610 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
2611 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2612 make_pass_graphite, pass_data_graphite_transforms,
2613 make_pass_graphite_transforms, gate_tree_parallelize_loops,
2614 tree_parallelize_loops, pass_data_parallelize_loops,
2615 make_pass_parallelize_loops): Move to other files.
2616 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
2618 * tree.h: Remove prototypes from tree-address.c.
2620 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2622 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
2623 (struct int_tree_map): Move to tree-hasher.h
2624 (SCALE, LABEL, PERCENT): Move to gimple.h
2625 * tree-flow-inline.h: Delete. Move functions to other files.
2626 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
2627 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
2628 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
2629 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
2630 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
2631 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
2632 * tree-hasher.h: Don't include tree-flow.h.
2633 (struct int_tree_map): Relocate from tree-flow.h.
2634 * tree-sra.c (contains_view_convert_expr_p): Relocate from
2635 tree-flow-inline.h and make static.
2636 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
2638 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
2639 tree-flow-inline.h and make static.
2640 * tree.h (is_global_var, may_be_aliased): Relocate from
2642 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
2643 * value-prof.c: No longer include tree-flow-inline.h.
2644 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
2646 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2648 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
2649 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
2650 (enum move_pos): Move to tree-ssa-loop-im.h
2651 * cfgloop.h: Move some prototypes.
2652 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
2653 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
2654 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
2655 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
2656 (loop_containing_stmt): Relocate from tree-flow-inline.h.
2657 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
2658 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
2659 (enum move_pos): Relocate here.
2660 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
2662 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
2663 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
2664 make_pass_lim): Relocate here from tree-ssa-loop.c.
2665 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
2667 (loop_edge_to_cancel, unloop_loops): Make static.
2668 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2669 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
2670 (tree_complete_unroll, gate_tree_complete_unroll,
2671 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
2672 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2673 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
2674 * tree-ssa-loop-ivopts.c: Remove local prototypes.
2675 (stmt_invariant_in_loop_p): Remove unused function.
2676 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
2677 * tree-ssa-loop-manip.h: New file. Add prototypes.
2678 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
2679 (gcov_type_to_double_int): Move to cfgloop.h.
2680 (double_int_cmp, bound_index,
2681 estimate_numbers_of_iterations_loop): Make static.
2682 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
2683 (max_loop_iterations): Factor out get_max_loop_iterations.
2684 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
2686 * tree-ssa-loop-niter.h: New file. Add prototypes.
2687 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
2688 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
2689 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
2690 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
2691 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
2692 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
2693 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
2694 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
2695 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
2696 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
2697 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2698 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
2699 pass_data_complete_unroll, make_pass_complete_unroll,
2700 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2701 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
2702 tree-ssa-loop-ivcanon.c.
2703 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
2704 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
2705 tree-ssa-loop-prefetch.c.
2706 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
2708 (get_lsm_tmp_name): Relocate and add suffix parameter.
2709 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
2710 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
2711 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
2712 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
2713 (get_estimated_loop_iterations): Factor out accessor from
2714 estimated_loop_iterations in tree-ssa-loop-niter.c.
2715 (get_max_loop_iterations): Factor out accessor from
2716 _max_loop_iterations in tree-ssa-niter.c.
2717 * loop-unroll.c (decide_unroll_constant_iterations,
2718 decide_unroll_runtime_iterations, decide_peel_simple,
2719 decide_unroll_stupid): Use new get_* accessors.
2721 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
2723 PR tree-optimization/20318
2724 * doc/extend.texi (returns_nonnull): New function attribute.
2725 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
2727 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
2728 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
2730 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
2733 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
2734 false if both components are bitfields.
2736 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2738 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
2739 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
2740 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
2741 * config/aarch64/aarch64.h
2742 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
2743 * config/aarch64/aarch64-simd-builtins.def
2744 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
2746 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2748 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
2750 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2752 * config/aarch64/arm_neon.h (vdiv_f64): Added.
2754 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2756 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
2757 (vneg_f64): New intrinsic.
2758 (vneg_s8): Asm replaced with C.
2759 (vneg_s16): Likewise.
2760 (vneg_s32): Likewise.
2761 (vneg_s64): New intrinsic.
2762 (vnegq_f32): Asm replaced with C.
2763 (vnegq_f64): Likewise.
2764 (vnegq_s8): Likewise.
2765 (vnegq_s16): Likewise.
2766 (vnegq_s32): Likewise.
2767 (vnegq_s64): Likewise.
2769 2013-10-09 Renlin Li <Renlin.Li@arm.com>
2771 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
2773 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2775 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
2776 packed stack special handling.
2777 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
2778 back to fixed stack slots for FPRs saved due to stdarg.
2780 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2782 * config/s390/s390.c (s390_frame_info): Restructure function.
2784 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2786 * config/s390/s390.c (struct s390_frame_layout): New field
2788 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
2789 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
2790 regs_ever_clobbered to char*.
2791 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
2792 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
2793 (s390_register_info): Enable FPR save slots. Move/Copy some
2794 functionality into ...
2795 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
2796 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
2798 (s390_frame_info): Do gpr slot allocation here now. stdarg does
2799 not imply a stack frame.
2800 (s390_init_frame_layout): Remove variable clobbered_regs.
2801 (s390_update_register_info): Remove function.
2802 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
2803 cannot be used for register renaming.
2804 (s390_hard_regno_scratch_ok): New function.
2805 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
2806 (s390_initial_elimination_offset): Change offset calculation of
2807 the return address pointer.
2808 (save_gprs): Deal with only r6 being saved from the call-saved regs.
2809 (restore_gprs): Set frame related flag.
2810 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
2811 (s390_emit_prologue): Call s390_register_info instead of
2812 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
2813 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
2814 (s390_optimize_prologue): Call s390_optimize_register_info.
2815 Try to remove also FPR slot save/restore INSNs. Remove frame
2816 related flags from restore INSNs.
2818 2013-10-08 DJ Delorie <dj@redhat.com>
2820 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
2823 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
2824 avoid conflict with the MI use of %c.
2825 * config/rl78/rl78-real.md: change %c to %C throughout.
2826 * config/rl78/rl78-virt.md: Likewise.
2828 2013-10-08 Jan Hubicka <jh@suse.cz>
2830 * config/i386/i386.c (ix86_option_override_internal): Switch
2831 to SSE math for -ffast-math when target ISA supports SSE2.
2833 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
2835 * tree-flow.h: Remove some prototypes.
2836 * tree.h: Remove some protypes, add a couple.
2837 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
2838 using_eh_for_cleanups_p): Add interface routines for front ends.
2839 * tree-eh.h: New file. Add protoptyes.
2840 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
2841 (add_stmt_to_eh_lp_fn): Make static.
2842 (lower_try_finally): Use new using_eh_for_cleanups_p.
2843 * emit-rtl.c: Include tree-eh.h.
2844 * gimple.h: Include tree-eh.h.
2846 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
2848 PR tree-optimization/58480
2849 * tree-vrp.c (infer_nonnull_range): New function.
2850 (infer_value_range): Call infer_nonnull_range.
2852 2013-10-08 Dehao Chen <dehao@google.com>
2854 PR tree-optimization/58619
2855 * tree-inline.c (copy_phis_for_bb): Combine location data
2858 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2861 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
2862 RTX_FRAME_RELATED_P on INSN.
2864 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2866 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
2867 (altivec_expand_vec_perm_const): Call it.
2869 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2871 * config/rs6000/vector.md (mov<mode>): Emit permuted move
2872 sequences for LE VSX loads and stores at expand time.
2873 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
2875 * config/rs6000/rs6000.c (rs6000_const_vec): New.
2876 (rs6000_gen_le_vsx_permute): New.
2877 (rs6000_gen_le_vsx_load): New.
2878 (rs6000_gen_le_vsx_store): New.
2879 (rs6000_gen_le_vsx_move): New.
2880 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
2881 (*vsx_le_perm_load_v4si): New.
2882 (*vsx_le_perm_load_v8hi): New.
2883 (*vsx_le_perm_load_v16qi): New.
2884 (*vsx_le_perm_store_v2di): New.
2885 (*vsx_le_perm_store_v4si): New.
2886 (*vsx_le_perm_store_v8hi): New.
2887 (*vsx_le_perm_store_v16qi): New.
2888 (*vsx_xxpermdi2_le_<mode>): New.
2889 (*vsx_xxpermdi4_le_<mode>): New.
2890 (*vsx_xxpermdi8_le_V8HI): New.
2891 (*vsx_xxpermdi16_le_V16QI): New.
2892 (*vsx_lxvd2x2_le_<mode>): New.
2893 (*vsx_lxvd2x4_le_<mode>): New.
2894 (*vsx_lxvd2x8_le_V8HI): New.
2895 (*vsx_lxvd2x16_le_V16QI): New.
2896 (*vsx_stxvd2x2_le_<mode>): New.
2897 (*vsx_stxvd2x4_le_<mode>): New.
2898 (*vsx_stxvd2x8_le_V8HI): New.
2899 (*vsx_stxvd2x16_le_V16QI): New.
2901 2013-10-07 Renlin Li <Renlin.Li@arm.com>
2903 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
2905 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2907 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
2908 loop to work also for 31bit ABI.
2909 Save the stack pointer for frame_size > 0.
2911 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2913 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
2914 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
2915 constraint letters from expanders.
2916 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
2917 retry count to general_operand.
2918 ("tabort"): Give operand 0 a mode.
2919 ("tabort_1"): Add mode and constraint letter for operand 0.
2920 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
2922 2013-10-04 Jeff Law <law@redhat.com>
2924 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
2926 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
2928 (thread_across_edge): Here. Call it.
2930 2013-10-04 Cary Coutant <ccoutant@google.com>
2932 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
2933 discarding a location list expression (or a piece of one).
2935 2013-10-03 Jan Hubicka <jh@suse.cz>
2937 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
2938 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
2939 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
2941 2013-10-03 Jan Hubicka <jh@suse.cz>
2943 * config/i386/i386.c (ix86_option_override_internal): Do not enable
2944 accumulate-outgoing-args when producing unwind info.
2946 2013-10-03 Wei Mi <wmi@google.com>
2948 * lra-constraints.c (insert_move_for_subreg): New function
2949 extracted from simplify_operand_subreg.
2950 (simplify_operand_subreg): Add reload for paradoxical subreg.
2952 2013-10-03 Rong Xu <xur@google.com>
2954 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
2955 the candidate of builtin_expect such that we should fix the
2956 size/time estimation.
2957 (estimate_function_body_sizes): Do the acutally size/time fix-up
2960 2013-10-03 Rong Xu <xur@google.com>
2962 * predict.c (tree_predict_by_opcode): Get the probability
2963 for builtin_expect from param builtin_expect_probability.
2964 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
2965 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
2966 * doc/invoke.texi: Add documentation for builtin-expect-probability.
2968 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
2971 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
2972 * calls.c (alloca_call_p): Use get_callee_fndecl.
2973 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
2974 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
2976 (vrp_visit_stmt): Remove duplicated code.
2978 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
2980 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
2981 ceildf2, btruncdf2, instead of vsx_* name.
2983 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
2984 iterators to only do V2DF and V4SF here. Move the DF code to
2985 rs6000.md where it is combined with SF mode. Replace <VSv> with
2986 just 'v' since only vector operations are handled with these insns
2987 after moving the DF support to rs6000.md.
2988 (vsx_sub<mode>3): Likewise.
2989 (vsx_mul<mode>3): Likewise.
2990 (vsx_div<mode>3): Likewise.
2991 (vsx_fre<mode>2): Likewise.
2992 (vsx_neg<mode>2): Likewise.
2993 (vsx_abs<mode>2): Likewise.
2994 (vsx_nabs<mode>2): Likewise.
2995 (vsx_smax<mode>3): Likewise.
2996 (vsx_smin<mode>3): Likewise.
2997 (vsx_sqrt<mode>2): Likewise.
2998 (vsx_rsqrte<mode>2): Likewise.
2999 (vsx_fms<mode>4): Likewise.
3000 (vsx_nfma<mode>4): Likewise.
3001 (vsx_copysign<mode>3): Likewise.
3002 (vsx_btrunc<mode>2): Likewise.
3003 (vsx_floor<mode>2): Likewise.
3004 (vsx_ceil<mode>2): Likewise.
3005 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
3006 (vsx_sminsf3): Likewise.
3007 (vsx_fmadf4): Likewise.
3008 (vsx_fmsdf4): Likewise.
3009 (vsx_nfmadf4): Likewise.
3010 (vsx_nfmsdf4): Likewise.
3011 (vsx_cmpdf_internal1): Likewise.
3013 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
3014 simpler to select whether a target has SPE or traditional floating
3015 point support in iterators.
3016 (TARGET_DF_SPE): Likewise.
3017 (TARGET_SF_FPR): Likewise.
3018 (TARGET_DF_FPR): Likewise.
3019 (TARGET_SF_INSN): Macros to say whether floating point support
3020 exists for a given operation for expanders.
3021 (TARGET_DF_INSN): Likewise.
3023 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
3024 combining of SF/DF mode operations, using both traditional and VSX
3032 (abs<mode>2): Combine SF/DF modes using traditional floating point
3033 instructions. Add support for using the upper DF registers with
3034 VSX support, and SF registers with power8-vector support. Update
3035 expanders for operations supported by both the SPE and traditional
3036 floating point units.
3037 (abs<mode>2_fpr): Likewise.
3038 (nabs<mode>2): Likewise.
3039 (nabs<mode>2_fpr): Likewise.
3040 (neg<mode>2): Likewise.
3041 (neg<mode>2_fpr): Likewise.
3042 (add<mode>3): Likewise.
3043 (add<mode>3_fpr): Likewise.
3044 (sub<mode>3): Likewise.
3045 (sub<mode>3_fpr): Likewise.
3046 (mul<mode>3): Likewise.
3047 (mul<mode>3_fpr): Likewise.
3048 (div<mode>3): Likewise.
3049 (div<mode>3_fpr): Likewise.
3050 (sqrt<mode>3): Likewise.
3051 (sqrt<mode>3_fpr): Likewise.
3052 (fre<Fs>): Likewise.
3053 (rsqrt<mode>2): Likewise.
3054 (cmp<mode>_fpr): Likewise.
3055 (smax<mode>3): Likewise.
3056 (smin<mode>3): Likewise.
3057 (smax<mode>3_vsx): Likewise.
3058 (smin<mode>3_vsx): Likewise.
3059 (negsf2): Delete SF operations that are merged with DF.
3066 (fmasf4_fpr): Likewise.
3067 (fmssf4_fpr): Likewise.
3068 (nfmasf4_fpr): Likewise.
3069 (nfmssf4_fpr): Likewise.
3070 (sqrtsf2): Likewise.
3071 (rsqrtsf_internal1): Likewise.
3072 (smaxsf3): Likewise.
3073 (sminsf3): Likewise.
3074 (cmpsf_internal1): Likewise.
3075 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
3076 (negdf2): Delete DF operations that are merged with SF.
3078 (nabsdf2): Likewise.
3084 (rsqrtdf_internal1): Likewise.
3085 (fmadf4_fpr): Likewise.
3086 (fmsdf4_fpr): Likewise.
3087 (nfmadf4_fpr): Likewise.
3088 (nfmsdf4_fpr): Likewise.
3089 (sqrtdf2): Likewise.
3090 (smaxdf3): Likewise.
3091 (smindf3): Likewise.
3092 (cmpdf_internal1): Likewise.
3093 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
3094 (btrunc<mode>2): Delete separate expander, and combine with the
3095 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
3096 (btrunc<mode>2_fpr): Likewise.
3097 (ceil<mode>2): Likewise.
3098 (ceil<mode>2_fpr): Likewise.
3099 (floor<mode>2): Likewise.
3100 (floor<mode>2_fpr): Likewise.
3101 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
3102 Add support for using the upper registers with VSX and
3103 power8-vector. Move insns to be closer to the define_expands. On
3104 VSX systems, prefer the traditional form of FMA over the VSX
3105 version, since the traditional form allows the target not to
3106 overlap with the inputs.
3107 (fms<mode>4_fpr): Likewise.
3108 (nfma<mode>4_fpr): Likewise.
3109 (nfms<mode>4_fpr): Likewise.
3111 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3112 Richard Earnshaw <richard.earnshaw@arm.com>
3114 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
3115 (struct mult_cost_table): Likewise.
3116 (struct mem_cost_table): Likewise.
3117 (struct fp_cost_table): Likewise.
3118 (struct vector_cost_table): Likewise.
3119 (cpu_cost_table): Likewise.
3120 * config/arm/arm.opt (mold-rts-costs): New option.
3121 (mnew-generic-costs): Likewise.
3122 * config/arm/arm.c (generic_extra_costs): New table.
3123 (cortexa15_extra_costs): Likewise.
3124 (arm_slowmul_tune): Use NULL as new costs.
3125 (arm_fastmul_tune): Likewise.
3126 (arm_strongarm_tune): Likewise.
3127 (arm_xscale_tune): Likewise.
3128 (arm_9e_tune): Likewise.
3129 (arm_v6t2_tune): Likewise.
3130 (arm_cortex_a5_tune): Likewise.
3131 (arm_cortex_a9_tune): Likewise.
3132 (arm_v6m_tune): Likewise.
3133 (arm_fa726te_tune): Likewise.
3134 (arm_cortex_a15_tune): Use cortex15_extra_costs.
3135 (arm_cortex_tune): Use generict_extra_costs.
3136 (shifter_op_p): New function.
3137 (arm_unspec_cost): Likewise.
3138 (LIBCALL_COST): Define.
3139 (arm_new_rtx_costs): New function.
3140 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
3141 table is available. Use old costs otherwise unless mnew-generic-costs
3143 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
3144 (cpu_cost_table): Declare.
3146 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
3149 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
3150 (*subs_mul_imm_<mode>)
3151 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
3152 (*sub_<shift>_<mode>)
3153 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
3154 Remove k constraint.
3156 2013-10-03 Ian Bolton <ian.bolton@arm.com>
3158 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
3160 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
3162 2013-10-02 Teresa Johnson <tejohnson@google.com>
3164 * predict.c (probably_never_executed): New function.
3165 (probably_never_executed_bb_p): Invoke probably_never_executed.
3166 (probably_never_executed_edge_p): Ditto.
3167 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
3168 Treat profile insanities conservatively.
3170 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
3172 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
3174 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
3176 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
3177 use it. Use smaller increase for scratch. Don't increase reject
3178 for early clobber scratch.
3179 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
3180 setting eliminated regs except setting fp from hfp.
3181 (lra_eliminate): Check lra_insn_recog_data on NULL.
3183 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3186 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
3187 setting -mvsx-timode by default until the underlying problem is fixed.
3188 (RS6000_CPU, power7 defaults): Likewise.
3190 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3192 * config/x-linux (host-linux.o): Remove header dependencies.
3193 Use $(COMPILE) and $(POSTCOMPILE).
3194 * config/t-linux-android (linux-android.o): Ditto.
3196 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3198 * Makefile.in (expmed.o-warn): Remove.
3200 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3202 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
3203 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
3205 2013-10-02 Teresa Johnson <tejohnson@google.com>
3207 * dojump.c (do_jump_1): Divide probability between
3208 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
3210 2013-10-02 Tom Tromey <tromey@redhat.com>
3212 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
3214 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3216 * tree-flow.h: Remove some prototypes.
3217 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
3218 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
3219 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
3220 mark_virtual_phi_result_for_renaming): Relocate here.
3221 * tree-into-ssa.h: Add prototypes.
3222 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
3223 single_pred_before_succ_order.
3224 (blocks_in_phiopt_order): Rename and move to cfganal.c.
3225 (nonfreeing_call_p) Move to gimple.c.
3226 * cfganal.c (single_pred_before_succ_order): Move and renamed from
3228 * basic-block.h (single_pred_before_succ_order): Add prototype.
3229 * gimple.c (nonfreeing_call_p): Relocate here.
3230 * gimple.h: Add prototype.
3231 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
3232 * tree-ssa-dom.h: New file. Relocate prototypes here.
3233 * tree-ssa.h: Include tree-ssa-dom.h.
3235 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3237 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
3238 Use $(COMPILE) and $(POSTCOMPILE).
3240 * config/alpha/x-alpha (driver-alpha.o): Ditto.
3242 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3244 * tree-flow.h: Remove some prototypes.
3245 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
3246 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3247 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
3248 propagate_tree_value*): Move from here to...
3249 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
3250 propagate_tree_value*): Relocate here.
3251 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3252 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
3253 * gimple-fold.c: Remove gimple-fold.h from include list.
3254 * tree-vrp.c: Remove gimple-fold.h from include list.
3255 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
3256 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
3257 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
3258 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
3259 * sese.c: Add tree-ssa-propagate.h to include list.
3261 2013-10-02 Richard Biener <rguenther@suse.de>
3263 * tree-loop-distribution.c: Include tree-vectorizer.h for
3265 (enum partition_kind): Remove PKIND_REDUCTION.
3266 (struct partition_s): Remove has_writes member, add reduction_p member.
3267 (partition_alloc): Adjust.
3268 (partition_builtin_p): Likewise.
3269 (partition_has_writes): Remove.
3270 (partition_reduction_p): New function.
3271 (partition_merge_into): Likewise.
3272 (generate_code_for_partition): Commonize builtin partition
3274 (rdg_cannot_recompute_vertex_p): Remove.
3275 (already_processed_vertex_p): Likewise.
3276 (rdg_flag_vertex): Do not set has_writes.
3277 (classify_partition): Adjust.
3278 (rdg_build_partitions): Do not set has_writes, treat all
3279 partitions as useful.
3280 (distribute_loop): Record number of library calls generated. Adjust.
3281 (tree_loop_distribution): Report number of loops and library
3282 calls generated as opt-info.
3284 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3286 * tree-flow.h: Include new .h files. Move prototypes.
3287 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
3288 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
3289 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
3290 * tree-pretty-print.h: Add prototypes from tree-flow.h.
3291 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
3292 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
3293 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
3294 * tree.h (get_ref_base_and_extent): Move prototype out.
3295 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
3297 * gimple-low.h: New File. Add prototypes from tree-flow.h.
3298 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
3299 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
3300 * tree-scalar-evolution.c: Include tree.h.
3301 * sese.c: Include tree.h.
3302 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
3303 * dwarf2out.c: Include tree-dfa.h.
3304 * tree-chrec.c: Include tree.h.
3305 * tree-data-ref.c: Include tree.h.
3307 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
3309 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
3312 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3314 * config/t-sol2 (sol2-c.o): Remove header dependencies.
3315 Use $(COMPILE) and $(POSTCOMPILE).
3316 (sol2-cxx.o): Likewise.
3317 (sol2-stubs.o): Likewise.
3319 * config/x-solaris (host-solaris.o): Likewise.
3321 * config/sparc/t-sparc (sparc.o): Remove.
3322 (sparc-c.o): Remove header dependencies.
3323 Use $(COMPILE) and $(POSTCOMPILE).
3324 * config/sparc/x-sparc: Likewise.
3326 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
3328 * config/arc/arc-opts.h: Add 2013 to Copyright years.
3329 * config/arc/arc700.md: Likewise.
3330 * config/arc/arc-modes.def: Likewise.
3331 * config/arc/arc-simd.h: Likewise.
3332 * config/arc/t-arc-uClibc: Likewise.
3333 * config/arc/t-arc-newlib: Likewise.
3335 2013-10-02 Renlin Li <renlin.li@arm.com>
3337 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
3339 (aarch64_expand_epilogue): Likewise.
3341 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3342 Yufeng Zhang <yufeng.zhang@arm.com>
3344 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
3346 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
3347 'base_in' represent a conversion and legal_cast_p_1 holds; set
3348 'base_in' with the returned value from get_unwidened.
3350 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3352 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
3353 plus_constant is not used.
3355 2013-10-01 Wei Mi <wmi@google.com>
3357 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
3358 * config/i386/i386.md: Add define_peephole2 to
3359 break partial reg stall for cvtss2sd/cvtsd2ss.
3361 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3363 * config/arc/arc.c (pass_arc_ifcvt::clone):
3364 Update for ctxt_ -> m_ctxt change.
3366 2013-10-01 Jeff Law <law@redhat.com>
3368 * tree-ssa-threadupdate.c (struct redirection_data): Delete
3369 outgoing_edge and intermediate_edge fields. Instead store the path.
3370 (redirection_data::hash): Hash on the last edge's destination index.
3371 (redirection_data::equal): Check the entire thread path.
3372 (lookup_redirectio_data): Corresponding changes.
3373 (create_edge_and_update_destination_phis): Likewise.
3374 (thread_single_edge): Likewise.
3376 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3377 Diego Novillo <dnovillo@google.com>
3379 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
3380 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
3381 (vld32wh_insn, vld32wl_insn): Delete commented-out old
3382 versions of these patterns.
3384 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
3385 (__builtin_arc_aligned): Likewise.
3387 * config/arc/arc.md: Expand adc_0 comment stating the intended
3388 purpose and why it isn't ready.
3389 Replace commented out call_value_via_label_mixed with a
3390 plain comment about bl_s.
3392 * config/arc/arc.c (stdio.h): Don't include directly.
3393 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
3394 Remove if (1) condition.
3395 (arc_encode_section_info): Fix comment.
3397 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3399 * config/arc/arc.c (arc_conditional_register_usage):
3400 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
3401 Also set reg_alloc_order for DMA config regs.
3403 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3404 Jeremy Bennett <jeremy.bennett@embecosm.com>
3406 * doc/install.texi (--with-cpu): Mention ARC.
3407 (arc-*-elf32): New paragraph.
3408 (arc-linux-uclibc): Likewise.
3409 * doc/md.texi (Machine Constraints): Add ARC part.
3410 * doc/invoke.texi: (menu): Add ARC Options.
3411 (Machine Dependent Options) <ARC Options>: Add synopsis.
3412 (node ARC Options): Add.
3413 * doc/extend.texi (long_call / short_call attribute): Add ARC.
3414 (ARC Built-in Functions): New section defining
3415 generic ARC built-in functions.
3416 (ARC SIMD Built-in Functions): New section defining SIMD specific
3418 (Declaring Attributes of Functions): Extended
3419 description of short_call and long_call attributes for ARC and
3420 added index entries.
3422 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
3423 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
3424 Joern Rennecke <joern.rennecke@embecosm.com>
3425 Muhammad Khurram Riaz <khurram.riaz@arc.com>
3426 Brendan Kehoe <brendan@zen.org>
3427 Michael Eager <eager@eagercon.com>
3428 Simon Cook <simon.cook@embecosm.com>
3429 Jeremy Bennett <jeremy.bennett@embecosm.com>
3431 * config/arc, common/config/arc: New directories.
3433 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3434 Brendan Kehoe <brendan@zen.org>
3435 Simon Cook <simon.cook@embecosm.com>
3437 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
3439 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3441 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
3442 * tree-ssa-coalesce.h: New. Move prototype to here.
3443 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
3444 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
3445 (gimple_can_coalesce_p): Move to...
3446 * gimple.c (gimple_can_coalesce_p): Here.
3448 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3450 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
3451 (dump_decl_set): Move to gimple.c.
3452 * gimple.h: Don't include tree-ssa-operands.h.
3453 (dump_decl_set): Add prototype.
3454 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3455 Move to gimple-ssa.h.
3456 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
3457 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
3458 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
3459 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
3460 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
3462 (dump_decl_set): Relocate here.
3463 * gimple-ssa.h: New file.
3464 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3465 Relocate from gimple.h.
3466 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
3467 * tree-ssa-operands.c (swap_ssa_operands): Rename from
3468 swap_tree_operands and remove non-ssa path.
3469 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
3470 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
3472 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
3473 vect_is_simple_reduction_1): Use swap_ssa_operands.
3474 * tree-flow.h: Move various prototypes to tree-phinodes.h.
3475 (enum need_phi_state): Move to tree-into-ssa.c.
3476 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3477 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
3478 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
3479 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
3480 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3481 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3482 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3483 num_imm_uses): Move to ssa-iterators.h.
3484 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
3485 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
3487 (op_iter_done, op_iter_next_def, op_iter_next_tree,
3488 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
3489 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
3490 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
3491 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
3492 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
3493 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
3494 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
3495 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
3496 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
3497 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
3498 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
3499 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
3500 (set_phi_nodes): Move to tree-phinodes.h.
3501 * tree-ssa-operands.h (enum ssa_op_iter_type,
3502 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
3503 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
3504 (dump_decl_set): Remove prototype.
3505 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
3506 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
3507 (set_phi_nodes): Relocate from tree-flow-inline.h.
3508 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
3510 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
3511 include list. Temporarily add gimple.h to include list.
3512 * ssa-iterators.h: New file.
3513 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3514 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
3515 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
3516 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
3517 Relocate from tree-ssa-operands.h.
3518 (delink_imm_use, link_imm_use_to_list, link_imm_use,
3519 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3520 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3521 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3522 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
3523 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
3524 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
3525 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
3526 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
3527 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
3528 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
3529 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
3530 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
3531 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
3532 Relocate from tree-flow-inline.h.
3533 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
3535 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
3538 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
3539 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
3540 Insert '\t' to output template.
3541 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
3542 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
3543 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
3544 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
3546 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
3548 * doc/install.texi (Host/target specific installation notes for GCC):
3549 Put @anchor before @heading.
3550 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
3551 Use @email for email addresses.
3553 2013-10-01 Jeff Law <law@redhat.com>
3555 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
3556 a vec. Only delete the path if we create one without successfully
3557 registering a jump thread.
3558 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
3560 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
3561 (paths): New vector of jump threading paths.
3562 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
3563 (THREAD_PATH): New accessor macro for the entire thread path.
3564 (lookup_redirection_data): Get intermediate and final outgoing edge
3565 from the thread path.
3566 (create_edge_and_update_destination_phis): Copy the threading path.
3567 (ssa_fix_duplicate_block_edges): Get edges and block types from the
3568 jump threading path.
3569 (ssa_redirect_edges): Get edges and block types from the jump threading
3570 path. Free the path vector.
3571 (thread_block): Get edges from the jump threading path. Look at the
3572 entire path to see if we thread to a loop exit. If we cancel a jump
3573 thread request, then free the path vector.
3574 (thread_single_edge): Get edges and block types from the jump threading
3575 path. Free the path vector.
3576 (thread_through_loop_header): Get edges and block types from the jump
3577 threading path. Free the path vector.
3578 (mark_threaded_blocks): Iterate over the vector of paths and store
3579 the path on the appropriate edge. Get edges and block types from the
3580 jump threading path.
3581 (mark_threaded_blocks): Get edges and block types from the jump
3582 threading path. Free the path vector.
3583 (thread_through_all_blocks): Use the vector of paths rather than
3584 a vector of 3-edge sets.
3585 (register_jump_thread): Accept pointer to a path vector rather
3586 than the path vector itself. Store the path vector for later use.
3589 2013-10-01 Jakub Jelinek <jakub@redhat.com>
3590 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3593 * config/s390/s390.c (s390_split_branches): Modify check for table
3595 (s390_chunkify_start): Rearrange table jump insn check in order to
3596 deal with compare and branch insns correctly.
3598 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
3602 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
3603 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
3604 define_insn_and_split.
3605 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
3606 (shiftsi3_compare): New pattern.
3608 * config/arm/unspecs.md (UNSPEC_RRX): New.
3610 2013-10-01 Alan Modra <amodra@gmail.com>
3612 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
3613 casing inout operands.
3615 2013-10-01 Richard Biener <rguenther@suse.de>
3617 PR tree-optimization/58553
3618 * tree-loop-distribution.c (struct partition_s): Add niter member.
3619 (classify_partition): Populate niter member for the partition
3620 and properly identify whether the relevant store happens before
3621 or after the loop exit.
3622 (generate_memset_builtin): Use niter member from the partition.
3623 (generate_memcpy_builtin): Likewise.
3625 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3627 * vec.h (vec_prefix, vec): Prefix member names with "m_".
3628 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
3630 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3632 * basic-block.h (edge_list): Prefix member names with "m_".
3633 * context.h (context): Likewise.
3634 * domwalk.h (dom_walker): Likewise.
3635 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
3636 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
3637 * hash-table.h (hash_table): Likewise.
3638 * machmode.h (bit_field_mode_iterator): Likewise.
3639 * pass_manager.h (pass_list): Likewise.
3640 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
3641 * tree-pass.h (pass_data): Likewise.
3642 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
3643 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
3644 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
3645 * asan.c (pass_data_asan): Update accordingly.
3646 * cfganal.c (control_dependences::find_control_dependence): Likewise.
3647 (control_dependences::control_dependences): Likewise.
3648 (control_dependences::~control_dependences): Likewise.
3649 (control_dependences::~control_dependences): Likewise.
3650 (control_dependences::get_edges_dependent_on): Likewise.
3651 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
3652 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
3653 * context.c (gcc::context::context): Likewise.
3654 * cprop.c (pass_rtl_cprop::clone): Likewise.
3655 * domwalk.c (dom_walker::walk): Likewise.
3656 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
3657 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
3658 * mode-switching.c (pass_mode_switching::clone): Likewise.
3659 * passes.c (opt_pass::opt_pass): Likewise.
3660 (pass_manager::pass_manager): Likewise.
3661 * predict.c (pass_strip_predict_hints::clone): Likewise.
3662 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
3663 (pass_split_all_insns::clone): Likewise.
3664 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
3666 (bit_field_mode_iterator::next_mode): Likewise.
3667 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
3668 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
3669 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
3670 * tree-complex.c (pass_lower_complex::clone): Likewise.
3671 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
3672 * tree-object-size.c (pass_object_sizes::clone): Likewise.
3673 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
3674 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
3675 (pass_fold_builtins::clone): Likewise.
3676 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
3677 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
3678 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
3679 (pass_cd_dce::clone): Likewise.
3680 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
3681 (pass_phi_only_cprop::clone): Likewise.
3682 * tree-ssa-dse.c (pass_dse::clone): Likewise.
3683 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
3684 * tree-ssa-loop.c (pass_lim::clone): Likewise.
3685 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
3686 * tree-ssa-pre.c (pass_fre::clone): Likewise.
3687 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
3688 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
3689 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
3690 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
3691 * tree-vrp.c (pass_vrp::clone): Likewise.
3692 * tsan.c (pass_tsan::clone): Likewise.
3694 2013-09-30 Jakub Jelinek <jakub@redhat.com>
3697 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
3698 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
3701 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
3702 optimization, punt if sign_bit_p looked through any zero extension.
3704 2013-09-30 Teresa Johnson <tejohnson@google.com>
3706 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
3707 Update redirected out edge count in joiner case.
3708 (ssa_redirect_edges): Common the joiner and non-joiner cases
3709 so that joiner case gets profile updates.
3711 2013-09-30 Richard Biener <rguenther@suse.de>
3713 PR tree-optimization/58554
3714 * tree-loop-distribution.c (classify_partition): Require
3715 unconditionally executed stores for memcpy and memset recognition.
3716 (tree_loop_distribution): Calculate dominance info.
3718 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3720 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
3721 (NO_PROFILE_COUNTERS): Likewise.
3722 (PROFILE_HOOK): Likewise.
3723 (FUNCTION_PROFILER): Likewise.
3724 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
3726 2013-09-30 Iain Sandoe <iain@codesourcery.com>
3728 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
3729 calls and defines in TARGET_MACHO conditional.
3730 (load_macho_picbase_di): Likewise.
3731 (reload_macho_picbase): Likewise.
3732 (reload_macho_picbase_si): Likewise.
3733 (reload_macho_picbase_di): Likewise.
3734 (nonlocal_goto_receiver): Likewise.
3736 2013-09-30 Nick Clifton <nickc@redhat.com>
3738 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
3739 that use the MSP430X ISA.
3740 (msp430_option_override): Scan -mmcu command line option for any
3741 MCU name that supports the MSP430X ISA.
3742 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
3743 -mmcu options which enable the MSP430X ISA.
3745 2013-09-30 Richard Biener <rguenther@suse.de>
3748 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
3749 before looking for setjmp-like calls.
3751 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3754 * config/darwin-protos.h (machopic_get_function_picbase): New.
3755 * config/darwin.c (machopic_get_function_picbase): New.
3756 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
3757 label for a new func. (load_macho_picbase_di): Likewise.
3758 (reload_macho_picbase): New expand.
3759 (reload_macho_picbase_si): New insn.
3760 (reload_macho_picbase_di): New insn.
3761 (nonlocal_goto_receiver): New define and split.
3762 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
3763 (unspecv enum): Add UNSPECV_NLGR.
3765 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3767 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
3768 that altivec registers are correctly sized on Darwin.
3770 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3772 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
3773 darwin-driver.o): Use COMPILE and POSTCOMPILE.
3774 * config/x-darwin (host-darwin.o): Likewise.
3775 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
3776 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
3777 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
3779 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
3781 * doc/invoke.texi: Fix usage of @tie{} command.
3783 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
3785 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
3787 2013-09-28 Jan Hubicka <jh@suse.cz>
3789 * config/i386/x86-tune.def: Add documentation for each of the options;
3792 2013-09-28 Jan Hubicka <jh@suse.cz>
3794 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
3796 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
3797 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
3798 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
3799 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
3800 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
3802 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
3804 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
3805 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
3806 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
3807 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
3808 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
3809 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
3810 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
3811 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
3812 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
3813 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
3814 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
3815 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
3816 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
3817 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
3818 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
3819 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
3820 gimple.h, godump.c, graphite-clast-to-gimple.c,
3821 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
3822 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
3823 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
3824 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3825 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
3826 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
3827 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
3828 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
3829 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
3830 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
3831 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
3832 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
3833 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
3834 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
3835 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
3836 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
3837 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
3838 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
3839 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
3840 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
3841 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
3842 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
3843 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
3844 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
3845 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3846 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3847 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
3848 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
3849 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
3850 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3851 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3852 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
3853 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
3854 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3855 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
3856 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
3857 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
3858 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
3859 whitespace before "(".
3861 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
3863 * expr.h (extract_bit_field): Remove packedp parameter.
3864 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
3865 from forward declaration.
3866 (store_split_bit_field): Remove packedp arg from calls to
3867 extract_fixed_bit_field.
3868 (extract_bit_field_1): Remove packedp parameter and packedp
3869 argument from recursive calls and calls to extract_fixed_bit_field.
3870 (extract_bit_field): Remove packedp parameter and corresponding
3871 arg to extract_bit_field_1.
3872 (extract_fixed_bit_field): Remove packedp parameter. Remove code
3874 (extract_split_bit_field): Remove packedp arg from call to
3875 extract_fixed_bit_field.
3876 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
3877 (copy_blkmode_from_reg): Likewise.
3878 (copy_blkmode_to_reg): Likewise.
3879 (read_complex_part): Likewise.
3880 (store_field): Likewise.
3881 (expand_expr_real_1): Likewise.
3882 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
3883 to extract_bit_field.
3884 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
3885 call to extract_bit_field.
3886 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
3887 call to extract_bit_field.
3888 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
3889 and special packedp behavior from -fstrict-volatile-bitfields
3892 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3894 * lra-eliminations.c (init_elim_table): Guard value_p.
3896 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3898 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
3899 DFmode, DImode, and SFmode in the upper VSX registers based on the
3900 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
3901 if -mpower8-vector. Combine -mvsx-timode handling with the rest
3902 of the VSX register handling.
3904 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
3906 (zero_extendsidi2_lfiwzx): Add support for loading into the
3907 Altivec registers with -mpower8-vector. Use wu/wv constraints to
3908 only do VSX memory options on Altivec registers.
3909 (extendsidi2_lfiwax): Likewise.
3910 (extendsfdf2_fpr): Likewise.
3911 (mov<mode>_hardfloat, SF/SD modes): Likewise.
3912 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
3913 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
3914 (movdi_internal64): Likewise.
3916 2013-09-27 Xinliang David Li <davidxl@google.com>
3918 * opts.c (finish_options): Adjust parameters
3919 according to vect cost model.
3920 (common_handle_option): Set dynamic vect cost
3922 targhooks.c (default_add_stmt_cost): Compute stmt cost
3924 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
3925 Use helper function.
3926 * tree-vectorizer.h (unlimited_cost_model): New function.
3927 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
3928 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
3930 (vect_enhance_data_refs_alignment): Ditto.
3931 * flag-types.h: New enum.
3932 * common/config/i386/i386-common.c (ix86_option_init_struct):
3933 No need to initialize vect_cost_model flag.
3934 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
3937 2013-09-27 Diego Novillo <dnovillo@google.com>
3939 * gimple.h (enum ssa_mode): Remove.
3941 2013-09-27 Paulo Matos <pmatos@broadcom.com>
3943 * cfgloop.h (number_of_loops): Fix typo in check for null.
3945 2013-09-27 Jakub Jelinek <jakub@redhat.com>
3948 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
3949 are children of outermost saved_cfun's loop, and set it up to
3950 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
3951 if loop != loop0 and SESE region contains bbs that belong to loop0.
3953 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3955 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
3956 (binary_scale_code_p, get_base_term, get_index_term): New functions.
3957 (set_address_segment, set_address_base, set_address_index)
3958 (set_address_disp): Accept the argument unconditionally.
3959 (baseness): Remove must_be_base_p and must_be_index_p checks.
3960 (decompose_normal_address): Classify as much as possible in the
3963 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3965 * cse.c (count_reg_usage): Handle INT_LIST.
3966 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
3967 * reginfo.c (reg_scan_mark_refs): Likewise.
3968 * reload1.c (eliminate_regs_1): Likewise.
3970 2013-09-27 Iain Sandoe <iain@codesourcery.com>
3973 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
3976 2013-09-27 Richard Biener <rguenther@suse.de>
3978 PR tree-optimization/58459
3979 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
3980 restriction not propagating into loops.
3982 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
3984 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
3985 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
3986 * doc/tree-ssa.texi (Walking use-def chains): Delete.
3988 2013-09-26 Richard Biener <rguenther@suse.de>
3990 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
3992 2013-09-26 Richard Biener <rguenther@suse.de>
3994 * alias.h (component_uses_parent_alias_set): Rename to ...
3995 (component_uses_parent_alias_set_from): ... this.
3996 * alias.c (component_uses_parent_alias_set): Rename to ...
3997 (component_uses_parent_alias_set_from): ... this and return
3999 (reference_alias_ptr_type_1): Use the result from
4000 component_uses_parent_alias_set_from instead of stripping
4001 components one at a time.
4002 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
4004 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
4006 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
4007 Move prototypes to...
4008 * tree-ssa-ter.h: New File. Move prototypes here.
4009 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
4010 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
4011 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
4012 from is_replaceable_p.
4013 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
4014 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
4015 (process_replaceable): Use ter_is_replaceable_p.
4016 (find_replaceable_in_bb): Use ter_is_replaceable_p.
4017 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
4018 newly refactored ssa_is_replaceable_p.
4019 * cfgexpand.c: Include tree-outof-ssa.h.
4020 * ssaexpand.h: Delete.
4022 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
4024 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
4025 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
4027 (create_gimple_tmp): Delete.
4028 (get_expr_type, build_assign, build_type_cast): Move to...
4029 * gimple-builder.c: New File.
4030 (get_expr_type): Relocate from gimple.c.
4031 (build_assign, build_type_cast): Change to only create ssanames.
4032 * gimple.h: Move prototypes to...
4033 * gimple-builder.h: New File. Here.
4034 * tree-ssa.h: And here.
4035 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
4036 count_uses_and_derefs): Relocate from gimple.c.
4037 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
4038 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
4039 * tree-ssa-math-opts (execute_cse_reciprocals): Use
4040 gimple_replace_ssa_lhs.
4041 * asan.c: Include gimple-builder.h.
4042 * Makefile.in: Add gimple-builder.o.
4044 2013-09-26 Richard Biener <rguenther@suse.de>
4046 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
4047 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
4048 (loe_visit_block): Use gcc_checking_assert.
4049 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
4050 gimple_assign_ssa_name_copy_p.
4051 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
4054 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
4056 * config/rs6000/t-rs6000 (rs6000.o): Remove.
4057 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
4059 2013-09-26 Richard Biener <rguenther@suse.de>
4061 PR tree-optimization/58539
4062 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
4063 the fact that debug statements are not taking part in loop-closed
4066 2013-09-26 Nick Clifton <nickc@redhat.com>
4068 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
4069 time warning message.
4070 (msp430_print_operand_raw): Delete unused letter parameter.
4071 (TARGET_PRINT_OPERAND_ADDRESS): Define.
4072 (msp430_print_operand_address): New function.
4073 (msp430_print_operand): Move address printing code from here to
4075 * config/msp430/msp430.md (movsipsi2): Add comment in generated
4077 (zero_extendpsisi2): Likewise.
4078 (extendpsisi2): New pattern.
4079 (andneghi3): New pattern.
4081 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
4083 * config/aarch64/aarch64.opt (mlra): New option.
4084 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
4085 (TARGET_LRA_P): Define.
4087 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
4089 * expr.c (expand_assignment): Remove obsolete comment.
4091 2013-09-25 Jeff Law <law@redhat.com>
4093 * tree-flow.h (thread_through_all_blocks): Prototype moved into
4094 tree-ssa-threadupdate.h.
4095 (register_jump_thread): Similarly.
4096 * tree-ssa-threadupdate.h: New header file.
4097 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
4098 * tree-vrp.c: Likewise.
4099 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
4100 (thread_around_empty_blocks): Change type of path vector argument to
4101 an edge,type pair from just an edge. Initialize both elements when
4102 appending to a jump threading path. Tweak references to elements
4104 (thread_across_edge): Similarly. Release memory for the elements
4106 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
4107 (dump_jump_thread_path): New function broken out from
4108 register_jump_thread.
4109 (register_jump_thread): Use dump_jump_thread_path. Change type of
4110 path vector entries. Search the path for NULL edges and dump
4111 the path if one is found. Tweak the conversion of path to 3-edge
4112 form to use the block copy type information embedded in the path.
4114 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
4116 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
4118 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
4119 Vladimir Makarov <vmakarov@redhat.com>
4121 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
4122 from the least significant bit.
4123 (strip_address_mutations): Add bitfield operations handling.
4124 (must_be_index_p): Add shifting and rotate operations handling.
4125 (set_address_base): Use must_be_base_p predicate.
4126 (set_address_index): Use must_be_index_p predicate.
4128 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
4129 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4130 Sergey Lega <sergey.s.lega@intel.com>
4131 Anna Tikhonova <anna.tikhonova@intel.com>
4132 Ilya Tocar <ilya.tocar@intel.com>
4133 Andrey Turetskiy <andrey.turetskiy@intel.com>
4134 Ilya Verbin <ilya.verbin@intel.com>
4135 Kirill Yukhin <kirill.yukhin@intel.com>
4136 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4138 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
4140 (ix86_expand_vector_move_misalign): Support new unaligned load and
4141 stores and use new names.
4142 (CODE_FOR_sse2_storedqu): Rename to ...
4143 (CODE_FOR_sse2_storedquv16qi): ... this.
4144 (CODE_FOR_sse2_loaddqu): Rename to ...
4145 (CODE_FOR_sse2_loaddquv16qi): ... this.
4146 (CODE_FOR_avx_loaddqu256): Rename to ...
4147 (CODE_FOR_avx_loaddquv32qi): ... this.
4148 (CODE_FOR_avx_storedqu256): Rename to ...
4149 (CODE_FOR_avx_storedquv32qi): ... this.
4150 * config/i386/i386.md (fpint_logic): New.
4151 * config/i386/sse.md (VMOVE): Extend for AVX512.
4155 (VI_UNALIGNED_LOADSTORE): Ditto.
4156 (sse2_avx_avx512f): Ditto.
4157 (sse2_avx2): Extend for AVX512.
4158 (sse4_1_avx2): Ditto.
4159 (avx2_avx512f): New.
4160 (sse): Extend for AVX512.
4163 (avxsizesuffix): Ditto.
4164 (sseintvecmode): Ditto.
4166 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
4167 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
4168 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
4169 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
4170 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
4171 (<sse2_avx_avx512f>_storedqu<mode): ... this.
4172 (<sse>_movnt<mode>): Replace constraint "x" with "v".
4173 (STORENT_MODE): Extend for AVX512.
4174 (*absneg<mode>2): Replace constraint "x" with "v".
4175 (*mul<mode>3): Ditto.
4176 (*ieee_smin<mode>3): Ditto.
4177 (*ieee_smax<mode>3): Ditto.
4178 (avx_cmp<mode>3): Replace VF with VF_128_256.
4179 (*<sse>_maskcmp<mode>3_comm): Ditto.
4180 (<sse>_maskcmp<mode>3): Ditto.
4181 (<sse>_andnot<mode>3): Extend for AVX512.
4182 (<code><mode>3, anylogic): Replace VF with VF_128_256.
4183 (<code><mode>3, fpint_logic): New.
4184 (*<code><mode>3): Extend for AVX512.
4185 (avx512flogicsuff): New.
4186 (avx512f_<logic><mode>): Ditto.
4187 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
4189 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
4190 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
4191 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
4192 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
4193 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
4194 (xop_vpermil2<mode>3): Ditto.
4195 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
4196 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
4197 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
4198 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
4200 2013-09-25 Tom Tromey <tromey@redhat.com>
4202 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
4203 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
4204 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
4205 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
4206 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
4207 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
4208 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
4209 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
4210 (GRAPHITE_HTAB_H): Remove.
4212 2013-09-25 Tom Tromey <tromey@redhat.com>
4214 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
4216 2013-09-25 Tom Tromey <tromey@redhat.com>
4218 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
4220 2013-09-25 Tom Tromey <tromey@redhat.com>
4222 * config/i386/t-i386 (i386.o): Remove.
4223 (i386-c.o): Use COMPILE and POSTCOMPILE.
4225 2013-09-25 Tom Tromey <tromey@redhat.com>
4227 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
4229 2013-09-25 Tom Tromey <tromey@redhat.com>
4231 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
4232 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
4233 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
4234 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
4235 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
4236 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
4237 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
4238 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
4239 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
4240 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
4241 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
4242 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
4243 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
4244 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
4245 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
4246 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
4247 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
4248 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
4249 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
4250 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
4251 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
4252 (tree-ssa-pre.o, tree-ssa-sccvn.o)
4253 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
4254 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
4255 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
4256 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
4257 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
4258 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
4259 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
4260 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
4261 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
4262 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
4263 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
4264 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
4265 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
4266 (graphite-clast-to-gimple.o, graphite-dependences.o)
4267 (graphite-interchange.o, graphite-poly.o)
4268 (graphite-scop-detection.o, graphite-sese-to-poly.o)
4269 (graphite-optimize-isl.o, tree-vect-loop.o)
4270 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
4271 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
4272 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
4273 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
4274 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
4275 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
4276 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
4277 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
4278 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
4279 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
4280 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
4281 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
4282 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
4283 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
4284 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
4285 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
4286 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
4287 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
4288 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
4289 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
4290 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
4291 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
4292 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
4293 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
4294 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
4295 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
4296 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
4297 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
4298 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
4299 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
4300 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
4301 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
4302 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
4303 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
4304 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
4305 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
4306 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
4307 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
4308 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
4309 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
4310 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
4311 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
4312 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
4313 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
4314 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
4315 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
4316 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
4317 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
4318 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
4319 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
4320 (gcov-dump.o): Remove.
4321 (default-c.o): Use COMPILE and POSTCOMPILE.
4322 (CFLAGS-gcc.o): New variable.
4323 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
4325 2013-09-25 Tom Tromey <tromey@redhat.com>
4327 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
4328 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
4329 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
4330 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
4331 (c-family/c-pragma.o, c-family/c-pretty-print.o)
4332 (c-family/c-semantics.o, c-family/c-ada-spec.o)
4333 (c-family/array-notation-common.o, c-family/stub-objc.o)
4334 (c-family/c-ubsan.o): Remove.
4336 2013-09-25 Tom Tromey <tromey@redhat.com>
4338 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
4340 2013-09-25 Tom Tromey <tromey@redhat.com>
4342 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
4343 to add -DENABLE_SHARED_LIBGCC.
4344 (gcc.o): Don't use subshell.
4346 2013-09-25 Tom Tromey <tromey@redhat.com>
4348 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
4349 * configure.ac: Don't invoke AM_PROG_CC_C_O.
4350 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
4351 * configure, config.in: Rebuild.
4353 2013-09-25 Tom Tromey <tromey@redhat.com>
4355 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
4356 (COMPILE, POSTCOMPILE): New variables.
4357 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
4358 (DEPFILES): New variable.
4359 Include ".Po" files.
4360 * configure.ac: Add checks for dependency checking.
4361 * configure, aclocal.m4: Regenerate.
4363 2013-09-25 Tom Tromey <tromey@redhat.com>
4365 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
4366 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
4368 2013-09-25 Tom Tromey <tromey@redhat.com>
4370 * Makefile.in (generated_files): Add options.h,
4371 target-hooks-def.h, insn-opinit.h,
4372 common/common-target-hooks-def.h, pass-instances.def,
4373 c-family/c-target-hooks-def.h.
4375 2013-09-25 Jeff Law <law@redhat.com>
4377 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
4378 than foo[foo.length () - 1] to access last member in a vec.
4379 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
4381 2013-09-25 Richard Biener <rguenther@suse.de>
4384 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
4386 2013-09-25 Jan Hubicka <jh@suse.cz>
4388 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
4391 2013-09-25 Marek Polacek <polacek@redhat.com>
4394 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
4395 when determining the type name.
4397 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
4399 * config/sh/sh.md: Fix formatting.
4401 2013-09-24 Xinliang David Li <davidxl@google.com>
4403 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
4404 max peel iterations parameter.
4405 * param.def: New parameter.
4406 * doc/invoke.texi: Document New parameter.
4408 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
4410 * gimple-pretty-print.c: Various whitespace tweaks.
4411 * tree-core.h: Likewise.
4412 * tree-pretty-print.c: Likewise.
4413 * tree-ssa-alias.c: Likewise.
4414 * tree-ssa-copy.c: Likewise.
4415 * tree-ssanames.c: Likewise.
4416 * tree-ssanames.h: Likewise.
4417 * tree-vrp.c: Likewise.
4419 2013-09-24 Alan Modra <amodra@gmail.com>
4423 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
4424 for output operands that disallow regs. Don't use EXPAND_WRITE on
4427 2013-09-24 Richard Biener <rguenther@suse.de>
4430 * tree.c (reference_alias_ptr_type): Move ...
4431 * alias.c (reference_alias_ptr_type): ... here and implement
4432 in terms of the new reference_alias_ptr_type_1.
4433 (ref_all_alias_ptr_type_p): New helper.
4434 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
4435 use ref_all_alias_ptr_type_p.
4436 (get_deref_alias_set): Add flag_strict_aliasing check here.
4437 (reference_alias_ptr_type_1): New function, split out from ...
4438 (get_alias_set): ... here.
4439 (alias_ptr_types_compatible_p): New function.
4440 * alias.h (reference_alias_ptr_type): Declare.
4441 (alias_ptr_types_compatible_p): Likewise.
4442 * tree.h (reference_alias_ptr_type): Remove.
4443 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
4444 to compare MEM_REF alias types.
4446 2013-09-24 Richard Biener <rguenther@suse.de>
4448 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
4450 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
4452 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
4453 reload helper function arrays into a single array reg_addr.
4454 (reload_fpr_gpr): Likewise.
4455 (reload_gpr_vsx): Likewise.
4456 (reload_vsx_gpr): Likewise.
4457 (struct rs6000_reg_addr): Likewise.
4458 (reg_addr): Likewise.
4459 (rs6000_debug_reg_global): Change rs6000_vector_reload,
4460 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
4461 (rs6000_init_hard_regno_mode_ok): Likewise.
4462 (rs6000_secondary_reload_direct_move): Likewise.
4463 (rs6000_secondary_reload): Likewise.
4465 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
4466 constraints: wu, ww, and wy. Repurpose wv constraint added during
4467 power8 changes. Put wg constraint in alphabetical order.
4469 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
4470 for future work to add ISA 2.07 VSX single precision support.
4471 (-mvsx-scalar-double): Change default from -1 to 1, update
4472 documentation comment.
4473 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
4474 (-mupper-regs-df): New debug switch to control whether DF values
4475 can go in the traditional Altivec registers.
4476 (-mupper-regs-sf): New debug switch to control whether SF values
4477 can go in the traditional Altivec registers.
4479 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
4481 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
4482 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
4483 Add new constraints, wu/ww/wy. Repurpose wv constraint.
4484 (rs6000_debug_legitimate_address_p): Print if we are running
4485 before, during, or after reload.
4486 (rs6000_secondary_reload): Add a comment.
4487 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
4489 * config/rs6000/constraints.md (wa constraint): Sort w<x>
4490 constraints. Update documentation string.
4491 (wd constraint): Likewise.
4492 (wf constraint): Likewise.
4493 (wg constraint): Likewise.
4494 (wn constraint): Likewise.
4495 (ws constraint): Likewise.
4496 (wt constraint): Likewise.
4497 (wx constraint): Likewise.
4498 (wz constraint): Likewise.
4499 (wu constraint): New constraint for ISA 2.07 SFmode scalar
4501 (ww constraint): Likewise.
4502 (wy constraint): Likewise.
4503 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
4504 in the previous submissions.
4505 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
4507 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
4509 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
4510 (REG_BR_PROB): Say that the probability is stored in an int_list.
4511 * reg-notes.def: Update commentary to mention INT_LIST.
4512 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
4513 (INT_LIST): New rtx.
4514 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
4515 * rtlanal.c (int_reg_note_p): New function.
4516 (alloc_reg_note): Assert that the note does not have an int argument.
4517 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
4518 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
4519 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
4520 rather than an INSN_LIST. Handle INT_LIST.
4521 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
4522 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
4523 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
4524 Manipulate them as ints rather than rtxes.
4525 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
4526 * regmove.c (copy_src_to_dest): Likewise.
4527 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
4529 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
4530 into the cases that need it.
4531 * config/arm/arm.c (arm_unwind_emit): Likewise.
4533 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
4534 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
4535 * loop-doloop.c (add_test, doloop_modify): Likewise.
4536 * loop-unswitch.c (compare_and_jump_seq): Likewise.
4537 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
4538 * predict.c (combine_predictions_for_insn): Likewise.
4539 * print-rtl.c (print_rtx): Handle INT_LIST.
4540 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
4541 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
4542 * config/arm/arm.c (emit_unlikely_jump): Likewise.
4543 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
4544 (ix86_split_fp_branch, predict_jump): Likewise.
4545 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
4546 * config/sh/sh.c (expand_cbranchsi4): Likewise.
4547 * config/spu/spu.c (ea_load_store_inline): Likewise.
4549 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
4550 value of a REG_BR_PROB note.
4551 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
4552 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
4553 * emit-rtl.c (try_split): Likewise.
4554 * predict.c (br_prob_note_reliable_p): Likewise.
4555 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
4556 * reorg.c (mostly_true_jump): Likewise.
4557 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
4558 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
4559 * config/i386/i386.c (ix86_print_operand): Likewise.
4560 * config/ia64/ia64.c (ia64_print_operand): Likewise.
4561 * config/mmix/mmix.c (mmix_print_operand): Likewise.
4562 * config/rs6000/rs6000.c (output_cbranch): Likewise.
4563 * config/s390/s390.c (s390_expand_tbegin): Likewise.
4564 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
4565 * config/sparc/sparc.c (output_cbranch): Likewise.
4566 * config/spu/spu.c (get_branch_target): Likewise.
4567 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
4568 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
4570 2013-09-23 Jan Hubicka <jh@suse.cz>
4572 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
4574 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
4575 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
4577 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
4578 (gimple_fold_call): Dump inconsistent devirtualizations; add
4579 sanity check for type based devirtualizations.
4580 * ipa-prop.c: Include ipa-utils.h
4581 (ipa_intraprocedural_devirtualization): Add sanity check.
4582 (try_make_edge_direct_virtual_call): Likewise.
4584 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
4586 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
4587 assignment statements.
4589 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
4591 * gimple-pretty-print.c (dump_ssaname_info): New function.
4592 (dump_gimple_phi): Call it.
4593 (pp_gimple_stmt_1): Likewise.
4594 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
4595 (range_info_def): Declare.
4596 * tree-pretty-print.c (pp_double_int): New function.
4597 (dump_generic_node): Call it.
4598 * tree-pretty-print.h (pp_double_int): Declare.
4599 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
4600 * tree-ssanames.h (range_info_def): New structure.
4601 (value_range_type): Move definition here.
4602 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
4604 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
4606 (set_range_info): New function.
4607 (get_range_info): Likewise.
4608 (duplicate_ssa_name_range_info): Likewise.
4609 (duplicate_ssa_name_fn): Check pointer type and call
4610 duplicate_ssa_name_range_info.
4611 * tree-ssa-copy.c (fini_copy_prop): Likewise.
4612 * tree-vrp.c (value_range_type): Remove definition, now in
4614 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
4615 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
4616 (SSA_NAME_RANGE_INFO): New macro.
4618 2013-09-23 Richard Biener <rguenther@suse.de>
4620 PR tree-optimization/58464
4621 * tree-ssa-pre.c (phi_trans_lookup): Remove.
4622 (phi_trans_add): Change to add conditionally on being not present.
4623 (phi_translate_1): Remove recursion detection here.
4624 (phi_translate): Pre-seed the cache with NULL to catch
4625 recursion here in a more generic way.
4626 (bitmap_find_leader): Adjust comment.
4627 (get_representative_for): Dump value-numbers.
4628 (create_expression_by_pieces): Likewise.
4629 (insert_into_preds_of_block): Likewise.
4631 2013-09-23 Christian Bruel <christian.bruel@st.com>
4634 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
4635 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
4637 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
4640 2013-09-20 Renlin Li <renlin.li@arm.com>
4642 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
4644 (aarch64_expand_epilogue): Likewise.
4645 (aarch64_legitimize_reload_address): Likewise.
4647 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
4649 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
4650 NULL_TREE before pushing them onto the vector. Likewise for labels.
4652 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
4654 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
4655 comparison operators when -fno-trapping-math is in effect.
4656 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
4657 comparison operators in TFmode and assert that unsupported operators
4659 (ia64_print_operand): Likewise.
4661 2013-09-21 Jan Hubicka <jh@suse.cz>
4663 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
4664 (sse_typeless_stores): Enable for core
4665 (sse_load0_by_pxor): Likewise.
4666 (four_jump_limit): Disable for core.
4667 (pad_returns): Likewise.
4668 (avoid_vector_decode): Likewise.
4669 (fuse_cmp_and_branch): Enable for cores.
4670 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
4672 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4675 * config/pa/pa.c (pa_option_override): Disable auto increment and
4676 decrement instructions until reload is completed.
4678 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
4679 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
4680 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
4682 2013-09-20 DJ Delorie <dj@redhat.com>
4683 Nick Clifton <nickc@redhat.com>
4685 * config/rl78/rl78.c: Various whitespace and comment tweaks.
4686 (need_to_save): Save bank 0 on interrupts.
4687 (characterize_address): Strip far address wrappers.
4688 (rl78_as_legitimate_address): Likewise.
4689 (transcode_memory_rtx): Likewise.
4690 (rl78_peep_movhi_p): Disable this peephole after devirt.
4691 (rl78_propogate_register_origins): Forget all origins when a
4693 * config/rl78/rl78-virt.md: Various whitespace tweaks.
4694 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
4696 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
4697 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
4698 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
4700 * config/rl78/constraints.md: Various whitespace and paren tweaks.
4702 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4704 * config/pa/pa.md: In "scc" insn patterns, change output template to
4705 handle const0_rtx in reg_or_0_operand operands.
4707 2013-09-20 Martin Husemann <martin@NetBSD.org>
4710 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
4711 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
4713 2013-09-20 Richard Biener <rguenther@suse.de>
4716 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
4717 remembering SSA name version and block index.
4718 (new_scev_info_str): Adjust.
4719 (hash_scev_info): Likewise. Also hash the block index.
4720 (eq_scev_info): Adjust.
4721 (find_var_scev_info): Likewise.
4722 (struct instantiate_cache_entry): Remove.
4723 (struct instantiate_cache_type): Use a htab to map name, block
4725 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4726 (get_instantiated_value_entry): Likewise.
4727 (hash_idx_scev_info, eq_idx_scev_info): New functions.
4728 (instantiate_scev_name): Adjust.
4730 2013-09-20 Jeff Law <law@redhat.com>
4732 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
4734 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
4736 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
4737 Call aarch64_simd_expand_args to update op[argc].
4739 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4741 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
4744 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4746 * gengtype.c (file_rules): Added rule for *.cc files.
4747 (get_output_file_with_visibility): Give fatal message when no
4750 2013-09-20 Renlin Li <renlin.li@arm.com>
4752 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
4753 (aarch64_expand_epilogue): Likewise.
4754 (aarch64_legitimize_reload_address): Likewise.
4756 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
4759 * expr.c (expand_assignment): Remove misalignp code path.
4761 2013-09-20 Marek Polacek <polacek@redhat.com>
4764 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
4765 TYPE_PRECISION. Add asserts.
4767 2013-09-20 Richard Biener <rguenther@suse.de>
4769 PR tree-optimization/58453
4770 * tree-loop-distribution.c (distribute_loop): Apply the cost
4771 model for -ftree-loop-distribute-patterns, too.
4773 2013-09-20 Richard Biener <rguenther@suse.de>
4776 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
4777 make type comparison less strict.
4779 2013-09-20 Alan Modra <amodra@gmail.com>
4781 * configure: Regenerate.
4782 * aclocal.m4: Regenerate.
4784 2013-09-20 Marek Polacek <polacek@redhat.com>
4787 * doc/extend.texi: Document that attribute used is meant to be used
4788 on variables with static storage duration.
4790 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4792 PR tree-optimization/58472
4793 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
4794 simd_lane_access set inv_p = false.
4795 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
4796 the simduid magic VAR_DECL.
4798 2013-09-19 Jan Hubicka <jh@suse.cz>
4800 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
4802 2013-09-17 Jeff Law <law@redhat.com>
4804 * tree-ssa-dom.c (record_temporary_equivalences): New function
4805 split out of dom_opt_dom_walker::after_dom_children.
4806 (dom_opt_dom_walker::thread_across_edge): Move common code
4807 in here from dom_opt_dom_walker::after_dom_children.
4808 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
4810 2013-09-19 Jan Hubicka <jh@suse.cz>
4812 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
4813 (TARGET_GENERIC): Use PROCESOR_GENERIC
4814 (enum processor_type): Unify generic32 and 64.
4815 * i386.md (cpu): Likewise.
4816 * x86-tune.def (use_leave): Enable for generic32.
4817 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
4818 * athlon.md: Change generic64 to generic in all occurences.
4819 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
4820 (ix86_target_macros_internal): Likewise.
4821 * driver-i386.c (host_detect_local_cpu): Likewise.
4822 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
4824 (generic_memcpy, generic_memset, generic_cost): This one.
4825 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
4826 (m_GENERIC32, m_GENERIC64): Remove.
4827 (m_GENERIC): Turn into one flag.
4828 (processor_target): Unify generic tunnings.
4829 (ix86_option_override_internal): Replace generic32/64 by generic.
4830 (ix86_issue_rate): Likewise.
4831 (ix86_adjust_cost): Likewise.
4833 2013-09-19 Jan Hubicka <jh@suse.cz>
4835 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
4836 of speculative flag.
4838 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4840 * omp-low.c (expand_omp_sections): Always pass len - 1 to
4841 GOMP_sections_start, even if !exit_reachable.
4843 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4845 * lra-constraints.c (need_for_all_save_p): Use macro
4846 HARD_REGNO_CALL_PART_CLOBBERED.
4847 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
4848 set up pseudo conflict hard regs.
4850 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4853 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
4856 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4858 PR rtl-optimization/58438
4859 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
4860 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
4861 from previous subpasses.
4863 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
4865 * arm.c (arm_get_frame_offsets): Validate architecture supports
4866 LDRD/STRD before accepting the tuning preference.
4867 (arm_expand_prologue): Likewise.
4868 (arm_expand_epilogue): Likewise.
4870 2013-09-18 Richard Biener <rguenther@suse.de>
4872 PR tree-optimization/58417
4873 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
4874 have chrecs with symbols defined in the loop as operands.
4875 (chrec_fold_multiply): Likewise.
4876 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
4877 parameters before folding binary operations.
4878 (struct instantiate_cache_entry_hasher): Remove.
4879 (struct instantiate_cache_type): Use a pointer-map.
4880 (instantiate_cache_type::instantiate_cache_type): New function.
4881 (instantiate_cache_type::get): Likewise.
4882 (instantiate_cache_type::set): Likewise.
4883 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4884 (get_instantiated_value_entry): Likewise.
4885 (global_cache): New global.
4886 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
4887 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
4888 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
4889 (instantiate_scev_name): Adjust.
4890 (instantiate_scev): Construct global instead of local cache.
4891 (resolve_mixers): Likewise.
4893 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
4894 Paolo Carlini <paolo.carlini@oracle.com>
4897 * doc/implement-cxx.texi: Fix references to the C++ standards.
4899 2013-09-18 Jakub Jelinek <jakub@redhat.com>
4901 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
4902 * tree-vect-data-refs.c (vect_analyze_data_refs): For
4903 simd_lane_access drs, update also DR_ALIGNED_TO.
4905 2013-09-18 Marek Polacek <polacek@redhat.com>
4908 * doc/extend.texi: Document no_sanitize_undefined attribute.
4909 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
4910 no_sanitize_undefined attribute.
4912 2013-09-18 Nick Clifton <nickc@redhat.com>
4914 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
4915 (ASM_DECLARE_FUNCTION_NAME): Define.
4917 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
4919 * compare-elim.c (find_comparison_dom_walker): New class
4920 (find_comparisons_in_bb): Rename to
4921 find_comparison_dom_walker::before_dom_children
4922 (find_comparisons): Adjust
4923 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
4925 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
4926 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
4927 (init_walk_dominator_tree): Remove declaration.
4928 (fini_walk_dominator_tree): Remove declaration.
4929 * fwprop.c (single_def_use_dom_walker): New class
4930 (single_def_use_enter_block): Convert to
4931 single_def_use_dom_walker::before_dom_children.
4932 (single_def_use_leave_block): Convert to
4933 single_def_use_dom_walker::after_dom_children.
4934 (build_single_def_use_links): Adjust.
4935 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
4937 (find_candidates_in_block): Convert to
4938 find_candidates_dom_walker::before_dom_children.
4939 (execute_strength_reduction): Adjust.
4940 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
4941 (sese_dom_walker): New class.
4942 (sese_dom_walker::sese_dom_walker): New constructor.
4943 (sese_dom_walker::~sese_dom_walker): New destructor.
4944 (build_sese_conditions_before): Convert to
4945 sese_dom_walker::before_dom_children.
4946 (build_sese_conditions_after): Convert to
4947 sese_dom_walker::after_dom_children.
4948 (build_poly_scop): Adjust
4949 * tree-into-ssa.c (rewrite_dom_walker): New class
4950 (rewrite_enter_block): Convert to
4951 rewrite_dom_walker::before_dom_children.
4952 (rewrite_leave_block): Convert to
4953 rewrite_dom_walker::after_dom_children.
4954 (rewrite_update_dom_walker): New class.
4955 (rewrite_update_enter_block): Convert to
4956 rewrite_update_dom_walker::before_dom_children.
4957 (rewrite_update_leave_block): Convert to
4958 rewrite_update_dom_walker::after_dom_children.
4959 (rewrite_blocks, rewrite_into_ssa): Adjust.
4960 (mark_def_dom_walker): New class.
4961 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
4962 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
4963 (mark_def_sites_blocks): Convert to
4964 mark_def_dom_walker::before_dom_children.
4965 (mark_def_site_blocks): Remove.
4966 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
4967 (tree_ssa_dominator_optimize): Adjust.
4968 (dom_thread_across_edge): Convert to method
4969 dom_opt_dom_walker::thread_across_edge.
4970 (dom_opt_enter_block): Convert to member function
4971 dom_opt_dom_walker::before_dom_children.
4972 (dom_opt_leave_block): Convert to member function
4973 dom_opt_dom_walker::after_dom_children.
4974 * tree-ssa-dse.c (dse_dom_walker): New class.
4975 (dse_enter_block): Convert to member function
4976 dse_dom_walker::before_dom_children.
4977 (tree_ssa_dse): Adjust.
4978 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
4979 (determine_invariantness_stmt): Convert to method
4980 invariantness_dom_walker::before_dom_children.
4981 (determine_invariantness): Remove
4982 (move_computations_dom_walker): New class.
4983 (move_computations_stmt): Convert to method
4984 move_computations_dom_walker::before_dom_children.
4985 (move_computations, tree_ssa_lim): Adjust.
4986 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
4987 (nt_init_block): Make method
4988 notrappping_dom_walker::before_dom_children.
4989 (nt_fini_block): Make
4990 method nontrapping_dom_walker::after_dom_children.
4991 (get_non_trapping): Adjust.
4992 * tree-ssa-pre.c (eliminate_dom_walker): New class.
4993 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
4994 (eliminate_leave_block): Make method.
4995 eliminate_dom_walker::after_dom_children.
4997 * tree-ssa-strlen.c (strlen_dom_walker): New class.
4998 (strlen_enter_block): Make method
4999 strlen_dom_walker::before_dom_children.
5000 (strlen_leave_block): Make
5001 method strlen_dom_walker::after_dom_children.
5002 (tree_ssa_strlen): Adjust.
5003 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
5004 (tree_ssa_uncprop): Adjust.
5005 (uncprop_leave_block): Make method
5006 uncprop_dom_walker::after_dom_children.
5007 (uncprop_leave_block): Make method
5008 uncprop_dom_walker::before_dom_children.
5010 2013-09-18 Bin Cheng <bin.cheng@arm.com>
5012 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
5013 branch in same basic block. Check both src and dest of the move insn.
5015 2013-09-17 Nick Clifton <nickc@redhat.com>
5017 * config/rl78/rl78-real.md (bf): New pattern.
5019 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
5020 (rl78_print_operand): Do not put a # before a %B.
5021 * config/rl78/rl78.opt: Tweak doc strings.
5023 2013-09-17 DJ Delorie <dj@redhat.com>
5025 * config/rl78/constraints.md (Wcv): Allow up to $r31.
5026 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
5027 (rl78_option_override): Likewise, if -mallregs.
5028 (is_virtual_register): Likewise.
5029 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
5030 (REGNO_OK_FOR_BASE_P): Likewise.
5031 * config/rl78/rl78.opt (-mallregs): New.
5033 2013-09-17 Nick Clifton <nickc@redhat.com>
5035 * config/rl78/rl78.c (need_to_save): Change return type to bool.
5036 For interrupt functions: save all call clobbered registers if the
5037 interrupt handler is not a leaf function.
5038 (rl78_expand_prologue): Always recompute the frame information.
5039 For interrupt functions: only select bank 0 if one of the bank 0
5040 registers is going to be psuhed.
5042 2013-09-17 DJ Delorie <dj@redhat.com>
5044 * config/rl78/constraints.md: For each W* constraint, rename to C*
5045 and create a W* constraint that checks for an optional ES: prefix
5047 * config/rl78/rl78.md (UNS_ES_ADDR): New.
5048 (es_addr): New. Used to wrap far addresses.
5049 * config/rl78/rl78-protos.h (rl78_es_addr): New.
5050 (rl78_es_base): New.
5051 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
5052 wrapped far addresses.
5053 (rl78_print_operand_1): Unwrap far addresses before processing.
5054 (rl78_lo16): Wrap far addresses in unspecs.
5055 (rl78_es_addr): New.
5056 (rl78_es_base): New.
5057 (insn_ok_now): Check for not-yet-wrapped far addresses.
5058 (transcode_memory_rtx): Properly re-wrap far addresses.
5060 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
5062 * config/sparc/t-rtems: Add leon3 multilibs.
5064 2013-09-17 Cong Hou <congh@google.com>
5066 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
5067 when checking the dot production pattern. The type of rhs operand
5068 of multiply is now checked correctly.
5070 2013-09-17 Jeff Law <law@redhat.com>
5072 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5073 edge implied equivalences into successor phis.
5074 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
5075 here from tree-ssa-threadedge.c.
5076 (mark_threaded_blocks): When threading through a joiner, if both
5077 successors of the joiner's clone reach the same block, verify the
5078 PHI arguments are equal. If not, cancel the jump threading request.
5079 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
5080 tree-ssa-threadupdate.c
5081 (thread_across_edge): Don't check PHI argument equality when
5082 threading through joiner block here.
5084 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
5086 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
5087 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
5088 (warn_uninit, warn_uninitialized_vars,
5089 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
5090 Move to tree-ssa-uninit.c.
5091 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
5092 (has_undefined_value_p): New. Pass dependant parts of
5093 ssa_undefined_value_p.
5094 (uninit_undefined_value_p): Use has_undefined_value_p.
5095 (warn_uninit, warn_uninitialized_vars,
5096 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
5097 Move from tree-ssa.c.
5098 * tree-ssa.h: Adjust prototypes.
5100 2013-09-17 Jan Hubicka <jh@suse.cz>
5103 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
5104 * ipa-inline.c (can_inline_edge_p): Do not downgrade
5105 FUNCTION_NOT_OPTIMIZED.
5106 * ipa-inline-analysis.c (compute_inline_parameters): Function
5107 not optimized is not inlinable unless it is alwaysinline.
5108 (inline_analyze_function): Force calls in not optimized
5109 function not inlinable.
5111 2013-09-17 Jan Hubicka <jh@suse.cz>
5114 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
5116 * ipa.c (function_and_variable_visibility): Likewise.
5117 * ipa-profile.c (ipa_profile): Likewise.
5119 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
5122 * cgraph.c (cgraph_function_body_availability): Check for ifunc
5123 attribute, and don't inline the resolver in this case.
5125 2013-09-17 Teresa Johnson <tejohnson@google.com>
5127 * coverage.c (get_coverage_counts): Add missing newline.
5129 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5131 PR tree-optimization/58088
5132 * fold-const.c (mask_with_trailing_zeros): New function.
5133 (fold_binary_loc): Make sure we don't recurse infinitely
5134 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
5135 Use mask_with_trailing_zeros where appropriate.
5137 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
5139 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
5140 of 'prev' var to get better distance estimation.
5142 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
5144 * tree-inline.h (struct copy_body_data): Add transform_parameter.
5145 * tree-inline.c (is_parameter_of): New predicate.
5146 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
5147 a parameter has been remapped.
5148 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
5149 (optimize_inline_calls): Initialize transform_parameter.
5150 (copy_gimple_seq_and_replace_locals): Likewise.
5151 (tree_function_versioning): Likewise.
5152 (maybe_inline_call_in_expr): Likewise.
5154 2013-09-17 Nick Clifton <nickc@redhat.com>
5156 * config/msp430/msp430-protos.h: Add prototypes for new functions.
5157 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
5159 (is_attr_func): New function.
5160 (msp430_is_interrupt_func): New function.
5161 (is_naked_func): New function.
5162 (is_reentrant_func): New function.
5163 (is_critical_func): New function.
5164 (msp430_start_function): Add annotations for function attributes.
5165 (msp430_attr): New function.
5166 (msp430_attribute_table): New.
5167 (msp430_function_section): New function.
5168 (TARGET_ASM_FUNCTION_SECTION): Define.
5169 (msp430_builtin): New enum.
5170 (msp430_init_builtins): New function.
5171 (msp430_builtin_devl): New function.
5172 (msp430_expand_builtin): New function.
5173 (TARGET_INIT_BUILTINS): Define.
5174 (TARGET_EXPAND_BUILTINS): Define.
5175 (TARGET_BUILTIN_DECL): Define.
5176 (msp430_expand_prologue): Add support for naked, interrupt,
5177 critical and reentrant functions.
5178 (msp430_expand_epilogue): Likewise.
5179 (msp430_print_operand): Handle 'O' character.
5180 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
5182 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
5183 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
5184 (pushm): Use a 'n' rather than an 'i' constraint.
5185 (msp_return): Add generation of the interrupt return instruction.
5186 (disable_interrupts): New pattern.
5187 (enable_interrupts): New pattern.
5188 (push_intr_state): New pattern.
5189 (pop_intr_state): New pattern.
5190 (bic_SR): New pattern.
5191 (bis_SR): New pattern.
5192 * doc/extend.texi: Document MSP430 function attributes and builtin
5195 2013-09-17 Richard Biener <rguenther@suse.de>
5197 PR tree-optimization/58432
5198 * tree-loop-distribution.c (tree_loop_distribution): Also
5199 scan PHIs for outside loop uses and seed a partition from them.
5201 2013-09-17 Bin Cheng <bin.cheng@arm.com>
5203 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
5204 (restructure_reference): Call backtrace_base_for_ref.
5206 2013-09-17 Alan Modra <amodra@gmail.com>
5209 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
5212 2013-09-16 DJ Delorie <dj@redhat.com>
5214 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
5215 vregs location for RL78/G10.
5216 (rl78_expand_prologue): Avoid SEL on G10.
5217 (rl78_expand_epilogue): Likewise.
5218 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
5219 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
5220 __RL78_G10__ when appropriate.
5221 (ASM_SPEC): Pass -mg10 along to the assembler.
5222 * config/rl78/rl78.md (sel_rb): Disable for G10.
5223 * config/rl78/rl78.opt: Add -mg10 option.
5224 * config/rl78/t-rl78: Add -mg10 multilib.
5226 2013-09-16 Xinliang David Li <davidxl@google.com>
5228 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
5229 * omp-low.c (omp_max_vf): Ditto.
5230 (expand_omp_simd): Ditto.
5231 * tree-vectorizer.c (vectorize_loops): Ditto.
5232 (gate_vect_slp): Ditto.
5233 (gate_increase_alignment): Ditto.
5234 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
5235 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
5236 (gate_tree_vectorize): Name change.
5237 (tree_vectorize): Ditto.
5238 (pass_vectorize::gate): Call new function.
5239 (pass_vectorize::execute): Ditto.
5240 * opts.c: O3 default setting change.
5241 (finish_options): Check new flag.
5242 * doc/invoke.texi: Document new flags.
5243 * common.opt: New flags.
5245 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
5247 * doc/tm.texi.in (Cond Exec Macros): Remove node.
5248 (Condition Code): Don't reference it.
5249 * doc/tm.texi: Regenerate.
5251 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
5254 * lra-constraints.c (undo_optional_reloads): Consider all optional
5255 reload even if it did not get a hard reg.
5257 2013-09-16 Teresa Johnson <tejohnson@google.com>
5259 * dumpfile.c (dump_loc): Remove newline emission.
5260 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
5261 emission to dump_printf_loc calls where missing.
5262 (vect_mark_for_runtime_alias_test): Ditto.
5263 (vect_analyze_data_ref_dependence): Ditto.
5264 (vect_analyze_data_ref_dependences): Ditto.
5265 (vect_slp_analyze_data_ref_dependence): Ditto.
5266 (vect_slp_analyze_data_ref_dependences): Ditto.
5267 (vect_compute_data_ref_alignment): Ditto.
5268 (vect_update_misalignment_for_peel): Ditto.
5269 (vect_verify_datarefs_alignment): Ditto.
5270 (vector_alignment_reachable_p): Ditto.
5271 (vect_get_data_access_cost): Ditto.
5272 (vect_enhance_data_refs_alignment): Ditto.
5273 (vect_find_same_alignment_drs): Ditto.
5274 (vect_analyze_data_refs_alignment): Ditto.
5275 (vect_analyze_group_access): Ditto.
5276 (vect_analyze_data_ref_access): Ditto.
5277 (vect_analyze_data_ref_accesses): Ditto.
5278 (vect_prune_runtime_alias_test_list): Ditto.
5279 (vect_analyze_data_refs): Ditto.
5280 (vect_create_addr_base_for_vector_ref): Ditto.
5281 (vect_create_data_ref_ptr): Ditto.
5282 (vect_grouped_store_supported): Ditto.
5283 (vect_grouped_load_supported): Ditto.
5284 * value-prof.c (check_counter): Ditto.
5285 (check_ic_target): Ditto.
5286 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
5287 (vect_recog_widen_mult_pattern): Ditto.
5288 (vect_recog_widen_sum_pattern): Ditto.
5289 (vect_recog_over_widening_pattern): Ditto.
5290 (vect_recog_widen_shift_pattern): Ditto.
5291 (vect_recog_rotate_pattern): Ditto.
5292 (vect_recog_vector_vector_shift_pattern): Ditto.
5293 (vect_recog_divmod_pattern): Ditto.
5294 (vect_recog_mixed_size_cond_pattern): Ditto.
5295 (vect_recog_bool_pattern): Ditto.
5296 (vect_pattern_recog_1): Ditto.
5297 (vect_pattern_recog): Ditto.
5298 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
5299 (vect_is_simple_iv_evolution): Ditto.
5300 (vect_analyze_scalar_cycles_1): Ditto.
5301 (vect_get_loop_niters): Ditto.
5302 (vect_analyze_loop_1): Ditto.
5303 (vect_analyze_loop_form): Ditto.
5304 (vect_analyze_loop_operations): Ditto.
5305 (vect_analyze_loop_2): Ditto.
5306 (vect_analyze_loop): Ditto.
5307 (report_vect_op): Ditto.
5308 (vect_is_slp_reduction): Ditto.
5309 (vect_is_simple_reduction_1): Ditto.
5310 (vect_get_known_peeling_cost): Ditto.
5311 (vect_estimate_min_profitable_iters): Ditto.
5312 (vect_model_reduction_cost): Ditto.
5313 (vect_model_induction_cost): Ditto.
5314 (get_initial_def_for_induction): Ditto.
5315 (vect_create_epilog_for_reduction): Ditto.
5316 (vectorizable_reduction): Ditto.
5317 (vectorizable_induction): Ditto.
5318 (vectorizable_live_operation): Ditto.
5319 (vect_loop_kill_debug_uses): Ditto.
5320 (vect_transform_loop): Ditto.
5321 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
5322 (vect_stmt_relevant_p): Ditto.
5323 (process_use): Ditto.
5324 (vect_mark_stmts_to_be_vectorized): Ditto.
5325 (vect_model_simple_cost): Ditto.
5326 (vect_model_promotion_demotion_cost): Ditto.
5327 (vect_model_store_cost): Ditto.
5328 (vect_get_store_cost): Ditto.
5329 (vect_model_load_cost): Ditto.
5330 (vect_get_load_cost): Ditto.
5331 (vect_init_vector_1): Ditto.
5332 (vect_get_vec_def_for_operand): Ditto.
5333 (vect_finish_stmt_generation): Ditto.
5334 (vectorizable_call): Ditto.
5335 (vectorizable_conversion): Ditto.
5336 (vectorizable_assignment): Ditto.
5337 (vectorizable_shift): Ditto.
5338 (vectorizable_operation): Ditto.
5339 (vectorizable_store): Ditto.
5340 (vectorizable_load): Ditto.
5341 (vectorizable_condition): Ditto.
5342 (vect_analyze_stmt): Ditto.
5343 (vect_transform_stmt): Ditto.
5344 (vect_is_simple_use): Ditto.
5345 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
5346 (vect_can_advance_ivs_p): Ditto.
5347 (vect_update_ivs_after_vectorizer): Ditto.
5348 (vect_do_peeling_for_loop_bound): Ditto.
5349 (vect_gen_niters_for_prolog_loop): Ditto.
5350 (vect_update_inits_of_drs): Ditto.
5351 (vect_create_cond_for_alias_checks): Ditto.
5352 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
5353 (vect_build_slp_tree_1): Ditto.
5354 (vect_supported_load_permutation_p): Ditto.
5355 (vect_analyze_slp_instance): Ditto.
5356 (vect_analyze_slp): Ditto.
5357 (vect_make_slp_decision): Ditto.
5358 (vect_detect_hybrid_slp): Ditto.
5359 (vect_bb_vectorization_profitable_p): Ditto.
5360 (vect_slp_analyze_bb_1): Ditto.
5361 (vect_update_slp_costs_according_to_vf): Ditto.
5362 (vect_get_mask_element): Ditto.
5363 (vect_transform_slp_perm_load): Ditto.
5364 (vect_schedule_slp_instance): Ditto.
5365 (vect_schedule_slp): Ditto.
5366 (vect_slp_transform_bb): Ditto.
5367 * profile.c (read_profile_edge_counts): Ditto.
5368 (compute_branch_probabilities): Ditto.
5369 * coverage.c (get_coverage_counts): Ditto.
5371 2013-09-16 Diego Novillo <dnovillo@google.com>
5373 * tree-core.h: Add missing comment lines from refactoring of tree.h.
5375 2013-09-16 Jan Hubicka <jh@suse.cz>
5377 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
5378 abstract functions; for static functions check the presence of body.
5380 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5382 * config/aarch64/aarch64-simd-builtins.def (fma): New.
5383 * config/aarch64/aarch64-simd.md
5384 (aarch64_mla_elt<mode>): New.
5385 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
5386 (aarch64_mls_elt<mode>): Likewise.
5387 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
5388 (aarch64_fma4_elt<mode>): Likewise.
5389 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
5390 (aarch64_fma4_elt_to_128v2df): Likewise.
5391 (aarch64_fma4_elt_to_64df): Likewise.
5392 (fnma<mode>4): Likewise.
5393 (aarch64_fnma4_elt<mode>): Likewise.
5394 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
5395 (aarch64_fnma4_elt_to_128v2df): Likewise.
5396 (aarch64_fnma4_elt_to_64df): Likewise.
5397 * config/aarch64/iterators.md (VDQSF): New.
5398 * config/aarch64/arm_neon.h
5399 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
5400 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
5402 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5404 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
5405 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
5406 (aarch64_mul3_elt_to_128df): Likewise.
5407 (aarch64_mul3_elt_to_64v2df): Likewise.
5408 * config/aarch64/iterators.md (VEL): Also handle DFmode.
5410 (VMUL_CHANGE_NLANES) Likewise.
5413 * config/aarch64/arm_neon.h
5414 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
5416 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5418 * config/aarch64/arm_neon.h
5419 (vcvtx_high_f32_f64): Fix parameters.
5421 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5422 Uros Bizjak <ubizjak@gmail.com>
5424 * config/alpha.c: Include tree-ssanames.h.
5426 2013-09-16 Richard Biener <rguenther@suse.de>
5428 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
5429 (dot_rdg_1): Handle control_dd.
5430 (create_edge_for_control_dependence): New function.
5431 (create_rdg_edges): Add control dependences if asked for.
5432 (build_rdg): Likewise.
5433 (generate_loops_for_partition): If there are not necessary
5434 control stmts remove all their dependencies.
5435 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
5436 (distribute_loop): Pass on control dependences.
5437 (tree_loop_distribution): Compute control dependences and remove
5438 restriction on number of loop nodes.
5440 2013-09-16 Jakub Jelinek <jakub@redhat.com>
5442 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
5445 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
5447 * cse.c (try_const_anchors): Punt on CC modes.
5449 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5451 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
5453 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
5456 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
5459 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5462 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
5464 (darwin_objc1_section): Likewise.
5465 (darwin_file_end): Emit Image Info section when required.
5467 2013-09-14 Jan Hubicka <jh@suse.cz>
5469 * tree-into-ssa.c (gate_into_ssa): New.
5470 (pass_data_build_ssa): Use it.
5471 * cgraph.h (expand_thunk): Update prototype.
5472 * cgraphunit.c (analyze_function): Expand thunks early.
5473 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
5474 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
5475 set lowered flag; do not add new function.
5476 (assemble_thunks_and_aliases): Update.
5477 * tree-ssa.c (gate_init_datastructures): New gate.
5478 (pass_data_init_datastructures): Use it.
5480 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5483 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
5484 xmm register set described in the psABI.
5486 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
5488 * dwarf2out.c (should_emit_struct_debug): Add check
5489 for type_decl variable is not NULL.
5491 2013-09-13 Jacek Caban <jacek@codeweavers.com>
5493 * config.gcc: Use new winnt-c.c target hooks
5494 * config/t-winnt: New file
5495 * config/winnt-c.c: New file
5496 * doc/tm.texi.in: Document new hook
5497 * doc/tm.texi: Regenerated
5499 2013-09-13 Jan Hubicka <jh@suse.cz>
5502 * ipa-inline.c (check_callers): New function.
5503 (check_caller_edge): Remove.
5504 (want_inline_function_to_all_callers_p): Also permit alises that are
5506 (inline_to_all_callers): Terminate the walk when devirtualization
5507 introduce new calls.
5509 2013-09-13 Jan Hubicka <jh@suse.cz>
5511 * ipa-inline-analysis.c (struct growth_data): Add node.
5512 (do_estimate_growth_1): Fix detection of recursion.
5514 2013-09-13 Jakub Jelinek <jakub@redhat.com>
5516 PR tree-optimization/58392
5517 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
5518 to avoid shadowing of outer loop variable. If
5519 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
5520 replace_by_duplicate_decl simduid of loops that have it set and
5521 set dest_cfun->has_simduid_loops and/or
5522 dest_cfun->has_force_vect_loops.
5523 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
5524 instead of maybe_lookup_decl.
5525 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
5526 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
5527 call. Copy over force_vect and copy and remap simduid. Set
5528 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
5529 (copy_cfg_body): Remove blocks_to_copy argument. Use
5530 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
5531 caller. Don't set cfun->has_simduid_loops and/or
5532 cfun->has_force_vect_loops here.
5533 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
5535 (expand_call_inline, tree_function_versioning): Adjust copy_body
5538 2013-09-13 Martin Jambor <mjambor@suse.cz>
5541 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
5542 the assert if the edge was a speculative one.
5544 2013-09-13 Richard Biener <rguenther@suse.de>
5546 * tree-data-ref.h (known_dependences_p): Move here ...
5547 * tree-loop-distribution.c (known_dependences_p): ... from here.
5548 (dump_rdg_component, debug_rdg_component): Remove.
5550 (generate_loops_for_partition): Use gimple_uid instead of
5551 relying on matching stmt visit order.
5552 (rdg_build_partitions): Take starting stmt vector.
5553 (ldist_gen): Merge into ...
5554 (distribute_loop): ... this function. Do not compute starting
5556 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
5558 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5560 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
5561 Set type attribute correctly. Set predicable_short_it attribute.
5562 (cmpsi_shiftsi): Remove %? from output template.
5564 2013-09-13 Richard Biener <rguenther@suse.de>
5566 * tree-loop-distribution.c (struct rdg_component,
5567 rdg_defs_used_in_other_loops_p, free_rdg_components,
5568 rdg_build_components): Remove.
5569 (stmts_from_loop): Do not record virtual PHIs.
5570 (generate_loops_for_partition): Skip virtual PHIs.
5571 (build_rdg_partition_for_component): Rename to ...
5572 (build_rdg_partition_for_vertex): ... this and adjust.
5573 (rdg_build_partitions): Take a vector of starting vertices
5574 instead of components. Remove unnecessary leftover handling.
5575 (ldist_gen): Do not build components or record other stores.
5576 (distribute_loop): Do not distribute loops containing stmts
5579 2013-09-13 Christian Bruel <christian.bruel@st.com>
5582 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
5584 2013-09-13 Kai Tietz <ktietz@redhat.com>
5586 * config.gcc: Separate cases for mingw and cygwin targets,
5587 and add 64-bit cygwin target case.
5589 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
5590 dll-export inline-functions.
5591 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
5594 2013-09-13 Jeff Law <law@redhat.com>
5598 2013-09-06 Jeff Law <law@redhat.com>
5600 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5601 edge implied equivalences into successor phis.
5603 2013-09-12 DJ Delorie <dj@redhat.com>
5605 * config/rl78/rl78-virt.md: Change from | to \; for asm line
5608 2013-09-12 Brooks Moses <bmoses@google.com>
5611 * Makefile.in: Do not install driver binaries in $(target)/bin.
5613 2013-09-12 DJ Delorie <dj@redhat.com>
5615 * config/rl78/rl78.opt (mrelax): New.
5616 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
5617 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
5619 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
5621 (rl78_expand_epilogue): Likewise.
5623 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
5626 * lra-eliminations.c (remove_reg_equal_offset_note): New.
5627 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
5628 pointer elimination with using remove_reg_equal_offset_note.
5630 2013-09-12 DJ Delorie <dj@redhat.com>
5632 * config/msp430/: New port.
5633 * config.gcc (msp430): Added.
5634 * doc/invoke.texi: Document MSP430 options.
5635 * doc/install.texi: Document msp430-elf
5636 * doc/md.texi: Document msp430-elf
5637 * doc/contrib.texi: Document msp430-elf
5639 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
5642 2013-09-12 Martin Jambor <mjambor@suse.cz>
5645 * ipa-prop.c (remove_described_reference): Give up if the edge in the
5646 reference descriptor is NULL.
5647 (ipa_edge_removal_hook): If owning a reference descriptor, set its
5650 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5652 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
5653 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
5654 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
5655 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
5656 make_temp_ssa_name): move to tree-ssanames.h
5657 * tree-ssa-alias.h: Move prototype.
5658 * tree-ssa.h: Include tree-ssanames.h.
5659 * tree-ssanames.c (FREE_SSANAMES): Move to here.
5660 * tree-ssanames.h: New. Move items from tree-flow*.h
5661 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
5663 2013-09-12 Richard Biener <rguenther@suse.de>
5665 PR tree-optimization/58404
5666 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
5667 propagate non-invariant addresses into dereferences wrapped
5668 in component references.
5670 2013-09-12 Richard Biener <rguenther@suse.de>
5672 PR tree-optimization/58402
5673 * passes.def: Move pass_late_warn_uninitialized later.
5675 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5677 * tree-ssa.h: New. Move content from tree-flow.h and
5679 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
5680 Move prototypes belonging to tree-ssa.c.
5681 * tree-flow-inline.h (redirect_edge_var_map_def,
5682 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
5684 * gimple.h: Adjust prototypes.
5685 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
5687 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
5688 * tree.h: Move prototype to tree-ssa.h.
5689 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
5690 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
5691 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
5692 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
5693 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
5694 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
5695 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
5696 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
5697 graphite-blocking.c, graphite-clast-to-gimple.c,
5698 graphite-dependences.c, graphite-interchange.c,
5699 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
5700 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
5701 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
5702 ipa-reference.c, ipa-split.c, ipa-utils.c,
5703 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
5704 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
5705 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
5706 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5707 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
5708 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
5709 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
5710 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
5711 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
5712 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
5713 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
5714 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
5715 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
5716 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
5717 value-prof.c, var-tracking.c,
5718 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
5720 2013-09-12 Richard Biener <rguenther@suse.de>
5722 PR tree-optimization/58396
5723 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
5724 (build_rdg): Take a loop-nest parameter, fix memleaks.
5725 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
5727 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
5729 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
5732 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
5734 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
5735 The FMA instruction names should have a 'v' prefix.
5737 2013-09-12 Richard Biener <rguenther@suse.de>
5739 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
5740 (dot_rdg): Use popen instead of system in optional code.
5741 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
5742 (already_processed_vertex_p): Adjust.
5743 (has_anti_or_output_dependence, predecessor_has_mem_write,
5744 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
5745 rdg_flag_uses): Remove.
5746 (rdg_flag_vertex): Simplify.
5747 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
5749 (build_rdg_partition_for_component): Process the first vertex
5750 of a component only.
5751 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
5753 2013-09-12 Alan Modra <amodra@gmail.com>
5755 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
5757 2013-09-11 DJ Delorie <dj@redhat.com>
5758 Nick Clifton <nickc@redhat.com>
5760 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
5761 (rl78_stack_based_mem): New.
5762 * config/rl78/constraints.md (Iv08): New.
5771 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
5773 (movsi): Change from expand to insn-and-split.
5774 (ashrsi3): Clobber AX.
5778 * config/rl78/rl78.md (CC_REG): Fix.
5779 (addsi3): Allow memory and immediate operands.
5780 (addsi3_internal): Split into...
5781 (addsi3_internal_virt): ...new, and ...
5782 (addsi3_internal_real): ...new.
5784 (subsi3_internal_virt): New.
5785 (subsi3_internal_real): New.
5786 (mulsi3): Add memory operand.
5787 (mulsi3_rl78): Likewise.
5788 (mulsi3_g13): Likewise.
5789 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
5790 (cbranchqi4_real): Add more constraint options.
5791 (cbranchhi4_real): Expand pattern.
5792 (cbranchhi4_real_signed): New.
5793 (cbranchhi4_real_inverted): New.
5794 (cbranchsi4_real_lt): New.
5795 (cbranchsi4_real_ge): New.
5796 (cbranchsi4_real_signed): New.
5797 (cbranchsi4_real): New.
5799 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
5801 (lshrsi3_virt): Likewise.
5802 (ashlsi3_virt): Likewise.
5803 (cbranchqi4_virt_signed): New.
5804 (cbranchhi4_virt_signed): New.
5805 (cbranchsi4_virt): New.
5806 * config/rl78/rl78.c: Whitespace fixes throughout.
5807 (move_elim_pass): New.
5808 (pass_data_rl78_move_elim): New.
5809 (pass_rl78_move_elim): New.
5810 (make_pass_rl78_move_elim): New.
5811 (rl78_devirt_info): Run devirt earlier.
5812 (rl78_move_elim_info): New.
5813 (rl78_asm_file_start): Register it.
5814 (rl78_split_movsi): New.
5815 (rl78_as_legitimate_address): Allow virtual base registers when
5817 (rl78_addr_space_convert): Remove spurious debug stuff.
5818 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
5819 (rl78_print_operand): More cases for not printing '#'.
5820 (rl78_expand_compare): Remove most of the logic.
5821 (content_memory): New.
5822 (clear_content_memory): New.
5823 (get_content_index): New.
5824 (get_content_name): New.
5825 (display_content_memory): New.
5826 (update_content): New.
5827 (record_content): New.
5828 (already_contains): New.
5829 (insn_ok_now): Re-recog insns with virtual registers.
5830 (add_postponed_content_update): New.
5831 (process_postponed_content_update): New.
5832 (gen_and_emit_move): New.
5833 (transcode_memory_rtx): Record new location content.
5834 Use gen_and_emit_move.
5835 (force_into_acc): New.
5836 (move_to_acc): Use gen_and_emit_move.
5837 (move_from_acc): Likewise.
5838 (move_acc_to_reg): Likewise.
5839 (move_to_x): Likewise.
5840 (move_to_hl): Likewise.
5841 (move_to_de): Likewise.
5842 (rl78_alloc_physical_registers_op1): Record location content.
5843 (has_constraint): New.
5844 (rl78_alloc_physical_registers_op2): Record location content.
5846 (rl78_alloc_physical_registers_ro1): Likewise.
5847 (rl78_alloc_physical_registers_cmp): Likewise.
5848 (rl78_alloc_physical_registers_umul): Likewise.
5849 (rl78_alloc_address_registers_macax): New.
5850 (rl78_alloc_physical_registers): Initialize and set location
5851 content memory as needed.
5852 (rl78_reorg): Make sure split2 is called.
5853 (rl78_rtx_costs): New.
5855 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
5857 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
5858 for (not (neg ...)) and (neg (not ...)) cases.
5860 2013-09-11 Richard Biener <rguenther@suse.de>
5863 * passes.def: Split critical edges before late uninit warning passes.
5864 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
5866 2013-09-11 Jakub Jelinek <jakub@redhat.com>
5868 PR tree-optimization/58385
5869 * fold-const.c (build_range_check): If both low and high are NULL,
5870 use omit_one_operand_loc to preserve exp side-effects.
5872 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5874 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
5876 2013-09-11 Richard Biener <rguenther@suse.de>
5878 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
5879 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
5880 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
5881 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
5882 stmts_from_loop, known_dependences_p, build_empty_rdg,
5883 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
5884 * tree-loop-distribution.c: ... here.
5885 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
5886 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
5887 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
5888 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
5889 * tree-loop-distribution.c: ... here.
5890 * tree-loop-distribution.c: Include gimple-pretty-print.h.
5891 (struct partition_s): Add loops member.
5892 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
5893 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
5894 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
5896 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5897 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5898 Sergey Lega <sergey.s.lega@intel.com>
5899 Anna Tikhonova <anna.tikhonova@intel.com>
5900 Ilya Tocar <ilya.tocar@intel.com>
5901 Andrey Turetskiy <andrey.turetskiy@intel.com>
5902 Ilya Verbin <ilya.verbin@intel.com>
5903 Kirill Yukhin <kirill.yukhin@intel.com>
5904 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5906 * config/i386/constraints.md (k): New.
5908 * config/i386/i386.c (const regclass_map): Add new mask registers.
5909 (dbx_register_map): Ditto.
5910 (dbx64_register_map): Ditto.
5911 (svr4_dbx_register_map): Ditto.
5912 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
5914 (ix86_preferred_reload_class): Disable constants for mask registers.
5915 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
5916 (ix86_hard_regno_mode_ok): Support new mask registers.
5917 (x86_order_regs_for_local_alloc): Ditto.
5918 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
5919 (FIXED_REGISTERS): Add new mask registers.
5920 (CALL_USED_REGISTERS): Ditto.
5921 (REG_ALLOC_ORDER): Ditto.
5922 (VALID_MASK_REG_MODE): New.
5923 (FIRST_MASK_REG): Ditto.
5924 (LAST_MASK_REG): Ditto.
5925 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
5926 (MAYBE_MASK_CLASS_P): New.
5927 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
5928 (REG_CLASS_CONTENTS): Ditto.
5929 (MASK_REGNO_P): New.
5930 (ANY_MASK_REG_P): Ditto.
5931 (HI_REGISTER_NAMES): Add new mask registers.
5932 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
5933 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
5935 (attribute "type"): Add mskmov, msklog.
5936 (attribute "length_immediate"): Support them.
5937 (attribute "memory"): Ditto.
5938 (attribute "prefix_0f"): Ditto.
5939 (*movhi_internal): Support new mask registers.
5940 (*movqi_internal): Ditto.
5941 (define_split): Split out clobber pattern is a logic
5942 insn on mask registers.
5943 (*k<logic><mode>): New.
5944 (*andhi_1): Extend to support mask regs.
5945 (*andqi_1): Extend to support mask regs.
5947 (define_split): Split and-not to and and not if operands
5949 (*<code><mode>_1): Separate HI mode to new pattern...
5950 (*<code>hi_1): This.
5951 (*<code>qi_1): Extend to support mask regs.
5953 (kortestzhi): Ditto.
5954 (kortestchi): Ditto.
5956 (*one_cmpl<mode>2_1): Remove HImode and handle it...
5957 (*one_cmplhi2_1): ...Here, now with mask registers support.
5958 (*one_cmplqi2_1): Support new mask registers.
5959 (HI/QImode arithmetics splitter): Don't split if mask registers
5961 (HI/QImode not splitter): Ditto.
5962 * config/i386/predicated.md (mask_reg_operand): New.
5963 (general_reg_operand): Ditto.
5965 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5967 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
5968 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
5970 2013-09-10 Jeff Law <law@redhat.com>
5972 PR tree-optimization/58380
5973 * tree-ssa-threadupdate.c (thread_block): Recognize another case
5974 of threading through a buried loop header.
5976 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
5977 return value for single successor case.
5979 2013-09-10 Jan Hubicka <jh@suse.cz>
5981 * ipa-devirt.c (ipa_devirt): Enable with LTO.
5983 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
5986 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
5987 support conditional execution.
5988 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
5990 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
5992 * lra.c (lra): Clear lra_optional_reload_pseudos before every
5994 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
5995 Check destination too to check move insn.
5996 (undo_optional_reloads): Add check that the original peudo did not
5997 changed its allocation and the optional reload was inherited on last
5998 inheritance pass. Break loop after deciding to keep optional reload.
5999 (lra_undo_inheritance): Add check that inherited pseudo still in
6002 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
6004 * config/aarch64/aarch64.md (generic_sched): New.
6005 * config/aarch64/aarch64-generic.md (load): Make conditional
6006 on generic_sched attribute.
6007 (nonload): Likewise.
6009 2013-09-10 Jan Hubicka <jh@suse.cz>
6011 * lto-cgraph.c: Include ipa-utils.h.
6012 (compute_ltrans_boundary): Also add possible targets into the boundary.
6014 2013-09-10 Jan Hubicka <jh@suse.cz>
6016 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
6017 VAR_DECL of vtable rather than full expression.
6019 2013-09-10 Jan Hubicka <jh@suse.cz>
6020 Paolo Carlini <paolo.carlini@oracle.com>
6022 * cgraphunit.c (analyze_functions): Save input_location, set it
6023 to UNKNOWN_LOCATION and restore it at the end.
6025 2013-09-10 Martin Jambor <mjambor@suse.cz>
6027 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
6028 represented by a thunk.
6030 2013-09-10 Jeff Law <law@redhat.com>
6032 PR tree-optimization/58343
6033 * tree-ssa-threadupdate.c (thread_block): Identify and disable
6034 jump threading requests through loop headers buried in the middle
6035 of a jump threading path.
6037 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
6038 in return value/type.
6040 2013-09-10 Jakub Jelinek <jakub@redhat.com>
6042 PR rtl-optimization/58365
6043 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
6044 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
6047 2013-09-10 Richard Biener <rguenther@suse.de>
6049 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
6050 * tree-data-ref.c (create_rdg_vertices): Collect all data
6051 references, signal failure to the caller, use data-ref API.
6052 (build_rdg): Compute data references only once. Maintain lifetime
6053 of data references and data dependences from within RDG.
6054 (free_rdg): Free dependence relations.
6055 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
6056 inventing extra dependences.
6057 (distribute_loop): Update for RDG API changes.
6059 2013-09-10 Kai Tietz <ktietz@redhat.com>
6061 * doc/invoke.texi (fms-extensions): Document changed
6062 behavior for ms-abi targets.
6063 * config/i386/i386.c (ix86_option_override_internal):
6064 Set default value of option -fms-extension for ms-abi targets.
6066 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
6068 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
6070 2013-09-10 Alan Modra <amodra@gmail.com>
6073 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
6075 2013-09-10 Alan Modra <amodra@gmail.com>
6077 * config/rs6000/predicates.md (add_cint_operand): New.
6078 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
6079 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
6080 using add_cint_operand.
6081 (largetoc_high_plus_aix): Likewise.
6083 2013-09-09 Jakub Jelinek <jakub@redhat.com>
6085 PR tree-optimization/58364
6086 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
6087 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
6088 the current range can't be an unconditional true or false.
6090 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
6092 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
6094 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
6096 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
6098 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
6101 * doc/invoke.texi (-Wdelete-incomplete): Document it.
6103 2013-09-09 Ian Bolton <ian.bolton@arm.com>
6105 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
6106 NO_REGS for immediate that can't be moved directly into FP_REGS.
6108 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6110 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
6111 comparison with negated operand.
6112 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
6115 2013-09-09 Richard Biener <rguenther@suse.de>
6118 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
6119 of a subloop record all its block as affecting loop-closed SSA form.
6121 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6123 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
6124 of an rtx/bitpos pair.
6125 (store_fixed_bit_field): Update accordingly.
6127 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6129 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
6131 * builtins.c (expand_errno_check): Likewise.
6132 * dwarf2cfi.c (init_return_column_size): Likewise.
6133 * except.c (sjlj_mark_call_sites): Likewise.
6134 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
6135 * lra-constraints.c (emit_inc): Likewise.
6136 * ree.c (combine_set_extension): Likewise.
6137 * regmove.c (fixup_match_2): Likewise.
6138 * reload1.c (inc_for_reload): Likewise.
6140 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6142 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
6143 (force_to_mode, simplify_shift_const_1, simplify_comparison):
6144 Use gen_int_mode with the mode of the associated simplify_* call.
6145 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
6146 * expmed.c (expand_shift_1): Likewise.
6147 * function.c (instantiate_virtual_regs_in_insn): Likewise.
6148 * loop-iv.c (iv_number_of_iterations): Likewise.
6149 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
6150 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
6152 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6154 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
6155 of the associated expand_* call.
6156 (asan_emit_stack_protection): Likewise.
6157 * builtins.c (round_trampoline_addr): Likewise.
6158 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
6159 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
6160 (emit_store_flag): Likewise.
6161 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
6162 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
6164 * function.c (instantiate_virtual_regs_in_insn): Likewise.
6165 * ifcvt.c (noce_try_store_flag_constants): Likewise.
6166 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
6167 * modulo-sched.c (generate_prolog_epilog): Likewise.
6168 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
6169 (expand_ctz, expand_ffs, expand_unop): Likewise.
6171 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6173 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
6174 of the associated gen_rtx_* call.
6175 * caller-save.c (init_caller_save): Likewise.
6176 * combine.c (find_split_point, make_extraction): Likewise.
6177 (make_compound_operation): Likewise.
6178 * dwarf2out.c (mem_loc_descriptor): Likewise.
6179 * explow.c (plus_constant, probe_stack_range): Likewise.
6180 * expmed.c (expand_mult_const): Likewise.
6181 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
6182 * reload1.c (init_reload): Likewise.
6183 * valtrack.c (cleanup_auto_inc_dec): Likewise.
6184 * var-tracking.c (adjust_mems): Likewise.
6185 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
6186 rather than gen_rtx_fmt_ee.
6188 2013-09-09 Jan Hubicka <jh@suse.cz>
6191 * value-prof.c (gimple_ic): Copy also abnormal edges.
6193 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6195 * asan.c (asan_shadow_cst): Use gen_int_mode.
6197 2013-09-08 Jan Hubicka <jh@suse.cz>
6199 * ipa-profile.c: Add toplevel comment.
6200 (ipa_propagate_frequency_1): Be more conservative when profile is read.
6201 (contains_hot_call_p): New function.
6202 (ipa_propagate_frequency): Set frequencies based on counts when
6204 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
6205 profile; do not tamper with profile after inlining if it is read.
6207 2013-09-08 Jan Hubicka <jh@suse.cz>
6209 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
6212 2013-09-08 Jan Hubicka <jh@suse.cz>
6214 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
6217 2013-09-08 Jeff Law <law@redhat.com>
6220 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
6223 2013-09-08 Andi Kleen <ak@linux.intel.com>
6225 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
6227 2013-09-08 Jan Hubicka <jh@suse.cz>
6229 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
6231 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
6233 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
6234 for non-debug insns.
6235 * lra.c (new_insn_reg): Take the containing insn as a parameter.
6236 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
6237 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
6240 2013-09-08 Jan Hubicka <jh@suse.cz>
6242 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
6243 targets and devirtualize to BUILT_IN_UNREACHABLE.
6244 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
6245 * ipa.c (walk_polymorphic_call_targets): New function.
6246 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
6247 functions; use the new timevar.
6248 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
6249 was removed from the program.
6250 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
6251 not consider it in the walk when its vtable is dead.
6252 (possible_polymorphic_call_targets_1): Pass anonymous flag to
6254 (devirt_variable_node_removal_hook): New function.
6255 (possible_polymorphic_call_targets): Also register
6256 devirt_variable_node_removal_hook.
6257 (ipa_devirt): Do not do non-speculative devirtualization.
6258 (gate_ipa_devirt): One execute if devirtualizing speculatively.
6260 2013-09-08 Jan Hubicka <jh@suse.cz>
6262 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
6263 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
6264 varpool_remove_variable_insertion_hook): Declare.
6265 * varpool.c (varpool_node_hook_list): New structure.
6266 (first_varpool_node_removal_hook,
6267 first_varpool_variable_insertion_hook): New variables.
6268 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
6269 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
6270 varpool_remove_variable_insertion_hook,
6271 varpool_call_variable_insertion_hooks): New functions.
6272 (varpool_remove_node): Use it.
6274 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
6277 * diagnostic.c (diagnostic_build_prefix): When s.file is
6278 "<built-in>" don't output line and column numbers.
6280 2013-09-06 Jan Hubicka <jh@suse.cz>
6282 * cgraphunit.c (expand_thunk): Get body before touching arguments.
6283 * lto-streamer-out.c: Stream thunks, too.
6284 * lto-streamer-in.c (input_function): Pop cfun here
6285 (lto_read_body): Instead of here.
6287 2013-09-06 Caroline Tice <cmtice@google.com>
6289 * doc/install.texi: Add documentation for the --enable-vtable-verify
6290 and the --disable-libvtv configure options.
6292 2013-09-06 Jeff Law <law@redhat.com>
6294 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
6295 edge implied equivalences into successor phis.
6297 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
6299 * resource.c (mark_referenced_resources): Handle COND_EXEC.
6301 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
6303 * resource.c (mark_target_live_regs): Compute resources taking
6304 into account if a call is predicated or not.
6306 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6308 * toplev.c (output_stack_usage): Be prepared for suffixes created by
6309 the compiler in the function names.
6311 2013-09-06 Jan Hubicka <jh@suse.cz>
6314 * ipa-inline.c (has_caller_p): New function.
6315 (want_inline_function_to_all_callers_p): Use it.
6316 (sum_callers, inline_to_all_callers): Break out from ...
6317 (ipa_inline): ... here.
6319 2013-09-06 Jan Hubicka <jh@suse.cz>
6321 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
6322 only when AVX is enabled.
6324 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6326 * config/aarch64/aarch64.md
6327 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
6328 is fpsimd_<load/store>2.
6329 (load_pair<mode>): Likewise.
6330 (store_pair<mode>): Likewise.
6332 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6334 * config/arm/types.md (type): Add "mrs" type.
6335 * config/aarch64/aarch64.md
6336 (aarch64_load_tp_hard): Make type "mrs".
6338 (load_tp_hard): Make type "mrs".
6339 * config/arm/cortex-a15.md: Update with new attributes.
6340 * config/arm/cortex-a5.md: Update with new attributes.
6341 * config/arm/cortex-a53.md: Update with new attributes.
6342 * config/arm/cortex-a7.md: Update with new attributes.
6343 * config/arm/cortex-a8.md: Update with new attributes.
6344 * config/arm/cortex-a9.md: Update with new attributes.
6345 * config/arm/cortex-m4.md: Update with new attributes.
6346 * config/arm/cortex-r4.md: Update with new attributes.
6347 * config/arm/fa526.md: Update with new attributes.
6348 * config/arm/fa606te.md: Update with new attributes.
6349 * config/arm/fa626te.md: Update with new attributes.
6350 * config/arm/fa726te.md: Update with new attributes.
6352 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6354 * config/aarch64/aarch64.md
6355 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
6356 (*movtf_aarch64): Likewise.
6358 (thumb1_movdi_insn): Use "multiple" for type where more than one
6359 instruction is used for a move.
6360 (*arm32_movhf): Likewise.
6361 (*thumb_movdf_insn): Likewise.
6363 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6365 * config/arm/types.md (type): Rename fcpys to fmov.
6367 (*arm_movsi_vfp): Rename type fcpys as fmov.
6368 (*thumb2_movsi_vfp): Likewise
6369 (*movhf_vfp_neon): Likewise
6370 (*movhf_vfp): Likewise
6371 (*movsf_vfp): Likewise
6372 (*thumb2_movsf_vfp): Likewise
6373 (*movsfcc_vfp): Likewise
6374 (*thumb2_movsfcc_vfp): Likewise
6375 * config/aarch64/aarch64-simd.md
6376 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
6377 * config/aarch64/aarch64.md
6378 (*movsi_aarch64): Replace type mov_reg with fmovs.
6379 (*movdi_aarch64): Likewise
6380 (*movsf_aarch64): Likewise
6381 (*movdf_aarch64): Likewise
6383 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
6384 * config/arm/iwmmxt.md
6385 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
6386 * config/arm/arm1020e.md: Update with new attributes.
6387 * config/arm/cortex-a15-neon.md: Update with new attributes.
6388 * config/arm/cortex-a5.md: Update with new attributes.
6389 * config/arm/cortex-a53.md: Update with new attributes.
6390 * config/arm/cortex-a7.md: Update with new attributes.
6391 * config/arm/cortex-a8-neon.md: Update with new attributes.
6392 * config/arm/cortex-a9.md: Update with new attributes.
6393 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6394 * config/arm/cortex-r4f.md: Update with new attributes.
6395 * config/arm/marvell-pj4.md: Update with new attributes.
6396 * config/arm/vfp11.md: Update with new attributes.
6398 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6400 * config/aarch64/aarch64.md
6401 (*madd<mode>): Fix type attribute.
6402 (*maddsi_uxtw): Likewise.
6403 (*msub<mode>): Likewise.
6404 (*msubsi_uxtw): Likewise.
6405 (<su_optab>maddsidi4): Likewise.
6406 (<su_optab>msubsidi4): Likewise.
6408 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6410 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
6411 * config/arm/arm.md (core_cycles): Remove fdiv.
6412 * config/arm/vfp.md:
6413 (*sqrtsf2_vfp): Update for attribute changes.
6414 (*sqrtdf2_vfp): Likewise.
6415 * config/aarch64/aarch64.md:
6416 (sqrt<mode>2): Update for attribute changes.
6417 * config/arm/arm1020e.md: Update with new attributes.
6418 * config/arm/cortex-a15-neon.md: Update with new attributes.
6419 * config/arm/cortex-a5.md: Update with new attributes.
6420 * config/arm/cortex-a53.md: Update with new attributes.
6421 * config/arm/cortex-a7.md: Update with new attributes.
6422 * config/arm/cortex-a8-neon.md: Update with new attributes.
6423 * config/arm/cortex-a9.md: Update with new attributes.
6424 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6425 * config/arm/cortex-r4f.md: Update with new attributes.
6426 * config/arm/marvell-pj4.md: Update with new attributes.
6427 * config/arm/vfp11.md: Update with new attributes.
6429 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6431 * config/arm/types.md
6432 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
6433 * config/aarch64/aarch64.md
6434 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
6436 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
6437 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
6438 (float<GPI:mode><GPF:mode>2): Likewise.
6440 (*truncsisf2_vfp): Update with new attributes.
6441 (*truncsidf2_vfp): Likewise.
6442 (fixuns_truncsfsi2): Likewise.
6443 (fixuns_truncdfsi2): Likewise.
6444 (*floatsisf2_vfp): Likewise.
6445 (*floatsidf2_vfp): Likewise.
6446 (floatunssisf2): Likewise.
6447 (floatunssidf2): Likewise.
6448 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
6449 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
6450 * config/arm/arm1020e.md: Update with new attributes.
6451 * config/arm/cortex-a15-neon.md: Update with new attributes.
6452 * config/arm/cortex-a5.md: Update with new attributes.
6453 * config/arm/cortex-a53.md: Update with new attributes.
6454 * config/arm/cortex-a7.md: Update with new attributes.
6455 * config/arm/cortex-a8-neon.md: Update with new attributes.
6456 * config/arm/cortex-a9.md: Update with new attributes.
6457 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6458 * config/arm/cortex-r4f.md: Update with new attributes.
6459 * config/arm/marvell-pj4.md: Update with new attributes.
6460 * config/arm/vfp11.md: Update with new attributes.
6462 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6464 * config/aarch64/arm_neon.h
6465 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
6466 (vqtbx<1,2,3,4><q>_s8): Likewise.
6468 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6470 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
6471 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
6472 (add<mode>3): Add type attribute.
6473 (add<mode>3): Likewise.
6474 (usadd<mode>3): Likewise.
6475 (ssadd<mode>3): Likewise.
6476 (sub<mode>3): Likewise.
6477 (sub<mode>3): Likewise.
6478 (ussub<mode>3): Likewise.
6479 (sssub<mode>3): Likewise.
6480 (ssmulsa3): Likewise.
6481 (usmulusa3): Likewise.
6482 (arm_usatsihi): Likewise.
6484 (*movdi_vfp): Add types for all instructions.
6485 (*movdi_vfp_cortexa8): Likewise.
6486 (*movhf_vfp_neon): Likewise.
6487 (*movhf_vfp): Likewise.
6488 (*movdf_vfp): Likewise.
6489 (*thumb2_movdf_vfp): Likewise.
6490 (*thumb2_movdfcc_vfp): Likewise.
6491 * config/arm/arm.md: Add type attribute to all insn patterns.
6492 (*thumb1_adddi3): Add type attribute.
6493 (*arm_adddi3): Likewise.
6494 (*adddi_sesidi_di): Likewise.
6495 (*adddi_zesidi_di): Likewise.
6496 (*thumb1_addsi3): Likewise.
6497 (addsi3_compare0): Likewise.
6498 (*addsi3_compare0_scratch): Likewise.
6499 (*compare_negsi_si): Likewise.
6500 (cmpsi2_addneg): Likewise.
6501 (*addsi3_carryin_<optab>): Likewise.
6502 (*addsi3_carryin_alt2_<optab>): Likewise.
6503 (*addsi3_carryin_clobercc_<optab>): Likewise.
6504 (*subsi3_carryin): Likewise.
6505 (*subsi3_carryin_const): Likewise.
6506 (*subsi3_carryin_compare): Likewise.
6507 (*subsi3_carryin_compare_const): Likewise.
6508 (*arm_subdi3): Likewise.
6509 (*thumb_subdi3): Likewise.
6510 (*subdi_di_zesidi): Likewise.
6511 (*subdi_di_sesidi): Likewise.
6512 (*subdi_zesidi_di): Likewise.
6513 (*subdi_sesidi_di): Likewise.
6514 (*subdi_zesidi_ze): Likewise.
6515 (thumb1_subsi3_insn): Likewise.
6516 (*arm_subsi3_insn): Likewise.
6517 (*anddi3_insn): Likewise.
6518 (*anddi_zesidi_di): Likewise.
6519 (*anddi_sesdi_di): Likewise.
6520 (*ne_zeroextracts): Likewise.
6521 (*ne_zeroextracts): Likewise.
6522 (*ite_ne_zeroextr): Likewise.
6523 (*ite_ne_zeroextr): Likewise.
6524 (*anddi_notdi_di): Likewise.
6525 (*anddi_notzesidi): Likewise.
6526 (*anddi_notsesidi): Likewise.
6527 (andsi_notsi_si): Likewise.
6528 (thumb1_bicsi3): Likewise.
6529 (*iordi3_insn): Likewise.
6530 (*iordi_zesidi_di): Likewise.
6531 (*iordi_sesidi_di): Likewise.
6532 (*thumb1_iorsi3_insn): Likewise.
6533 (*xordi3_insn): Likewise.
6534 (*xordi_zesidi_di): Likewise.
6535 (*xordi_sesidi_di): Likewise.
6536 (*arm_xorsi3): Likewise.
6537 (*andsi_iorsi3_no): Likewise.
6538 (*smax_0): Likewise.
6539 (*smax_m1): Likewise.
6540 (*arm_smax_insn): Likewise.
6541 (*smin_0): Likewise.
6542 (*arm_smin_insn): Likewise.
6543 (*arm_umaxsi3): Likewise.
6544 (*arm_uminsi3): Likewise.
6545 (*minmax_arithsi): Likewise.
6546 (*minmax_arithsi_): Likewise.
6547 (*satsi_<SAT:code>): Likewise.
6548 (arm_ashldi3_1bit): Likewise.
6549 (arm_ashrdi3_1bit): Likewise.
6550 (arm_lshrdi3_1bit): Likewise.
6551 (*arm_negdi2): Likewise.
6552 (*thumb1_negdi2): Likewise.
6553 (*arm_negsi2): Likewise.
6554 (*thumb1_negsi2): Likewise.
6555 (*negdi_extendsid): Likewise.
6556 (*negdi_zero_extend): Likewise.
6557 (*arm_abssi2): Likewise.
6558 (*thumb1_abssi2): Likewise.
6559 (*arm_neg_abssi2): Likewise.
6560 (*thumb1_neg_abss): Likewise.
6561 (one_cmpldi2): Likewise.
6562 (extend<mode>di2): Likewise.
6563 (*compareqi_eq0): Likewise.
6564 (*arm_extendhisi2addsi): Likewise.
6565 (*arm_movdi): Likewise.
6566 (*thumb1_movdi_insn): Likewise.
6567 (*arm_movt): Likewise.
6568 (*thumb1_movsi_insn): Likewise.
6569 (pic_add_dot_plus_four): Likewise.
6570 (pic_add_dot_plus_eight): Likewise.
6571 (tls_load_dot_plus_eight): Likewise.
6572 (*thumb1_movhi_insn): Likewise.
6573 (*thumb1_movsf_insn): Likewise.
6574 (*movdf_soft_insn): Likewise.
6575 (*thumb_movdf_insn): Likewise.
6576 (cbranchsi4_insn): Likewise.
6577 (cbranchsi4_scratch): Likewise.
6578 (*negated_cbranchsi4): Likewise.
6579 (*tbit_cbranch): Likewise.
6580 (*tlobits_cbranch): Likewise.
6581 (*tstsi3_cbranch): Likewise.
6582 (*cbranchne_decr1): Likewise.
6583 (*addsi3_cbranch): Likewise.
6584 (*addsi3_cbranch_scratch): Likewise.
6585 (*arm_cmpdi_insn): Likewise.
6586 (*arm_cmpdi_unsig): Likewise.
6587 (*arm_cmpdi_zero): Likewise.
6588 (*thumb_cmpdi_zero): Likewise.
6589 (*deleted_compare): Likewise.
6590 (*mov_scc): Likewise.
6591 (*mov_negscc): Likewise.
6592 (*mov_notscc): Likewise.
6593 (*cstoresi_eq0_thumb1_insn): Likewise.
6594 (cstoresi_nltu_thumb1): Likewise.
6595 (cstoresi_ltu_thu): Likewise.
6596 (thumb1_addsi3_addgeu): Likewise.
6597 (*arm_jump): Likewise.
6598 (*thumb_jump): Likewise.
6599 (*check_arch2): Likewise.
6600 (arm_casesi_internal): Likewise.
6601 (thumb1_casesi_dispatch): Likewise.
6602 (*arm_indirect_jump): Likewise.
6603 (*thumb1_indirect_jump): Likewise.
6605 (*and_scc): Likewise.
6606 (*ior_scc): Likewise.
6607 (*compare_scc): Likewise.
6608 (*cond_move): Likewise.
6609 (*cond_arith): Likewise.
6610 (*cond_sub): Likewise.
6611 (*cmp_ite0): Likewise.
6612 (*cmp_ite1): Likewise.
6613 (*cmp_and): Likewise.
6614 (*cmp_ior): Likewise.
6615 (*ior_scc_scc): Likewise.
6616 (*ior_scc_scc_cmp): Likewise.
6617 (*and_scc_scc): Likewise.
6618 (*and_scc_scc_cmp): Likewise.
6619 (*and_scc_scc_nod): Likewise.
6620 (*negscc): Likewise.
6621 (movcond_addsi): Likewise.
6622 (movcond): Likewise.
6623 (*ifcompare_plus_move): Likewise.
6624 (*if_plus_move): Likewise.
6625 (*ifcompare_move_plus): Likewise.
6626 (*if_move_plus): Likewise.
6627 (*ifcompare_arith_arith): Likewise.
6628 (*if_arith_arith): Likewise.
6629 (*ifcompare_arith_move): Likewise.
6630 (*if_arith_move): Likewise.
6631 (*ifcompare_move_arith): Likewise.
6632 (*if_move_arith): Likewise.
6633 (*ifcompare_move_not): Likewise.
6634 (*if_move_not): Likewise.
6635 (*ifcompare_not_move): Likewise.
6636 (*if_not_move): Likewise.
6637 (*ifcompare_shift_move): Likewise.
6638 (*if_shift_move): Likewise.
6639 (*ifcompare_move_shift): Likewise.
6640 (*if_move_shift): Likewise.
6641 (*ifcompare_shift_shift): Likewise.
6642 (*ifcompare_not_arith): Likewise.
6643 (*ifcompare_arith_not): Likewise.
6644 (*if_arith_not): Likewise.
6645 (*ifcompare_neg_move): Likewise.
6646 (*if_neg_move): Likewise.
6647 (*ifcompare_move_neg): Likewise.
6648 (*if_move_neg): Likewise.
6649 (prologue_thumb1_interwork): Likewise.
6650 (*cond_move_not): Likewise.
6651 (*sign_extract_onebit): Likewise.
6652 (*not_signextract_onebit): Likewise.
6653 (stack_tie): Likewise.
6654 (align_4): Likewise.
6655 (align_8): Likewise.
6656 (consttable_end): Likewise.
6657 (consttable_1): Likewise.
6658 (consttable_2): Likewise.
6659 (consttable_4): Likewise.
6660 (consttable_8): Likewise.
6661 (consttable_16): Likewise.
6662 (*thumb1_tablejump): Likewise.
6663 (prefetch): Likewise.
6664 (force_register_use): Likewise.
6665 (thumb_eh_return): Likewise.
6666 (load_tp_hard): Likewise.
6667 (load_tp_soft): Likewise.
6668 (tlscall): Likewise.
6669 (*arm_movtas_ze): Likewise.
6670 (*arm_rev): Likewise.
6671 (*arm_revsh): Likewise.
6672 (*arm_rev16): Likewise.
6673 * config/arm/thumb2.md
6674 (*thumb2_smaxsi3): Likewise.
6675 (*thumb2_sminsi3): Likewise.
6676 (*thumb32_umaxsi3): Likewise.
6677 (*thumb2_uminsi3): Likewise.
6678 (*thumb2_negdi2): Likewise.
6679 (*thumb2_abssi2): Likewise.
6680 (*thumb2_neg_abss): Likewise.
6681 (*thumb2_movsi_insn): Likewise.
6682 (tls_load_dot_plus_four): Likewise.
6683 (*thumb2_movhi_insn): Likewise.
6684 (*thumb2_mov_scc): Likewise.
6685 (*thumb2_mov_negs): Likewise.
6686 (*thumb2_mov_negs): Likewise.
6687 (*thumb2_mov_nots): Likewise.
6688 (*thumb2_mov_nots): Likewise.
6689 (*thumb2_movsicc_): Likewise.
6690 (*thumb2_movsfcc_soft_insn): Likewise.
6691 (*thumb2_indirect_jump): Likewise.
6692 (*thumb2_and_scc): Likewise.
6693 (*thumb2_ior_scc): Likewise.
6694 (*thumb2_ior_scc_strict_it): Likewise.
6695 (*thumb2_cond_move): Likewise.
6696 (*thumb2_cond_arith): Likewise.
6697 (*thumb2_cond_ari): Likewise.
6698 (*thumb2_cond_sub): Likewise.
6699 (*thumb2_negscc): Likewise.
6700 (*thumb2_movcond): Likewise.
6701 (thumb2_casesi_internal): Likewise.
6702 (thumb2_casesi_internal_pic): Likewise.
6703 (*thumb2_alusi3_short): Likewise.
6704 (*thumb2_mov<mode>_shortim): Likewise.
6705 (*thumb2_addsi_short): Likewise.
6706 (*thumb2_subsi_short): Likewise.
6707 (thumb2_addsi3_compare0): Likewise.
6708 (*thumb2_cbz): Likewise.
6709 (*thumb2_cbnz): Likewise.
6710 (*thumb2_one_cmplsi2_short): Likewise.
6711 (*thumb2_negsi2_short): Likewise.
6712 (*orsi_notsi_si): Likewise.
6713 * config/arm/arm1020e.md: Update with new attributes.
6714 * config/arm/arm1026ejs.md: Update with new attributes.
6715 * config/arm/arm1136jfs.md: Update with new attributes.
6716 * config/arm/arm926ejs.md: Update with new attributes.
6717 * config/arm/cortex-a15.md: Update with new attributes.
6718 * config/arm/cortex-a5.md: Update with new attributes.
6719 * config/arm/cortex-a53.md: Update with new attributes.
6720 * config/arm/cortex-a7.md: Update with new attributes.
6721 * config/arm/cortex-a8.md: Update with new attributes.
6722 * config/arm/cortex-a9.md: Update with new attributes.
6723 * config/arm/cortex-m4.md: Update with new attributes.
6724 * config/arm/cortex-r4.md: Update with new attributes.
6725 * config/arm/fa526.md: Update with new attributes.
6726 * config/arm/fa606te.md: Update with new attributes.
6727 * config/arm/fa626te.md: Update with new attributes.
6728 * config/arm/fa726te.md: Update with new attributes.
6730 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6732 * config/aarch64/aarch64-simd.md
6733 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
6734 <vwx> iterator to ensure correct register choice.
6735 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
6736 (aarch64_sqdmull_n<mode>): Likewise.
6737 (aarch64_sqdmull2_n<mode>_internal): Likewise.
6738 * config/aarch64/arm_neon.h
6739 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
6740 (vml<as><q>_n_<su>16): Likewise.
6741 (vml<as>l_high_lane<q>_<su>16): Likewise.
6742 (vml<as>l_high_n_<su>16): Likewise.
6743 (vml<as>l_lane<q>_<su>16): Likewise.
6744 (vml<as>l_n_<su>16): Likewise.
6745 (vmul<q>_lane<q>_<su>16): Likewise.
6746 (vmul<q>_n_<su>16): Likewise.
6747 (vmull_lane<q>_<su>16): Likewise.
6748 (vmull_n_<su>16): Likewise.
6749 (vmull_high_lane<q>_<su>16): Likewise.
6750 (vmull_high_n_<su>16): Likewise.
6751 (vqrdmulh<q>_n_s16): Likewise.
6753 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
6755 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
6756 have the correct lane parameter.
6758 2013-09-06 Richard Biener <rguenther@suse.de>
6760 * cfganal.c (control_dependences::~control_dependences):
6761 Properly free all of the vector.
6763 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
6766 * config/i386/i386.c (ix86_conditional_register_usage):
6767 Proper initialize extended SSE registers.
6769 2013-09-06 Jan Hubicka <jh@suse.cz>
6771 PR tree-optimization/58311
6772 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
6774 2013-09-06 Jan Hubicka <jh@suse.cz>
6776 * Makefile.in (tree-sra.o): Update dependencies.
6777 * tree-sra.c: Include ipa-utils.h
6778 (scan_function): Use recursive_call_p.
6779 (has_caller_p): New function.
6780 (cgraph_for_node_and_aliases): Count also callers of aliases.
6782 2013-09-06 Jan Hubicka <jh@suse.cz>
6785 * cgraph.h (symtab_semantically_equivalent_p): Declare.
6786 * tree-tailcall.c: Include ipa-utils.h.
6787 (find_tail_calls): Use it.
6788 * ipa-pure-const.c (check_call): Likewise.
6789 * ipa-utils.c (recursive_call_p): New function.
6790 * ipa-utils.h (recursive_call_p): Dclare.
6791 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
6792 (symtab_semantically_equivalent_p): New function.
6793 * Makefile.in (tree-tailcall.o): Update dependencies.
6795 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6797 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
6800 2013-09-06 Richard Biener <rguenther@suse.de>
6802 * lto-streamer.h (lto_global_var_decls): Remove.
6803 * Makefile.in (OBJS): Remove lto-symtab.o.
6804 (lto-symtab.o): Remove.
6805 (GTFILES): Remove lto-symtab.c
6806 * lto-symtab.c: Move to lto/
6808 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6810 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
6811 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
6812 (UNSPEC_FPINT_RINT): New constant definitions.
6813 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
6814 definition with 2 attributes.
6815 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
6816 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
6819 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6821 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
6822 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
6823 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
6824 attribute to "z196_alone".
6825 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
6828 2013-09-06 Richard Biener <rguenther@suse.de>
6830 * basic-block.h (class control_dependences): New.
6831 * tree-ssa-dce.c (control_dependence_map): Remove.
6833 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
6834 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
6835 find_pdom, find_control_dependence, find_all_control_dependences):
6837 (mark_control_dependent_edges_necessary,
6838 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
6839 tree_dce_done, perform_tree_ssa_dce): Adjust.
6840 * cfganal.c (set_control_dependence_map_bit,
6841 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
6842 find_all_control_dependences): Move from tree-ssa-dce.c and
6843 implement as methods of control_dependences class.
6844 (control_dependences::control_dependences): New.
6845 (control_dependences::~control_dependences): Likewise.
6846 (control_dependences::get_edges_dependent_on): Likewise.
6847 (control_dependences::get_edge): Likewise.
6849 2013-09-04 Jan Hubicka <jh@suse.cz>
6851 * tree.c (types_same_for_odr): Drop overactive check.
6852 * ipa-devirt.c (hash_type_name): Likewise.
6854 2013-09-04 Jan Hubicka <jh@suse.cz>
6856 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
6857 (analyze_functions): ... here.
6859 2013-09-04 Jan Hubicka <jh@suse.cz>
6862 * cgraphunit.c (analyze_functions): Clear AUX fields
6863 after processing; initialize assembler name has.
6865 2013-09-05 Jeff Law <law@redhat.com>
6867 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
6868 from thread_around_empty_block. Record threading path into PATH.
6869 Recurse if threading through the initial block is successful.
6870 (thread_across_edge): Corresponding changes to slightly simplify.
6872 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6874 * config/aarch64/aarch64.md
6875 (type): Remove frecpe, frecps, frecpx.
6876 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
6877 fix to be a TARGET_SIMD instruction.
6878 (aarch64_frecps): Remove.
6879 * config/aarch64/aarch64-simd.md
6880 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
6881 (aarch64_frecps<mode>): Handle all float/vector of float modes.
6883 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6884 Sofiane Naci <sofiane.naci@arm.com>
6886 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
6887 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
6888 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
6889 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
6890 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
6891 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
6892 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
6893 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
6894 into "clz, "rbit". Rename "shift" to "shift_imm".
6895 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
6896 changes. Update for attribute changes all occurrences of arlo_* and
6898 * config/arm/arm-fixed.md: Update for attribute changes
6899 all occurrences of arlo_* types.
6900 * config/arm/thumb2.md: Update for attribute changes all occurrences
6902 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
6903 (cortexa7_older_only): Likewise.
6904 (cortexa7_younger): Likewise.
6905 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
6906 (1020alu_shift_op): Likewise.
6907 (1020alu_shift_reg_op): Likewise.
6908 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
6909 (alu_shift_op): Likewise.
6910 (alu_shift_reg_op): Likewise.
6911 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
6912 (11_alu_shift_op): Likewise.
6913 (11_alu_shift_reg_op): Likewise.
6914 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
6915 (9_alu_shift_reg_op): Likewise.
6916 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
6918 (cortex_a15_alu_shift): Likewise.
6919 (cortex_a15_alu_shift_reg): Likewise.
6920 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
6922 (cortex_a5_alu_shift): Likewise.
6923 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
6925 (cortex_a53_alu_shift): Likewise.
6926 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
6928 (cortex_a7_alu_reg): Likewise.
6929 (cortex_a7_alu_shift): Likewise.
6930 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
6932 (cortex_a8_alu_shift): Likewise.
6933 (cortex_a8_alu_shift_reg): Likewise.
6934 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
6935 (cortex_a9_dp_shift): Likewise.
6936 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
6938 * config/arm/cortex-r4.md
6939 (cortex_r4_alu): Update for attribute changes.
6940 (cortex_r4_mov): Likewise.
6941 (cortex_r4_alu_shift_reg): Likewise.
6942 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
6943 (526_alu_shift_op): Likewise.
6944 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
6945 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
6946 (626te_alu_shift_op): Likewise.
6947 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
6948 (726te_alu_shift_op): Likewise.
6949 (726te_alu_shift_reg_op): Likewise.
6950 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
6951 (mp626_alu_shift_op): Likewise.
6952 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
6953 (pj4_alu_conds): Likewise.
6954 (pj4_shift): Likewise.
6955 (pj4_shift_conds): Likewise.
6956 (pj4_alu_shift): Likewise.
6957 (pj4_alu_shift_conds): Likewise.
6958 * config/aarch64/aarch64.md: Update for attribute change
6959 all occurrences of arlo_* and shift* types.
6961 2013-09-05 Mike Stump <mikestump@comcast.net>
6963 * tree.h: Move documentation for tree_function_decl to tree-core.h
6964 with the declaration.
6966 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
6969 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
6970 looking for widest mode.
6972 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
6974 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
6976 2013-09-05 Richard Biener <rguenther@suse.de>
6978 PR tree-optimization/58137
6979 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
6980 Do not create vectors of pointers.
6981 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
6982 types for the components of the vector initializer.
6983 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
6984 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
6986 2013-09-05 Martin Jambor <mjambor@suse.cz>
6988 * ipa-prop.c (remove_described_reference): Accept missing references,
6989 return false if that hppens, otherwise return true.
6990 (cgraph_node_for_jfunc): New function.
6991 (try_decrement_rdesc_refcount): Likewise.
6992 (try_make_edge_direct_simple_call): Use them.
6993 (ipa_edge_removal_hook): Remove references from rdescs.
6994 (ipa_edge_duplication_hook): Clone rdescs and their references
6995 when the new edge has the same caller as the old one.
6996 * cgraph.c (cgraph_resolve_speculation): Remove speculative
6997 reference before removing any edges.
6999 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
7001 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
7003 * thumb2.md (thumb2_storewb_parisi): New pattern.
7005 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
7007 * config/aarch64/aarch64-option-extensions.def: Add
7008 AARCH64_OPT_EXTENSION of 'crc'.
7009 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
7010 (AARCH64_ISA_CRC): Ditto.
7011 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
7012 description of the CRC extension.
7014 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
7016 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
7017 * config/rs6000/linux.h: Ditto.
7018 * alpha/linux.h: Ditto.
7019 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
7020 no_c99_libc_has_function.
7021 * config/c6x/uclinux-elf.h: Ditto.
7022 * config/lm32/uclinux-elf.h: Ditto.
7023 * config/m68k/uclinux.h: Ditto.
7024 * config/moxie/uclinux.h: Ditto.
7025 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
7026 (crisv32-*-linux*, cris-*-linux*): Ditto.
7027 * config/bfin/bfin.c: Include "tm_p.h".
7029 2013-09-05 Richard Biener <rguenther@suse.de>
7031 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
7032 check for a definition without a basic-block.
7034 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
7035 Sofiane Naci <sofiane.naci@arm.com>
7037 * config/aarch64/aarch64.md
7038 (*movti_aarch64): Rename r_2_f and f_2_r.
7039 (*movsf_aarch64): Likewise.
7040 (*movdf_aarch64): Likewise.
7041 (*movtf_aarch64): Likewise.
7042 (aarch64_movdi_<mode>low): Likewise.
7043 (aarch64_movdi_<mode>high): Likewise.
7044 (aarch64_mov<mode>high_di): Likewise.
7045 (aarch64_mov<mode>low_di): Likewise.
7046 (aarch64_movtilow_tilow): Likewise.
7047 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
7048 values to config/arm/types.md
7049 (attribute "conds"): Update for attribute change.
7050 (anddi3_insn): Likewise.
7051 (iordi3_insn): Likewise.
7052 (xordi3_insn): Likewise.
7053 (one_cmpldi2): Likewise.
7054 * config/arm/types.md (type): Add Neon types.
7055 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
7056 use "type" attribute.
7057 (movmisalign<mode>_neon_store): Likewise.
7058 (movmisalign<mode>_neon_load): Likewise.
7059 (vec_set<mode>_internal): Likewise.
7060 (vec_setv2di_internal): Likewise.
7061 (vec_extract<mode>): Likewise.
7062 (vec_extractv2di): Likewise.
7063 (add<mode>3_neon): Likewise.
7064 (adddi3_neon): Likewise.
7065 (sub<mode>3_neon): Likewise.
7066 (subdi3_neon): Likewise.
7067 (mul<mode>3_neon): Likewise.
7068 (mul<mode>3add<mode>_neon): Likewise.
7069 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
7070 (fma<VCVTF:mode>4)): Likewise.
7071 (fma<VCVTF:mode>4_intrinsic): Likewise.
7072 (fmsub<VCVTF:mode>4)): Likewise.
7073 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
7074 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
7075 (ior<mode>3): Likewise.
7076 (and<mode>3): Likewise.
7077 (anddi3_neon): Likewise.
7078 (orn<mode>3_neon): Likewise.
7079 (orndi3_neon): Likewise.
7080 (bic<mode>3_neon): Likewise.
7081 (bicdi3_neon): Likewise.
7082 (xor<mode>3): Likewise.
7083 (one_cmpl<mode>2): Likewise.
7084 (abs<mode>2): Likewise.
7085 (neg<mode>2): Likewise.
7086 (umin<mode>3_neon): Likewise.
7087 (umax<mode>3_neon): Likewise.
7088 (smin<mode>3_neon): Likewise.
7089 (smax<mode>3_neon): Likewise.
7090 (vashl<mode>3): Likewise.
7091 (vashr<mode>3_imm): Likewise.
7092 (vlshr<mode>3_imm): Likewise.
7093 (ashl<mode>3_signed): Likewise.
7094 (ashl<mode>3_unsigned): Likewise.
7095 (neon_load_count): Likewise.
7096 (ashldi3_neon_noclobber): Likewise.
7097 (signed_shift_di3_neon): Likewise.
7098 (unsigned_shift_di3_neon): Likewise.
7099 (ashrdi3_neon_imm_noclobber): Likewise.
7100 (lshrdi3_neon_imm_noclobber): Likewise.
7101 (widen_ssum<mode>3): Likewise.
7102 (widen_usum<mode>3): Likewise.
7103 (quad_halves_<code>v4si): Likewise.
7104 (quad_halves_<code>v4sf): Likewise.
7105 (quad_halves_<code>v8hi): Likewise.
7106 (quad_halves_<code>v16qi): Likewise.
7107 (reduc_splus_v2di): Likewise.
7108 (neon_vpadd_internal<mode>): Likewise.
7109 (neon_vpsmin<mode>): Likewise.
7110 (neon_vpsmax<mode>): Likewise.
7111 (neon_vpumin<mode>): Likewise.
7112 (neon_vpumax<mode>): Likewise.
7113 (ss_add<mode>_neon): Likewise.
7114 (us_add<mode>_neon): Likewise.
7115 (ss_sub<mode>_neon): Likewise.
7116 (us_sub<mode>_neon): Likewise.
7117 (neon_vadd<mode>_unspec): Likewise.
7118 (neon_vaddl<mode>): Likewise.
7119 (neon_vaddw<mode>): Likewise.
7120 (neon_vhadd<mode>): Likewise.
7121 (neon_vqadd<mode>): Likewise.
7122 (neon_vaddhn<mode>): Likewise.
7123 (neon_vmul<mode>): Likewise.
7124 (neon_vmla<mode>): Likewise.
7125 (neon_vmlal<mode>): Likewise.
7126 (neon_vmls<mode>): Likewise.
7127 (neon_vmlsl<mode>): Likewise.
7128 (neon_vqdmulh<mode>): Likewise.
7129 (neon_vqdmlal<mode>): Likewise.
7130 (neon_vqdmlsl<mode>): Likewise.
7131 (neon_vmull<mode>): Likewise.
7132 (neon_vqdmull<mode>): Likewise.
7133 (neon_vsub<mode>_unspec): Likewise.
7134 (neon_vsubl<mode>): Likewise.
7135 (neon_vsubw<mode>): Likewise.
7136 (neon_vqsub<mode>): Likewise.
7137 (neon_vhsub<mode>): Likewise.
7138 (neon_vsubhn<mode>): Likewise.
7139 (neon_vceq<mode>): Likewise.
7140 (neon_vcge<mode>): Likewise.
7141 (neon_vcgeu<mode>): Likewise.
7142 (neon_vcgt<mode>): Likewise.
7143 (neon_vcgtu<mode>): Likewise.
7144 (neon_vcle<mode>): Likewise.
7145 (neon_vclt<mode>): Likewise.
7146 (neon_vcage<mode>): Likewise.
7147 (neon_vcagt<mode>): Likewise.
7148 (neon_vtst<mode>): Likewise.
7149 (neon_vabd<mode>): Likewise.
7150 (neon_vabdl<mode>): Likewise.
7151 (neon_vaba<mode>): Likewise.
7152 (neon_vabal<mode>): Likewise.
7153 (neon_vmax<mode>): Likewise.
7154 (neon_vmin<mode>): Likewise.
7155 (neon_vpaddl<mode>): Likewise.
7156 (neon_vpadal<mode>): Likewise.
7157 (neon_vpmax<mode>): Likewise.
7158 (neon_vpmin<mode>): Likewise.
7159 (neon_vrecps<mode>): Likewise.
7160 (neon_vrsqrts<mode>): Likewise.
7161 (neon_vqabs<mode>): Likewise.
7162 (neon_vqneg<mode>): Likewise.
7163 (neon_vcls<mode>): Likewise.
7164 (clz<mode>2): Likewise.
7165 (popcount<mode>2): Likewise.
7166 (neon_vrecpe): Likewise.
7167 (neon_vrsqrte): Likewise.
7168 (neon_vget_lane<mode>_sext_internal): Likewise.
7169 (neon_vget_lane<mode>_zext_internal): Likewise.
7170 (neon_vdup_n<mode>): Likewise.
7171 (neon_vdup_nv2di): Likewise.
7172 (neon_vdpu_lane<mode>_internal): Likewise.
7173 (neon_vswp<mode>): Likewise.
7174 (float<mode><V_cvtto>2): Likewise.
7175 (floatuns<mode><V_cvtto>2): Likewise.
7176 (fix_trunc<mode><V_cvtto>)2): Likewise
7177 (fixuns_trunc<mode><V_cvtto)2): Likewise.
7178 (neon_vcvt<mode>): Likewise.
7179 (neon_vcvtv4sfv4hf): Likewise.
7180 (neon_vcvtv4hfv4sf): Likewise.
7181 (neon_vcvt_n<mode>): Likewise.
7182 (neon_vmovn<mode>): Likewise.
7183 (neon_vqmovn<mode>): Likewise.
7184 (neon_vqmovun<mode>): Likewise.
7185 (neon_vmovl<mode>): Likewise.
7186 (neon_vmul_lane<mode>): Likewise.
7187 (neon_vmull_lane<mode>): Likewise.
7188 (neon_vqdmull_lane<mode>): Likewise.
7189 (neon_vqdmulh_lane<mode>): Likewise.
7190 (neon_vmla_lane<mode>): Likewise.
7191 (neon_vmlal_lane<mode>): Likewise.
7192 (neon_vqdmlal_lane<mode>): Likewise.
7193 (neon_vmls_lane<mode>): Likewise.
7194 (neon_vmlsl_lane<mode>): Likewise.
7195 (neon_vqdmlsl_lane<mode>): Likewise.
7196 (neon_vext<mode>): Likewise.
7197 (neon_vrev64<mode>): Likewise.
7198 (neon_vrev32<mode>): Likewise.
7199 (neon_vrev16<mode>): Likewise.
7200 (neon_vbsl<mode>_internal): Likewise.
7201 (neon_vshl<mode>): Likewise.
7202 (neon_vqshl<mode>): Likewise.
7203 (neon_vshr_n<mode>): Likewise.
7204 (neon_vshrn_n<mode>): Likewise.
7205 (neon_vqshrn_n<mode>): Likewise.
7206 (neon_vqshrun_n<mode>): Likewise.
7207 (neon_vshl_n<mode>): Likewise.
7208 (neon_vqshl_n<mode>): Likewise.
7209 (neon_vqshlu_n<mode>): Likewise.
7210 (neon_vshll_n<mode>): Likewise.
7211 (neon_vsra_n<mode>): Likewise.
7212 (neon_vsri_n<mode>): Likewise.
7213 (neon_vsli_n<mode>): Likewise.
7214 (neon_vtbl1v8qi): Likewise.
7215 (neon_vtbl2v8qi): Likewise.
7216 (neon_vtbl3v8qi): Likewise.
7217 (neon_vtbl4v8qi): Likewise.
7218 (neon_vtbx1v8qi): Likewise.
7219 (neon_vtbx2v8qi): Likewise.
7220 (neon_vtbx3v8qi): Likewise.
7221 (neon_vtbx4v8qi): Likewise.
7222 (neon_vtrn<mode>_internal): Likewise.
7223 (neon_vzip<mode>_internal): Likewise.
7224 (neon_vuzp<mode>_internal): Likewise.
7225 (neon_vld1<mode>): Likewise.
7226 (neon_vld1_lane<mode>): Likewise.
7227 (neon_vld1_dup<mode>): Likewise.
7228 (neon_vld1_dupv2di): Likewise.
7229 (neon_vst1<mode>): Likewise.
7230 (neon_vst1_lane<mode>): Likewise.
7231 (neon_vld2<mode>): Likewise.
7232 (neon_vld2_lane<mode>): Likewise.
7233 (neon_vld2_dup<mode>): Likewise.
7234 (neon_vst2<mode>): Likewise.
7235 (neon_vst2_lane<mode>): Likewise.
7236 (neon_vld3<mode>): Likewise.
7237 (neon_vld3qa<mode>): Likewise.
7238 (neon_vld3qb<mode>): Likewise.
7239 (neon_vld3_lane<mode>): Likewise.
7240 (neon_vld3_dup<mode>): Likewise.
7241 (neon_vst3<mode>): Likewise.
7242 (neon_vst3qa<mode>): Likewise.
7243 (neon_vst3qb<mode>): Likewise.
7244 (neon_vst3_lane<mode>): Likewise.
7245 (neon_vld4<mode>): Likewise.
7246 (neon_vld4qa<mode>): Likewise.
7247 (neon_vld4qb<mode>): Likewise.
7248 (neon_vld4_lane<mode>): Likewise.
7249 (neon_vld4_dup<mode>): Likewise.
7250 (neon_vst4<mode>): Likewise.
7251 (neon_vst4qa<mode>): Likewise.
7252 (neon_vst4qb<mode>): Likewise.
7253 (neon_vst4_lane<mode>): Likewise.
7254 (neon_vec_unpack<US>_lo_<mode>): Likewise.
7255 (neon_vec_unpack<US>_hi_<mode>): Likewise.
7256 (neon_vec_<US>mult_lo_<mode>): Likewise.
7257 (neon_vec_<US>mult_hi_<mode>): Likewise.
7258 (neon_vec_<US>shiftl_<mode>): Likewise.
7259 (neon_unpack<US>_<mode>): Likewise.
7260 (neon_vec_<US>mult_<mode>): Likewise.
7261 (vec_pack_trunc_<mode>): Likewise.
7262 (neon_vec_pack_trunk_<mode>): Likewise.
7263 (neon_vabd<mode>_2): Likewise.
7264 (neon_vabd<mode>_3): Likewise.
7265 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
7266 (thumb2_movsi_vfp): Likewise.
7267 (movdi_vfp): Likewise.
7268 (movdi_vfp_cortexa8): Likewise.
7269 (movhf_vfp_neon): Likewise.
7270 (movhf_vfp): Likewiwse.
7271 (movsf_vfp): Likewiwse.
7272 (thumb2_movsf_vfp): Likewiwse.
7273 (movdf_vfp): Likewise.
7274 (thumb2_movdf_vfp): Likewise.
7275 (movsfcc_vfp): Likewise.
7276 (thumb2_movsfcc_vfp): Likewise.
7277 (movdfcc_vfp): Likewise.
7278 (thumb2_movdfcc_vfp): Likewise.
7279 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
7280 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
7281 (v10_v2c): Likewise.
7282 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
7284 (cortex_a15_neon_int_2): Likewise.
7285 (cortex_a15_neon_int_3): Likewise.
7286 (cortex_a15_neon_int_4): Likewise.
7287 (cortex_a15_neon_int_5): Likewise.
7288 (cortex_a15_neon_vqneg_vqabs): Likewise.
7289 (cortex_a15_neon_vmov): Likewise.
7290 (cortex_a15_neon_vaba): Likewise.
7291 (cortex_a15_neon_vaba_qqq): Likewise.
7292 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7293 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7294 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
7295 scalar_64_32_long_scalar): Likewise.
7296 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7297 (cortex_a15_neon_mla_qqq_8_16): Likewise.
7298 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7299 lotype_qdd_64_32_long): Likewise.
7300 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7301 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7302 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
7303 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7304 (cortex_a15_neon_shift_1): Likewise.
7305 (cortex_a15_neon_shift_2): Likewise.
7306 (cortex_a15_neon_shift_3): Likewise.
7307 (cortex_a15_neon_vshl_ddd): Likewise.
7308 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7309 (cortex_a15_neon_vsra_vrsra): Likewise.
7310 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
7311 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
7312 (cortex_a15_neon_fp_vmul_ddd): Likewise.
7313 (cortex_a15_neon_fp_vmul_qqd): Likewise.
7314 (cortex_a15_neon_fp_vmla_ddd): Likewise.
7315 (cortex_a15_neon_fp_vmla_qqq): Likewise.
7316 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
7317 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
7318 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7319 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7320 (cortex_a15_neon_bp_simple): Likewise.
7321 (cortex_a15_neon_bp_2cycle): Likewise.
7322 (cortex_a15_neon_bp_3cycle): Likewise.
7323 (cortex_a15_neon_vld1_1_2_regs): Likewise.
7324 (cortex_a15_neon_vld1_3_4_regs): Likewise.
7325 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7326 (cortex_a15_neon_vld2_4_regs): Likewise.
7327 (cortex_a15_neon_vld3_vld4): Likewise.
7328 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7329 (cortex_a15_neon_vst1_3_4_regs): Likewise.
7330 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
7331 (cortex_a15_neon_vst3_vst4): Likewise.
7332 (cortex_a15_neon_vld1_vld2_lane): Likewise.
7333 (cortex_a15_neon_vld3_vld4_lane" 10
7334 (cortex_a15_neon_vst1_vst2_lane): Likewise.
7335 (cortex_a15_neon_vst3_vst4_lane): Likewise.
7336 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
7337 (cortex_a15_neon_ldm_2): Likewise.0
7338 (cortex_a15_neon_stm_2): Likewise.
7339 (cortex_a15_neon_mcr): Likewise.
7340 (cortex_a15_neon_mcr_2_mcrr): Likewise.
7341 (cortex_a15_neon_mrc): Likewise.
7342 (cortex_a15_neon_mrrc): Likewise.
7343 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
7345 (cortex_a15_alu_shift): Likewise.
7346 (cortex_a15_alu_shift_reg): Likewise.
7347 (cortex_a15_mult32): Likewise.
7348 (cortex_a15_mult64): Likewise.
7349 (cortex_a15_block): Likewise.
7350 (cortex_a15_branch): Likewise.
7351 (cortex_a15_load1): Likewise.
7352 (cortex_a15_load3): Likewise.
7353 (cortex_a15_store1): Likewise.
7354 (cortex_a15_store3): Likewise.
7355 (cortex_a15_call): Likewise.
7356 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
7357 (cortex_a5_f2r): Likewise.
7358 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
7360 (cortex_a53_f2r): Likewise.
7361 * config/arm/cortex-a7.md
7362 (cortex_a7_branch): Update for attribute change.
7363 (cortex_a7_call): Likewise.
7364 (cortex_a7_alu_imm): Likewise.
7365 (cortex_a7_alu_reg): Likewise.
7366 (cortex_a7_alu_shift): Likewise.
7367 (cortex_a7_mul): Likewise.
7368 (cortex_a7_load1): Likewise.
7369 (cortex_a7_store1): Likewise.
7370 (cortex_a7_load2): Likewise.
7371 (cortex_a7_store2): Likewise.
7372 (cortex_a7_load3): Likewise.
7373 (cortex_a7_store3): Likewise.
7374 (cortex_a7_load4): Likewise.
7375 (cortex_a7_store4): Likewise.
7376 (cortex_a7_fpalu): Likewise.
7377 (cortex_a7_fconst): Likewise.
7378 (cortex_a7_fpmuls): Likewise.
7379 (cortex_a7_neon_mul): Likewise.
7380 (cortex_a7_fpmacs): Likewise.
7381 (cortex_a7_neon_mla: Likewise.
7382 (cortex_a7_fpmuld: Likewise.
7383 (cortex_a7_fpmacd: Likewise.
7384 (cortex_a7_fpfmad: Likewise.
7385 (cortex_a7_fdivs: Likewise.
7386 (cortex_a7_fdivd: Likewise.
7387 (cortex_a7_r2f: Likewise.
7388 (cortex_a7_f2r: Likewise.
7389 (cortex_a7_f_flags: Likewise.
7390 (cortex_a7_f_loads: Likewise.
7391 (cortex_a7_f_loadd: Likewise.
7392 (cortex_a7_f_stores: Likewise.
7393 (cortex_a7_f_stored: Likewise.
7394 (cortex_a7_neon): Likewise.
7395 * config/arm/cortex-a8-neon.md
7396 (cortex_a8_neon_mrc): Update for attribute change.
7397 (cortex_a8_neon_mrrc): Likewise.
7398 (cortex_a8_neon_int_1): Likewise.
7399 (cortex_a8_neon_int_2): Likewise.
7400 (cortex_a8_neon_int_3): Likewise.
7401 (cortex_a8_neon_int_4): Likewise.
7402 (cortex_a8_neon_int_5): Likewise.
7403 (cortex_a8_neon_vqneg_vqabs): Likewise.
7404 (cortex_a8_neon_vmov): Likewise.
7405 (cortex_a8_neon_vaba): Likewise.
7406 (cortex_a8_neon_vaba_qqq): Likewise.
7407 (cortex_a8_neon_vsma): Likewise.
7408 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7409 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7410 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7411 long_scalar): Likewise.
7412 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7413 (cortex_a8_neon_mla_qqq_8_16): Likewise.
7414 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7415 long_scalar_qdd_64_32_long): Likewise.
7416 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7417 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7418 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
7419 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7420 (cortex_a8_neon_shift_1): Likewise.
7421 (cortex_a8_neon_shift_2): Likewise.
7422 (cortex_a8_neon_shift_3): Likewise.
7423 (cortex_a8_neon_vshl_ddd): Likewise.
7424 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7425 (cortex_a8_neon_vsra_vrsra): Likewise.
7426 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
7427 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
7428 (cortex_a8_neon_fp_vsum): Likewise.
7429 (cortex_a8_neon_fp_vmul_ddd): Likewise.
7430 (cortex_a8_neon_fp_vmul_qqd): Likewise.
7431 (cortex_a8_neon_fp_vmla_ddd): Likewise.
7432 (cortex_a8_neon_fp_vmla_qqq): Likewise.
7433 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
7434 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
7435 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7436 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7437 (cortex_a8_neon_bp_simple): Likewise.
7438 (cortex_a8_neon_bp_2cycle): Likewise.
7439 (cortex_a8_neon_bp_3cycle): Likewise.
7440 (cortex_a8_neon_ldr): Likewise.
7441 (cortex_a8_neon_str): Likewise.
7442 (cortex_a8_neon_vld1_1_2_regs): Likewise.
7443 (cortex_a8_neon_vld1_3_4_regs): Likewise.
7444 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7445 (cortex_a8_neon_vld2_4_regs): Likewise.
7446 (cortex_a8_neon_vld3_vld4): Likewise.
7447 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7448 (cortex_a8_neon_vst1_3_4_regs): Likewise.
7449 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
7450 (cortex_a8_neon_vst3_vst4): Likewise.
7451 (cortex_a8_neon_vld1_vld2_lane): Likewise.
7452 (cortex_a8_neon_vld3_vld4_lane): Likewise.
7453 (cortex_a8_neon_vst1_vst2_lane): Likewise.
7454 (cortex_a8_neon_vst3_vst4_lane): Likewise.
7455 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
7456 (cortex_a8_neon_mcr): Likewise.
7457 (cortex_a8_neon_mcr_2_mcrr): Likewise.
7458 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
7459 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
7461 (ca9_neon_mrrc): Likewise.
7462 (cortex_a9_neon_int_1): Likewise.
7463 (cortex_a9_neon_int_2): Likewise.
7464 (cortex_a9_neon_int_3): Likewise.
7465 (cortex_a9_neon_int_4): Likewise.
7466 (cortex_a9_neon_int_5): Likewise.
7467 (cortex_a9_neon_vqneg_vqabs): Likewise.
7468 (cortex_a9_neon_vmov): Likewise.
7469 (cortex_a9_neon_vaba): Likewise.
7470 (cortex_a9_neon_vaba_qqq): Likewise.
7471 (cortex_a9_neon_vsma): Likewise.
7472 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7473 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7474 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7475 long_scalar): Likewise.
7476 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7477 (cortex_a9_neon_mla_qqq_8_16): Likewise.
7478 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7479 long_scalar_qdd_64_32_long): Likewise.
7480 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7481 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7482 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
7483 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7484 (cortex_a9_neon_shift_1): Likewise.
7485 (cortex_a9_neon_shift_2): Likewise.
7486 (cortex_a9_neon_shift_3): Likewise.
7487 (cortex_a9_neon_vshl_ddd): Likewise.
7488 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7489 (cortex_a9_neon_vsra_vrsra): Likewise.
7490 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
7491 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
7492 (cortex_a9_neon_fp_vsum): Likewise.
7493 (cortex_a9_neon_fp_vmul_ddd): Likewise.
7494 (cortex_a9_neon_fp_vmul_qqd): Likewise.
7495 (cortex_a9_neon_fp_vmla_ddd): Likewise.
7496 (cortex_a9_neon_fp_vmla_qqq): Likewise.
7497 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
7498 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
7499 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7500 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7501 (cortex_a9_neon_bp_simple): Likewise.
7502 (cortex_a9_neon_bp_2cycle): Likewise.
7503 (cortex_a9_neon_bp_3cycle): Likewise.
7504 (cortex_a9_neon_ldr): Likewise.
7505 (cortex_a9_neon_str): Likewise.
7506 (cortex_a9_neon_vld1_1_2_regs): Likewise.
7507 (cortex_a9_neon_vld1_3_4_regs): Likewise.
7508 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7509 (cortex_a9_neon_vld2_4_regs): Likewise.
7510 (cortex_a9_neon_vld3_vld4): Likewise.
7511 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7512 (cortex_a9_neon_vst1_3_4_regs): Likewise.
7513 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
7514 (cortex_a9_neon_vst3_vst4): Likewise.
7515 (cortex_a9_neon_vld1_vld2_lane): Likewise.
7516 (cortex_a9_neon_vld3_vld4_lane): Likewise.
7517 (cortex_a9_neon_vst1_vst2_lane): Likewise.
7518 (cortex_a9_neon_vst3_vst4_lane): Likewise.
7519 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
7520 (cortex_a9_neon_mcr): Likewise.
7521 (cortex_a9_neon_mcr_2_mcrr): Likewise.
7522 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
7523 (cortex_a9_fps): Likewise.
7524 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
7526 (cortex_m4_fmuls): Likewise.
7527 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
7529 (cortex_r4_mrc): Likewise.
7530 * config/arm/iterators.md: Update comment referring to neon_type.
7531 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
7532 (iwmmxt_movsi_insn): Likewise.
7533 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
7535 (pj4_core_to_vfp): Likewise.
7536 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
7538 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
7539 (vfp_fstore): Likewise.
7540 * doc/md.texi: Change references to neon_type to refer to type.
7542 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7544 * tree.h (DECL_BUILT_IN): Fix typo in comment.
7546 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
7548 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
7551 2013-09-04 Easwaran Raman <eraman@google.com>
7554 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
7555 (build_and_add_sum): Use it.
7556 (appears_later_in_bb): Simplify code.
7558 2013-09-04 Teresa Johnson <tejohnson@google.com>
7560 * dumpfile.c (dump_finish): Don't close stderr/stdout.
7562 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
7564 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
7566 2013-09-04 Jan Hubicka <jh@suse.cz>
7568 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
7569 * ipa-devirt.c: Include diganostic.h
7570 (odr_type_d): Add types and types_set.
7571 (hash_type_name): Work for types with vtables during LTO.
7572 (odr_hasher::remove): Fix comment; destroy types_set.
7573 (add_type_duplicate): New function,
7574 (get_odr_type): Use it.
7575 (dump_type_inheritance_graph): Dump type duplicates.
7576 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
7578 * tree.c (types_same_for_odr): Give exact answers on types with
7581 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7583 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
7584 explaining their differences.
7586 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
7588 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
7590 2013-09-03 Jeff Law <law@redhat.com>
7592 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
7593 when not threading through a joiner block. Pass joiner/no joiner
7594 state to register_jump_thread.
7595 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
7596 state from argument rather than implying on path length.
7597 Dump the entire jump thread path into debugging dump.
7598 * tree-flow.h (register_jump_thread): Update prototype.
7600 2013-08-29 Xinliang David Li <davidxl@google.com>
7602 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7603 Remove a trivial gcc_assert.
7605 2013-08-29 Xinliang David Li <davidxl@google.com>
7607 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
7608 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
7609 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7610 Delay base decl alignment adjustment.
7611 * tree-vectorizer.c (vect_destroy_datarefs): New function.
7612 * tree-vectorizer.h: New data structure.
7613 (set_dr_misalignment): New function.
7614 (dr_misalignment): Ditto.
7615 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
7616 (vectorizable_load): Ditto.
7617 (ensure_base_align): New function.
7618 (vectorize_loops): Add dbg_cnt support.
7619 (execute_vect_slp): Ditto.
7620 * dbgcnt.def: New debug counter.
7621 * Makefile: New dependency.
7623 2013-09-03 Meador Inge <meadori@codesourcery.com>
7627 2013-08-30 Meador Inge <meadori@codesourcery.com>
7629 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7631 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
7633 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
7634 function descriptor.
7636 2013-09-03 Richard Biener <rguenther@suse.de>
7638 * tree-affine.c (add_elt_to_tree): Fix association issue,
7639 avoid useless converts and make sure to always return a
7640 properly typed result.
7642 2013-09-03 Richard Biener <rguenther@suse.de>
7645 * fold-const.c (negate_expr_p): Fix division case.
7646 (negate_expr): Likewise.
7648 2013-09-03 Richard Biener <rguenther@suse.de>
7651 * tree-streamer-out.c: Include tm.h.
7652 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
7654 2013-09-03 Jan Hubicka <jh@suse.cz>
7656 * tree-profile.c (tree_profiling): Cleanup CFG when done.
7658 2013-09-03 Alan Modra <amodra@gmail.com>
7660 * config.gcc (powerpc*-*-linux*): Add support for little-endian
7661 multilibs to big-endian target and vice versa.
7662 * config/rs6000/t-linux64: Use := assignment on all vars.
7663 (MULTILIB_EXTRA_OPTS): Remove fPIC.
7664 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
7665 * config/rs6000/t-linux64le: New file.
7666 * config/rs6000/t-linux64bele: New file.
7667 * config/rs6000/t-linux64lebe: New file.
7669 2013-09-02 Jan Hubicka <jh@suse.cz>
7671 * ipa-inline-transform.c (inline_transform): Do not
7672 optimize_inline_calls when not optimizing.
7674 2013-09-02 Jan Hubicka <jh@suse.cz>
7676 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
7677 duplicated nodes for assembler names.
7678 * symtab.c (symtab_unregister_node): Do not attempt to unlink
7679 hard registers from assembler name hash.
7681 2013-09-02 Jan Hubicka <jh@suse.cz>
7683 * ipa-split.c (execute_split_functions): Split externally visible
7684 functions called once.
7686 2013-09-02 Martin Jambor <mjambor@suse.cz>
7689 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
7690 linked list. When finding the correct duplicate, also consider also
7691 the caller in additon to its inlined_to node.
7693 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
7695 * config/aarch64/aarch64-simd-builtins.def
7696 (dup_lane_scalar): Remove.
7697 * config/aarch64/aarch64-simd.md
7698 (aarch64_simd_dup): Add 'w->w' alternative.
7699 (aarch64_dup_lane<mode>): Allow for VALL.
7700 (aarch64_dup_lane_scalar<mode>): Remove.
7701 (aarch64_dup_lane_<vswap_width_name><mode>): New.
7702 (aarch64_get_lane_signed<mode>): Add w->w altenative.
7703 (aarch64_get_lane_unsigned<mode>): Likewise.
7704 (aarch64_get_lane<mode>): Likewise.
7705 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
7706 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
7707 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
7708 (VCON): Change container of V2SF.
7709 (vswap_width_name): Likewise.
7710 * config/aarch64/arm_neon.h
7711 (__aarch64_vdup_lane_any): New.
7712 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7713 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
7714 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7716 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
7719 * expr.c (emit_move_complex): Do not move complex FP values as parts if
7720 the source or the destination is a single hard register.
7722 2013-09-02 Richard Biener <rguenther@suse.de>
7725 * tree-scalar-evolution.c (instantiate_scev_name): Allow
7728 2013-09-02 Richard Biener <rguenther@suse.de>
7730 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
7731 arithmetic to sizetype.
7733 2013-09-02 Bin Cheng <bin.cheng@arm.com>
7735 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
7736 Find auto-increment use both before and after candidate.
7738 2013-09-02 Marek Polacek <polacek@redhat.com>
7740 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
7742 2013-09-01 Jan Hubicka <jh@suse.cz>
7744 * Makefile.in: Add ipa-profile.o
7745 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
7746 * cgraph.c (struct cgraph_propagate_frequency_data,
7747 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
7748 ipa-profile.c; replace cgraph_ by ipa_ prefix.
7749 * cgraph.h (cgraph_propagate_frequency): Remove.
7750 * ipa-inline-analysis.c: Include ipa-utils.h;
7751 drop duplicated cfgloop.h.
7752 (inline_update_callee_summaries): Update.
7753 * ipa-profile.c: New file.
7754 * ipa-utils.h (ipa_propagate_frequency): Declare.
7755 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
7756 data-streamer.h, value-prof.h.
7757 (symtab_remove_unreachable_nodes): Update profile.
7758 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
7759 account_time_size, cmp_counts, dump_histogram,
7760 ipa_profile_generate_summary, ipa_profile_write_summary,
7761 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
7762 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
7763 Move to ipa-profile.c.
7765 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
7767 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
7769 2013-09-01 Jan Hubicka <jh@suse.cz>
7771 * common.opt (fdevirtualize-speculatively): New function.
7772 * invoke.texi (fdevirtualize-speculatively): Document.
7773 * ipa-devirt.c: Include ipa-inline.h
7774 (likely_target_p): New function.
7775 (ipa_devirt): New function.
7776 (gate_ipa_devirt): New function.
7777 (pass_data_ipa_devirt): New static var.
7778 (pass_ipa_devirt): Likewise.
7779 (make_pass_ipa_devirt): New function.
7780 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
7781 (common_handle_option): Disable devirtualization when
7782 value range profiling is available.
7783 * passes.def (pass_ipa_devirt): Add.
7784 * timever.def (TV_IPA_DEVIRT): New timevar.
7785 * tree-pass.h (make_pass_ipa_devirt):
7787 2013-09-01 Iain Sandoe <iain@codesourcery.com>
7789 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
7790 include sanitize(undefined).
7792 2013-08-31 Diego Novillo <dnovillo@google.com>
7794 * Makefile.in (TREE_CORE_H): Define.
7796 (GTFILES): Add tree-core.h.
7797 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
7799 * tree-core.h: New file.
7800 Move all data structures, enum, typedefs, global
7801 declarations and constants from ...
7804 2013-08-31 Jan Hubicka <jh@suse.cz>
7806 * bulitins.c (expand_builtin): Do not early exit for gcov
7807 instrumented functions.
7809 2013-08-31 Marek Polacek <polacek@redhat.com>
7811 * ubsan.c: Include tm_p.h.
7813 2013-08-31 Jan Hubicka <jh@suse.cz>
7815 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
7818 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
7819 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
7820 * tree-cfg.c (verify_gimple_label): ... here.
7821 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
7822 (ipa_merge_profiles): New function.
7823 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
7824 (lto_input_function_body): Likewise.
7825 * ipa-utils.h (ipa_merge_profiles): Declare.
7826 * lto-streamer.h (lto_input_function_body): Update prototype.
7827 (emit_label_in_global_context_p): Remove.
7828 * lto-symtab.c: Include ipa-utils.h
7829 (lto_cgraph_replace_node): Use ipa_merge_profiles.
7831 2013-08-31 Jan Hubicka <jh@suse.cz>
7833 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
7835 2013-08-31 Jan Hubicka <jh@suse.cz>
7837 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
7839 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
7841 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
7842 "cmp" RTX before signed_comparison_operator check to account
7845 2013-08-30 Jan Hubicka <jh@suse.cz>
7847 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
7848 (detect_type_change_1): Rename to ...
7849 (detect_type_change): ... this one; early return on non-polymorphic
7851 (detect_type_change_ssa): Add comp_type parameter; update
7852 use of detect_type_change.
7853 (compute_complex_assign_jump_func): Add param_type parameter;
7854 update use of detect_type_change_ssa.
7855 (compute_complex_ancestor_jump_func): Likewise.
7856 (ipa_get_callee_param_type): New function.
7857 (ipa_compute_jump_functions_for_edge): Compute parameter type;
7858 update calls to the jump function computation functions.
7860 2013-08-30 Teresa Johnson <tejohnson@google.com>
7861 Steven Bosscher <steven@gcc.gnu.org>
7863 * cfgrtl.c (fixup_new_cold_bb): New routine.
7864 (commit_edge_insertions): Invoke fixup_partitions.
7865 (find_partition_fixes): New routine.
7866 (fixup_partitions): Ditto.
7867 (verify_hot_cold_block_grouping): Update comments.
7868 (rtl_verify_edges): Invoke find_partition_fixes.
7869 (rtl_verify_bb_pointers): Update comments.
7870 (rtl_verify_bb_layout): Ditto.
7871 * basic-block.h (probably_never_executed_edge_p): Declare.
7872 (fixup_partitions): Ditto.
7873 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
7874 * bb-reorder.c (sanitize_hot_paths): New function.
7875 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
7877 * predict.c (probably_never_executed_edge_p): New routine.
7878 * cfg.c (check_bb_profile): Add partition insanity warnings.
7880 2013-08-30 Meador Inge <meadori@codesourcery.com>
7882 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7884 2013-08-30 Marek Polacek <polacek@redhat.com>
7886 * Makefile.in (ubsan.o): Add.
7887 (c-family/c-ubsan.o): Add.
7888 (builtins.o): Add ubsan.h dependency.
7889 * ubsan.h: New file.
7890 * ubsan.c: New file.
7891 * common.opt: Add -fsanitize=undefined option.
7892 (flag_sanitize): Add variable.
7893 (fsanitize=): Add option. Add Driver.
7894 (fsanitize=thread): Remove option.
7895 (fsanitize=address): Likewise.
7896 (static-libubsan): New option.
7897 * doc/invoke.texi: Document the new flag and -static-libubsan.
7898 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
7899 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
7900 * builtin-attrs.def (ATTR_COLD): Define.
7901 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7902 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
7903 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
7904 * flag-types.h (sanitize_code): New enum.
7905 * opts.c (common_handle_option): Parse command line arguments
7906 of -fsanitize=. Add -fsanitize=unreachable option.
7907 * varasm.c (get_variable_section): Adjust.
7908 (assemble_noswitch_variable): Likewise.
7909 (assemble_variable): Likewise.
7910 (output_constant_def_contents): Likewise.
7911 (categorize_decl_for_section): Likewise.
7912 (place_block_symbol): Likewise.
7913 (output_object_block): Likewise.
7914 * builtins.def: Likewise.
7915 * toplev.c (compile_file): Likewise.
7916 (process_options): Likewise.
7917 * cppbuiltin.c: Likewise.
7918 * tsan.c (tsan_pass): Likewise.
7919 (tsan_gate): Likewise.
7920 (tsan_gate_O0): Likewise.
7921 * cfgexpand.c (partition_stack_vars): Likewise.
7922 (expand_stack_vars): Likewise.
7923 (defer_stack_allocation): Likewise.
7924 (expand_used_vars): Likewise.
7925 * cfgcleanup.c (old_insns_match_p): Likewise.
7926 * asan.c (asan_finish_file): Likewise.
7927 (asan_instrument): Likewise.
7928 (gate_asan): Likewise.
7929 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
7930 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7931 (asan_global_struct): Use pointer_sized_int_node instead
7932 calling build_nonstandard_integer_type.
7933 (initialize_sanitizer_builtins): Likewise.
7934 (asan_finish_file): Likewise.
7935 * gcc.c: Document %{%:function(args):X}.
7936 (static_spec_functions): Add sanitize.
7937 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
7938 store funcval != NULL there.
7939 (do_spec_1): Adjust handle_spec_function caller.
7940 (handle_braces): Allow %:function(args) as condition.
7941 (sanitize_spec_function): New function.
7942 (ADD_STATIC_LIBUBSAN_LIBS): Define.
7943 (LIBUBSAN_SPEC): Likewise.
7944 (LIBUBSAN_EARLY_SPEC): Likewise.
7945 (SANITIZER_SPEC): Handle libubsan.
7946 (SANITIZER_EARLY_SPEC): Likewise.
7947 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
7948 instead of fsanitize=address.
7949 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
7950 instead of fsanitize=address*.
7951 * builtins.c: Include ubsan.h.
7952 (fold_builtin_0): Instrument __builtin_unreachable.
7953 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
7954 instead of flag_asan.
7955 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
7956 (pointer_sized_int_node): Define.
7957 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
7959 2013-08-30 Mike Stump <mikestump@comcast.net>
7961 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
7964 2013-08-29 Jan Hubicka <jh@suse.cz>
7966 * cgraph.c (cgraph_function_body_availability): Handle weakref
7968 * passes.def: Remove pass_fixup_cfg.
7969 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
7970 track when we need to remove functions.
7971 (gate_ipa_inline): Execute inlining always; add comment why.
7972 (pass_data_ipa_inline): Remove TODO_remove_functions.
7973 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
7974 do not produce summaries.
7975 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
7976 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
7977 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
7978 constant pool and vtable.
7980 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
7982 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
7983 New arm_neon.h's internal macros to specify 64-bit constants.
7984 Avoid using stdint.h's macros.
7986 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
7988 * recog.c (verify_changes): Verify that changes[i].old is non-zero
7989 before applying REG_P.
7991 2013-08-30 Jakub Jelinek <jakub@redhat.com>
7993 PR tree-optimization/58277
7994 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
7995 after seeing too many stmts with vdef in between dombb and current
7996 bb, invalidate everything.
7998 2013-08-30 Richard Biener <rguenther@suse.de>
8000 * fold-const.c (fold_single_bit_test): Fix overflow test.
8002 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
8004 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
8005 and which can live in a register, always retrieve the value on entry.
8006 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
8007 passed by invisible reference specially.
8008 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
8009 (vt_add_function_parameter): Correctly deal with a parameter passed by
8010 invisible reference.
8012 2013-08-30 Jan Hubicka <jh@suse.cz>
8014 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
8016 2013-08-30 Richard Biener <rguenther@suse.de>
8018 PR tree-optimization/58228
8019 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
8020 allow invariant loads in nested loop vectorization.
8022 2013-08-30 Richard Biener <rguenther@suse.de>
8024 PR tree-optimization/58223
8025 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
8026 (has_anti_or_output_dependence): ... this and adjust to also
8027 look for output dependences.
8028 (mark_nodes_having_upstream_mem_writes): Adjust.
8029 (rdg_flag_uses): Likewise.
8031 2013-08-30 Richard Biener <rguenther@suse.de>
8033 PR tree-optimization/58010
8034 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
8035 assert that we have a loop-closed PHI.
8037 2013-08-29 Jan Hubicka <jh@suse.cz>
8039 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
8040 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
8041 * lto-section-in.c (lto_free_function_in_decl_state): New function.
8042 (lto_free_function_in_decl_state_for_node): New function.
8044 2013-08-29 Xinliang David Li <davidxl@google.com>
8046 * loop-unroll.c (report_unroll_peel): Minor message change.
8047 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
8048 Emit alignment peeling message with default -fopt-info.
8049 (vect_loop_versioning): Emit loop version info message.
8050 * tree-vectorizer.c (vectorize_loops): Minor message change.
8051 (execute_vect_slp): Ditto.
8053 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
8055 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
8056 of the clone from the DECL_NAME of the original function.
8058 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
8060 * passes.c (register_pass): Add overload.
8061 * tree-pass.h (register_pass): Forward declare it. Add comment.
8063 2013-08-29 Jan Hubicka <jh@suse.cz>
8065 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
8066 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
8067 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
8068 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
8069 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
8070 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
8071 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
8072 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
8074 2013-08-29 Teresa Johnson <tejohnson@google.com>
8076 * dumpfile.c (dump_loc): Output column number.
8077 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
8078 * doc/invoke.texi: Document optall -fopt-info flag.
8079 * profile.c (read_profile_edge_counts): Use new dump framework.
8080 (compute_branch_probabilities): Ditto.
8081 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
8082 when pass not in any opt group.
8083 * pass_manager.h (pass_manager::get_pass_profile): New method.
8084 * value-prof.c (check_counter): Use new dump framework.
8085 (check_ic_target): Ditto.
8086 * coverage.c (get_coverage_counts): Ditto.
8087 (coverage_init): Setup new dump framework.
8089 2013-08-29 Richard Biener <rguenther@suse.de>
8091 PR tree-optimization/58246
8092 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
8093 handle the dominance check inside a basic-block.
8095 2013-08-29 Richard Biener <rguenther@suse.de>
8098 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
8099 of default defs that appear in abnormal PHI nodes.
8101 2013-08-29 Richard Biener <rguenther@suse.de>
8103 PR tree-optimization/57685
8104 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
8105 single-use operands to avoid exponential complexity.
8107 2013-08-28 Dehao Chen <dehao@google.com>
8109 * ipa-inline.c (edge_badness): Fix integer underflow.
8111 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
8113 * gtm-builtins.def (_ITM_free): Declare leaf.
8115 2013-08-28 Jakub Jelinek <jakub@redhat.com>
8118 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
8119 (*tls_local_dynamic_base_64_largepic): Likewise.
8120 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
8121 Remove predicate from call operand.
8122 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
8123 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
8125 2013-08-28 Jeff Law <law@redhat.com>
8127 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
8128 checks for the number of predecessors and successors allowed.
8129 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
8130 which require copying a joiner block if there is a request which
8131 is a subpath that requires no joiner block copying.
8133 2013-08-28 Jan Hubicka <jh@suse.cz>
8135 * lto-streamer-out.c (DFS_write_tree_body): Drop
8136 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
8137 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
8138 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
8139 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
8140 Do not read DECL_ERROR_ISSUED.
8141 (unpack_ts_decl_with_vis_value_fields): Do not read
8143 (lto_input_ts_binfo_tree_pointers): Do not read
8144 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
8145 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
8146 write DECL_ERROR_ISSUED..
8147 (pack_ts_decl_with_vis_value_fields): Do not write
8149 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
8150 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
8151 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
8152 * tree.h (tree_decl_common): Update comment.
8153 (DECL_ERROR_ISSUED): Remove.
8155 2013-08-28 Jakub Jelinek <jakub@redhat.com>
8158 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
8160 2013-08-28 Jan Hubicka <jh@suse.cz>
8162 * builtins.def (free): Declare leaf.
8164 2013-08-27 David Malcolm <dmalcolm@redhat.com>
8167 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
8168 * configure: Regenerate.
8170 2013-08-27 Martin Jambor <mjambor@suse.cz>
8172 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
8173 (ipa_ancestor_jf_data): Likewise.
8174 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
8175 (ipa_get_jf_pass_through_type_preserved): New function.
8176 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
8177 (ipa_get_jf_ancestor_type_preserved): New function.
8178 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
8179 (ipa_get_jf_ancestor_result): Likewise.
8180 (propagate_vals_accross_pass_through): Use
8181 ipa_get_jf_pass_through_result to do all the value mappings.
8182 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
8183 type_preserved flag.
8184 (ipa_set_jf_cst_copy): New function.
8185 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
8186 (ipa_set_jf_arith_pass_through): Likewise.
8187 (ipa_set_ancestor_jf): Likewise.
8188 (compute_complex_assign_jump_func): Set type_preserved instead of
8190 (ipa_compute_jump_functions_for_edge): Likewise.
8191 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
8192 (update_jump_functions_after_inlining): Update type_preserved.
8193 Explicitely create jump functions when combining one with pass_through.
8194 (ipa_write_jump_function): Stream the type_preserved flags.
8195 (ipa_read_jump_function): Likewise.
8197 2013-08-27 Jakub Jelinek <jakub@redhat.com>
8198 Aldy Hernandez <aldyh@redhat.com>
8200 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
8201 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
8202 * function.h (struct function): Add has_force_vect_loops and
8204 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
8205 * gimple.c (gimple_build_omp_critical): Add KIND argument and
8207 * gimple.def: Update CLAUSES comments.
8208 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
8209 (gimple_build_omp_for): Add argument to prototype.
8210 (gimple_omp_for_kind): New.
8211 (gimple_omp_for_set_kind): New.
8212 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
8213 GOVD_DATA_SHARE_CLASS.
8214 (enum omp_region_type): Add ORT_SIMD.
8215 (gimple_add_tmp_var): Handle ORT_SIMD.
8216 (gimplify_var_or_parm_decl): Same.
8217 (is_gimple_stmt): Same.
8218 (omp_firstprivatize_variable): Same.
8219 (omp_add_variable): Only use splay_tree_insert if lookup failed.
8220 (omp_notice_variable): Handle ORT_SIMD.
8221 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
8222 (omp_check_private): Handle ORT_SIMD.
8223 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
8225 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
8226 Handle OMP_CLAUSE_LASTPRIVATE.
8227 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
8229 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
8230 (gimplify_expr): Handle OMP_SIMD.
8231 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
8232 (expand_GOMP_SIMD_VF): New.
8233 (expand_GOMP_SIMD_LAST_LANE): New.
8234 * internal-fn.def (GOMP_SIMD_LANE): New.
8235 (GOMP_SIMD_VF): New.
8236 (GOMP_SIMD_LAST_LANE): New.
8237 * omp-low.c: Include target.h.
8238 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
8240 (check_omp_nesting_restrictions): Same.
8242 (lower_rec_simd_input_clauses): New.
8243 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
8245 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
8246 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
8247 (expand_omp_build_assign): New.
8248 (expand_omp_for_init_counts): New.
8249 (expand_omp_for_init_vars): New.
8250 (extract_omp_for_update_vars): New.
8251 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
8252 and rewrite accordingly.
8253 (expand_omp_simd): New.
8254 (expand_omp_for): Use expand_omp_simd.
8255 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
8256 (lower_omp_for): Do not lower the body.
8257 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
8259 * tree-flow.h (find_omp_clause): Remove prototype.
8260 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
8261 forcing vectorization of the loop, or if flag_tree_vectorize.
8262 (gate_tree_if_conversion): Similarly.
8263 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
8264 gimple_build_omp_for.
8265 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
8266 * tree-parloops (create_parallel_loop): Pass kind argument to
8267 gimple_build_omp_for.
8268 * tree-pretty-print.c (dump_omp_clause): Add cases for
8269 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
8270 OMP_CLAUSE__SIMDUID_.
8271 (dump_generic_node): Handle OMP_SIMD.
8272 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
8273 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
8274 unroll OMP_SIMD loops here.
8275 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
8276 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
8278 (vect_analyze_data_refs): Handle simd loops.
8279 * tree-vect-loop.c (vectorizable_live_operation): Handle
8281 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
8282 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
8283 (vectorizable_load): Same.
8284 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
8285 (struct simduid_to_vf): New.
8286 (simduid_to_vf::hash): New.
8287 (simduid_to-vf::equal): New.
8288 (struct simd_array_to_simduid): New.
8289 (simd_array_to_simduid::hash): New.
8290 (simd_array_to_simduid::equal): New.
8291 (adjust_simduid_builtins): New.
8292 (struct note_simd_array_uses_struct): New.
8293 (note_simd_array_uses_cb): New.
8294 (note_simd_array_uses): New.
8295 (vectorize_loops): Handle simd hints and adjust simd builtins
8297 * tree-vectorizer.h (struct _stmt_vec_info): Add
8298 simd_lane_access_p field.
8299 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
8300 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
8301 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
8302 (omp_clause_code_name): Same.
8303 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
8304 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
8305 * tree.def (OMP_SIMD): New entry.
8306 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
8307 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
8308 (OMP_CLAUSE_DECL): Adjust range for new clauses.
8309 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
8310 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
8311 (OMP_CLAUSE_LINEAR_STEP): New.
8312 (OMP_CLAUSE_SAFELEN_EXPR): New.
8313 (OMP_CLAUSE__SIMDUID__DECL): New.
8314 (find_omp_clause): New prototype.
8316 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
8318 * config/i386/driver-i386.c (host_detect_local_cpu): Update
8319 Haswell processor detection.
8321 2013-08-27 Christian Widmer <shadow@umbrox.de>
8324 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
8325 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
8326 AVX2 capable processors.
8328 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
8330 * config/aarch64/arm_neon.h: Replace all inline asm implementations
8331 of vget_low_* with implementations in terms of other intrinsics.
8333 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
8336 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
8337 values to -1, 0 and 1.
8339 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
8341 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
8342 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
8343 (<optab><mode>3_insn): Remove.
8344 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
8345 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
8346 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
8347 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
8348 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
8349 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
8350 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
8351 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
8352 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
8353 (ror<mode>3_insn): Likewise.
8354 * config/aarch64/predicates.md (aarch64_simd_register): New.
8356 2013-08-27 Richard Biener <rguenther@suse.de>
8358 PR tree-optimization/57521
8359 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
8360 one edge is non-critical.
8361 (find_phi_replacement_condition): Make sure to use a non-critical
8362 edge. Cleanup and remove old bug workarounds.
8363 (bb_postdominates_preds): Remove.
8364 (if_convertible_loop_p_1): Do not compute post-dominators.
8365 (combine_blocks): Do not free post-dominators.
8366 (main_tree_if_conversion): Likewise.
8367 (pass_data_if_conversion): Add TODO_verify_ssa.
8369 2013-08-27 DJ Delorie <dj@redhat.com>
8371 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
8373 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
8375 * function.c (assign_parm_find_data_types): Set passed_mode and
8376 nominal_mode to the TYPE_MODE of nominal_type for the built
8377 pointer type in case of the struct-pass-by-reference.
8379 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8381 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
8382 (UINT16_TYPE): Change default to "unsigned int".
8384 * config/avr/avr.opt (mfract-convert-truncate): New option.
8385 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
8386 is set, round negative fractional integers according to n1169
8387 when converting to integer types.
8389 2013-08-26 Jan Hubicka <jh@suse.cz>
8391 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
8392 methods can not be called indirectly when their address is not taken.
8394 2013-08-26 Jan Hubicka <jh@suse.cz>
8396 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
8399 2013-08-26 Jan Hubicka <jh@suse.cz>
8401 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
8404 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8406 * reload.c (find_valid_class): Allow classes that do not include
8407 FIRST_PSEUDO_REGISTER - 1.
8409 2013-08-26 Jan Hubicka <jh@suse.cz>
8411 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
8412 fix edge count/frequency when speculation failed; fix type check
8413 for the direct call.
8415 2013-08-26 Jan Hubicka <jh@suse.cz>
8417 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
8419 2013-08-26 Jan Hubicka <jh@suse.cz>
8421 * ipa-inline-transform.c (inline_transform): Be ready for basic block
8422 to be changed by edge redirection.
8424 2013-08-26 Jan Hubicka <jh@suse.cz>
8426 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
8427 formating; add sanity check.
8428 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
8429 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
8430 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
8431 (resolve_noninline_speculation): Update callee keys, too.
8433 2013-08-26 Jan Hubicka <jh@suse.cz>
8435 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
8436 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
8438 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8440 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
8443 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
8445 * config/i386/i386.c (ix86_debug_options): Remove prototype.
8446 (x86_64_elf_select_section): Ditto.
8447 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
8449 (ix86_pass_by_reference): Ditto.
8450 (output_set_got): Ditto.
8451 (ix86_unary_operator_ok): Ditto.
8452 (ix86_expand_builtin): Ditto.
8454 2013-08-23 Jan Hubicka <jh@suse.cz>
8456 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
8458 2013-08-23 Jan Hubicka <jh@suse.cz>
8460 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
8461 (tree_decl_with_vis): Add FINAL field.
8463 2013-08-23 Jeff Law <law@redhat.com>
8465 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
8466 the debugging dump when the expression is fully redundant.
8468 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8470 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
8471 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
8472 * pretty-print.c (pp_formatted_text_data): Likewise.
8473 (pp_write_text_to_stream): Likewise.
8474 (pp_write_text_as_dot_label_to_stream): Likewise.
8475 (pp_append_r): Likewise.
8476 (pp_format): Likewise.
8477 (pp_flush): Likewise.
8478 (pp_clear_output_area): Likewise.
8479 (pp_append_text): Likewise.
8480 (pp_formatted_text): Likewise.
8481 (pp_remaining_character_count_for_line): Likewise.
8482 (pp_newline): Likewise.
8483 (pp_character): Likewise.
8484 (output_buffer::~output_buffer): Define.
8485 (pretty_printer::~pretty_printer): Destruct output buffer.
8486 * pretty-print.h (output_buffer::~output_buffer): Declare.
8487 (pretty_printer::~pretty_printer): Declare virtual.
8489 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
8492 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
8493 HOST_WIDE_INT_M1U): New macros.
8494 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
8495 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
8496 unsigned -1 for lshift.
8497 * cse.c (cse_insn): Likewise.
8498 * double-int.c (rshift_double, lshift_double): Likewise.
8499 * builtins.c (fold_builtin_bitop): Likewise.
8500 * combine.c (force_to_mode): Likewise.
8501 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
8502 * simplify-rtx.c (simplify_const_unary_operation,
8503 simplify_const_binary_operation): Likewise.
8504 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
8505 check_va_list_escapes): Likewise.
8506 * rtlanal.c (nonzero_bits1): Likewise.
8507 * expmed.c (expand_smod_pow2): Likewise.
8508 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
8510 2013-08-23 Jan Hubicka <jh@suse.cz>
8512 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
8513 as having address taken.
8515 2013-08-23 Jan Hubicka <jh@suse.cz>
8517 * ipa-utils.h (method_class_type): Declare.
8518 * ipa-devirt.c (method_class_type): Export.
8520 * cgraphunit.c (analyze_functions): Do basic devirtualization;
8521 do not walk base classes of anonymous types.
8523 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
8525 PR rtl-optimization/58220
8527 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
8528 handle SEQUENCE insns properly.
8530 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8532 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
8534 (pp_newline_and_indent): Likewise.
8535 (pp_separate_with): Likewise.
8536 * pretty-print.c (pp_newline_and_flush): Define.
8537 (pp_newline_and_indent): Likewise.
8538 (pp_separate_with): Likewise.
8540 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8543 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
8544 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
8546 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
8548 * config/i386/predicates.md (ext_sse_reg_operand): New.
8549 * config/i386/i386.md (*movti_internal): Use
8550 predicate to determine if EVEX is needed.
8551 (*movsi_internal): Ditto.
8552 (*movdf_internal): Ditto.
8553 (*movsf_internal): Ditto.
8554 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
8556 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8558 PR tree-optimization/58209
8559 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
8560 (find_tail_calls): Give up for pointer result types if m is non-NULL.
8561 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
8562 emit POINTER_PLUS_EXPR.
8563 (create_tailcall_accumulator): For pointer result type accumulate in
8566 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
8568 * configure.ac: Add backslashes missing from the last change.
8569 * configure: Regenerate.
8571 2013-08-22 Jan Hubicka <jh@susue.cz>
8573 * ipa.c (function_and_variable_visibility): First remember function
8574 was global and then make it local.
8576 2013-08-22 Julian Brown <julian@codesourcery.com>
8578 * configure.ac: Add aarch64 to list of arches which use "nop" in
8580 * configure: Regenerate.
8582 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8584 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
8585 gnu_libc_has_function.
8586 * config/s390/tpf.h: Likewise.
8588 2013-08-22 Jan Hubicka <jh@susue.cz>
8590 * timevar.c (validate_phases): Add cast.
8592 2013-08-22 Jan Hubicka <jh@susue.cz>
8594 * timevar.c (validate_phases): Use size_t for memory.
8595 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
8597 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
8599 * pretty-print.h (output_buffer::output_buffer): Declare.
8600 (pretty_printer::pretty_printer): Likewise.
8601 (pp_construct): Remove.
8602 * pretty-print.c (output_buffer::output_buffer): Define.
8603 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
8604 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
8605 (print_gimple_expr): Likewise.
8606 (print_gimple_seq): Likewise.
8607 (gimple_dump_bb): Likewise.
8608 * sched-vis.c (dump_value_slim): Likewise.
8609 (dump_insn_slim): Likewise.
8610 (dump_rtl_slim): Likewise.
8611 (str_pattern_slim): Likewise.
8612 * tree-mudflap.c (mf_varname_tree): Likewise.
8613 * graph.c (print_graph_cfg): Likewise.
8614 (start_graph_dump): Likewise.
8615 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
8617 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
8618 pretty printer initialization.
8619 * coretypes.h (diagnostic_context): Remove superflous type alias
8621 (pretty_printer): Likewise. Declare directly as a class.
8622 (pretty_print_info): Remove declaration as class.
8623 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
8624 and pp_clear_output_area.
8625 (asan_add_global): Likewise.
8627 2013-08-22 Jan Hubicka <jh@suse.cz>
8629 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
8630 * ipa-utils.h (update_type_inheritance_graph): Declare.
8631 (possible_polymorphic_call_target_p): Declare.
8632 (possible_polymorphic_call_target_p): New.
8633 * ipa-devirt.c: Update toplevel comments.
8634 (cached_polymorphic_call_targets): Move up.
8635 (odr_type_d): Move ID down.
8636 (polymorphic_type_binfo_p): Update comment.
8637 (odr_hasher::remove): Likewise;
8638 (get_odr_type): Set anonymous_namespace.
8639 (dump_odr_type): Dump it.
8640 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
8641 (maybe_record_node): Record node in cached_polymorphic_call_targets.
8642 (record_binfo): Add comment.
8643 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
8645 (devirt_node_removal_hook): Do not iCE when cache is freed.
8646 (possible_polymorphic_call_target_p): New predicate.
8647 (update_type_inheritance_graph): New function.
8649 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
8650 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8651 Sergey Lega <sergey.s.lega@intel.com>
8652 Anna Tikhonova <anna.tikhonova@intel.com>
8653 Ilya Tocar <ilya.tocar@intel.com>
8654 Andrey Turetskiy <andrey.turetskiy@intel.com>
8655 Ilya Verbin <ilya.verbin@intel.com>
8656 Kirill Yukhin <kirill.yukhin@intel.com>
8657 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8659 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
8660 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
8661 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
8662 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
8663 (OPTION_MASK_ISA_AVX2_UNSET): Update.
8664 (OPTION_MASK_ISA_AVX512F_UNSET): New.
8665 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
8666 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
8667 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
8668 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
8669 OPT_mavx512pf, OPT_mavx512er cases.
8670 * config/i386/constraints.md (v): New constraint.
8671 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
8672 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
8673 (bit_AVX512CD): New.
8674 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8675 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
8676 * config/i386/i386-c.c (ix86_target_macros_internal):
8677 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
8679 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
8680 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
8681 * config/i386/i386.c (regclass_map, dbx_register_map)
8682 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
8683 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
8684 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
8686 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
8687 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
8688 -mavx512pf options. Fix formatting.
8689 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
8690 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
8691 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
8692 -mavx512cd, -mavx512pf options.
8693 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
8694 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
8695 (ix86_preferred_output_reload_class): Replace SSE_REGS with
8697 (ix86_hard_regno_mode_ok): Support 512-bit registers.
8698 (ix86_set_reg_reg_cost): Ditto.
8699 (x86_order_regs_for_local_alloc): Ditto.
8700 (MAX_VECT_LEN): Extend to 64-byte.
8701 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
8702 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
8703 (TARGET_AVX512ER, TARGET_AVX512CD): New.
8704 (BIGGEST_ALIGNMENT): Extend to 512-bits.
8705 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
8706 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
8707 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
8708 (SSE_REG_MODE_P): Support new modes.
8709 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
8710 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
8711 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
8712 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
8713 (REG_CLASS_CONTENTS): Add new registers.
8714 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
8715 (EXT_REX_SSE_REGNO_P): New.
8716 (HI_REGISTER_NAMES): Add new registers.
8717 * config/i386/i386.md: Define constants for new registers.
8718 (mode): Add new 512-bit modes.
8719 (prefix): Support evex prefix.
8720 (isa): Support avx512f, noavx512f, fma_avx512f.
8721 (ssemodesuffix): Add new 512-bit modes.
8723 (*movxi_internal_avx512f): Ditto.
8724 (*movdi_internal): Replace constraint "x" with the new constraint "v".
8726 (*movsi_internal): Likewise.
8727 (*movdf_internal): Likewise.
8728 (*movsf_internal): Likewise.
8729 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
8730 (<code><mode>3): Likewise.
8731 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
8733 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
8734 with the new constraint "v".
8735 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
8737 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
8738 with the new constraint "v".
8739 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
8740 (<sse2>_storedqu<avxsizesuffix>): Likewise.
8741 (*<plusminus_insn><mode>3): Likewise.
8742 (<sse>_vm<plusminus_insn><mode>3): Likewise.
8743 (*mul<mode>3): Likewise.
8744 (<sse>_vmmul<mode>3): Likewise.
8745 (<sse>_div<mode>3): Likewise.
8746 (<sse>_vmdiv<mode>3): Likewise.
8747 (<sse>_sqrt<mode>2): Likewise.
8748 (<sse>_vmsqrt<mode>2): Likewise.
8749 (*<code><mode>3_finite): Likewise.
8750 (*<code><mode>3) <smaxmin>: Likewise.
8751 (<sse>_vm<code><mode>3): Likewise.
8752 (*<code><mode>3) <any_logic>: Likewise.
8753 (*fma_fmadd_<mode>): Likewise.
8754 (*fma_fmsub_<mode>): Likewise.
8755 (*fma_fnmadd_<mode>): Likewise.
8756 (*fma_fnmsub_<mode>): Likewise.
8757 (*fma_fmaddsub_<mode>): Likewise.
8758 (*fma_fmsubadd_<mode>): Likewise.
8759 (*fmai_fmadd_<mode>): Likewise.
8760 (*fmai_fmsub_<mode>): Likewise.
8761 (*fmai_fnmadd_<mode>): Likewise.
8762 (*fmai_fnmsub_<mode>): Likewise.
8763 (sse_cvtsi2ss): Likewise.
8764 (sse_cvtsi2ssq): Likewise.
8765 (sse_cvtss2si): Likewise.
8766 (sse_cvtss2si_2): Likewise.
8767 (sse_cvtss2siq): Likewise.
8768 (sse_cvtss2siq_2): Likewise.
8769 (sse_cvttss2si): Likewise.
8770 (sse_cvtss2siq_2): Likewise.
8771 (float<sseintvecmodelower><mode>2): Likewise.
8772 (sse2_cvtsd2si_2): Likewise.
8773 (sse2_cvtsd2siq_2): Likewise.
8774 (*<plusminus_insn><mode>3): Likewise.
8775 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
8776 (*<sse4_1_avx2>_mul<mode>3): Likewise.
8777 (ashr<mode>3): Likewise.
8778 (<shift_insn><mode>3): Likewise.
8779 (avx2_<code><mode>3): Likewise.
8780 (*avx2_<code><mode>3): Likewise.
8781 (*andnot<mode>3): Likewise.
8782 (*<code><mode>3) <any_logic>: Likewise.
8783 (abs<mode>2): Likewise.
8784 (avx2_permvar<mode>): Likewise.
8785 (avx2_perm<mode>_1): Likewise.
8786 (*avx_vpermilp<mode>): Likewise.
8787 (avx_vpermilvar<mode>3): Likewise.
8788 (avx2_ashrv<mode>): Likewise.
8789 (avx2_<shift_insn>v<mode>): Likewise.
8790 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
8792 * doc/rtl.texi: Document XImode.
8794 2013-08-21 Jeff Law <law@redhat.com>
8796 * tree-flow.h (register_jump_thread): Pass vector of edges
8797 instead of each important edge.
8798 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
8799 thread path into a vector and pass that to register_jump_thread.
8800 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
8801 passed in edge vector to the current 3-edge form.
8804 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8806 * dce.c (fini_dce): Call df_analyze again just in case
8807 delete_unmarked_insns removed anything.
8809 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
8811 * reload.h (struct reg_equivs): Rename to ..
8812 (struct reg_equivs_s): .. this.
8814 2013-08-20 Martin Liska <marxin.liska@gmail.com>
8816 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
8818 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8820 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
8822 2013-08-21 Jeff Law <law@redhat.com>
8824 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
8825 simplify assignments too. If the RHS collapses to a singleton
8826 range, then return the value for the range.
8828 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
8830 * config/i386/sse.md (V16): Rename to...
8832 (mov<mode>): Update iterator name.
8833 (*mov<mode>_internal): Ditto.
8834 (push<mode>1): Ditto.
8835 (movmisalign<mode>): Ditto.
8837 2013-08-20 Jan Hubicka <jh@suse.cz>
8840 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
8841 entry for direct edges.
8842 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
8844 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8846 Revert my last two changes, r201865 and r201864:
8849 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8851 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8852 instances can own GC refs.
8854 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8855 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8856 (gcc::context::gt_pch_nx): Likewise.
8857 (gcc::context::gt_pch_nx): Likewise.
8858 * ggc.h (gt_ggc_mx <T>): New.
8859 (gt_pch_nx_with_op <T>): New.
8860 (gt_pch_nx <T>): New.
8861 * passes.c (opt_pass::gt_ggc_mx): New.
8862 (opt_pass::gt_pch_nx): New.
8863 (opt_pass::gt_pch_nx_with_op): New.
8864 (pass_manager::gt_ggc_mx): New.
8865 (pass_manager::gt_pch_nx): New.
8866 (pass_manager::gt_pch_nx_with_op): New.
8867 (pass_manager::operator new): Use
8868 ggc_internal_cleared_alloc_stat rather than xcalloc.
8869 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8870 (pass_manager::gt_ggc_mx): New.
8871 (pass_manager::gt_pch_nx): New.
8872 (pass_manager::gt_pch_nx_with_op): New.
8873 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8874 (opt_pass::operator new): New.
8875 (opt_pass::gt_ggc_mx): New.
8876 (opt_pass::gt_pch_nx): New.
8877 (opt_pass::gt_pch_nx_with_op): New.
8880 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8882 * Makefile.in (GTFILES): Add context.h.
8883 * context.c (gcc::context::operator new): New.
8884 (gcc::context::gt_ggc_mx): New.
8885 (gcc::context::gt_pch_nx): New.
8886 (gcc::context::gt_pch_nx): New.
8887 * context.h (gcc::context): Add GTY((user)) marking.
8888 (gcc::context::operator new): New.
8889 (gcc::context::gt_ggc_mx): New.
8890 (gcc::context::gt_pch_nx): New.
8891 (gcc::context::gt_pch_nx): New.
8892 (g): Add GTY marking.
8893 (gt_ggc_mx (gcc::context *)): New.
8894 (gt_pch_nx (gcc::context *)): New.
8895 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8896 void *cookie)): New.
8897 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8899 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8901 * dce.c (fini_dce): Call df_analyze again just in case
8902 delete_unmarked_insns removed anything.
8904 2013-08-20 Teresa Johnson <tejohnson@google.com>
8906 PR rtl-optimizations/57451
8907 * final.c (reemit_insn_block_notes): Prevent lexical blocks
8908 from crossing split section boundaries.
8910 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8912 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
8913 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
8914 with MULTLIB_DEFAULTS.
8916 2013-08-20 Nick Clifton <nickc@redhat.com>
8918 * target.def (narrow_volatile_bitfield): Note that the default
8919 value is false, not !TARGET_STRICT_ALIGN.
8920 * doc/tm.texi: Regenerate.
8922 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
8924 Fix LIB_SPEC for systems without libpthread.
8926 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
8927 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
8929 * config/i386/linux-common.h: Likewise.
8930 * config/mips/linux-common.h: Likewise.
8932 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
8934 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
8937 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8939 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8940 instances can own GC refs.
8942 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8943 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8944 (gcc::context::gt_pch_nx): Likewise.
8945 (gcc::context::gt_pch_nx): Likewise.
8946 * ggc.h (gt_ggc_mx <T>): New.
8947 (gt_pch_nx_with_op <T>): New.
8948 (gt_pch_nx <T>): New.
8949 * passes.c (opt_pass::gt_ggc_mx): New.
8950 (opt_pass::gt_pch_nx): New.
8951 (opt_pass::gt_pch_nx_with_op): New.
8952 (pass_manager::gt_ggc_mx): New.
8953 (pass_manager::gt_pch_nx): New.
8954 (pass_manager::gt_pch_nx_with_op): New.
8955 (pass_manager::operator new): Use
8956 ggc_internal_cleared_alloc_stat rather than xcalloc.
8957 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8958 (pass_manager::gt_ggc_mx): New.
8959 (pass_manager::gt_pch_nx): New.
8960 (pass_manager::gt_pch_nx_with_op): New.
8961 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8962 (opt_pass::operator new): New.
8963 (opt_pass::gt_ggc_mx): New.
8964 (opt_pass::gt_pch_nx): New.
8965 (opt_pass::gt_pch_nx_with_op): New.
8967 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8969 * Makefile.in (GTFILES): Add context.h.
8970 * context.c (gcc::context::operator new): New.
8971 (gcc::context::gt_ggc_mx): New.
8972 (gcc::context::gt_pch_nx): New.
8973 (gcc::context::gt_pch_nx): New.
8974 * context.h (gcc::context): Add GTY((user)) marking.
8975 (gcc::context::operator new): New.
8976 (gcc::context::gt_ggc_mx): New.
8977 (gcc::context::gt_pch_nx): New.
8978 (gcc::context::gt_pch_nx): New.
8979 (g): Add GTY marking.
8980 (gt_ggc_mx (gcc::context *)): New.
8981 (gt_pch_nx (gcc::context *)): New.
8982 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8983 void *cookie)): New.
8984 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8986 2013-08-20 Alan Modra <amodra@gmail.com>
8989 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
8990 (rs6000_emit_epilogue): Likewise.
8992 2013-08-19 Dehao Chen <dehao@google.com>
8994 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
8996 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
8997 Jakub Jelinek <jakub@redhat.com>
8999 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
9000 (BUILT_IN_FABSD64): Likewise.
9001 (BUILT_IN_FABSD128): Likewise.
9002 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
9003 (fold_builtin_1): Likewise.
9004 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
9005 destination and source operands.
9006 (*abstd2_fpr): Likewise.
9007 (*nabstd2_fpr): Likewise.
9009 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
9011 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
9014 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
9016 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
9019 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
9021 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
9022 * builtins.c (default_libc_has_function): New.
9023 (gnu_libc_has_function): Ditto.
9024 (no_c99_libc_has_function): Ditto.
9025 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
9026 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
9027 (fold_builtin_sincos): Likewise.
9028 (fold_builtin_cexp): Likewise.
9029 * builtins.def (DEF_C94_BUILTIN): Likewise.
9030 (DEF_C99_BUILTIN): Likewise.
9031 (DEF_C99_C90RES_BUILTIN): Likewise.
9032 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
9033 definitions to using this define.
9034 * config/darwin-protos.h (darwin_libc_has_function): New.
9035 * config/darwin.c (darwin_libc_has_function): Ditto.
9036 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
9037 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
9038 * config/darwin.h: Ditto.
9039 * config/elfos.h: Ditto.
9040 * config/freebsd.h: Ditto.
9041 * config/i386/cygming.h: Ditto.
9042 * config/i386/djgpp.h: Ditto.
9043 * config/i386/i386-interix.h: Ditto.
9044 * config/microblaze/microblaze.h: Ditto.
9045 * config/mmix/mmix.h: Ditto.
9046 * config/gnu-user.h: Ditto.
9047 * config/ia64/hpux.h: Ditto.
9048 * config/pa/pa-hpux.h: Ditto.
9049 * config/pdp11/pdp11.h: Ditto.
9050 * config/picochip/picochip.h: Ditto.
9051 * config/linux.h: Ditto.
9052 * config/netbsd.h: Ditto.
9053 * config/openbsd.h: Ditto.
9054 * config/rs6000/aix43.h: Ditto.
9055 * config/rs6000/aix51.h: Ditto.
9056 * config/rs6000/aix52.h: Ditto.
9057 * config/rs6000/aix53.h: Ditto.
9058 * config/rs6000/aix61.h: Ditto.
9059 * config/rs6000/darwin.h: Ditto.
9060 * config/rs6000/linux.h: Ditto.
9061 * config/rs6000/linux64.h: Ditto.
9062 * config/s390/tpf.h: Ditto.
9063 * config/sol2-10.h: Ditto.
9064 * config/sol2.h: Ditto.
9065 * config/vms/vms.h: Ditto.
9066 * config/vxworks.h: Ditto.
9067 * config/linux-android.c (linux_android_libc_has_function):
9068 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
9069 * config/linux-protos.h (linux_android_libc_has_function):
9071 * config/i386/i386.c (ix86_libc_has_function): New.
9072 * config/i386/i386-protos.h
9073 (ix86_libc_has_function): New declaration.
9074 * config/i386/i386.md
9075 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
9076 ("isinf<mode>2): Likewise.
9077 * convert.c (convert_to_integer): Using new target hook
9078 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
9079 TARGET_C99_FUNCTIONS.
9080 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
9081 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
9082 * coretypes.h (function_class): New enum for different
9083 classes of functions.
9084 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
9085 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
9086 (TARGET_HAS_SINCOS): Likewise.
9087 (TARGET_LIBC_HAS_FUNCTION): New.
9088 * doc/tm.texi: Regenerated.
9089 * targhooks.h (default_libc_has_function): New declaration.
9090 (no_c99_libc_has_function): Ditto.
9091 (gnu_libc_has_function): Ditto.
9092 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
9093 and TARGET_HAS_SINCOS.
9095 2013-08-18 Jan Hubicka <jh@suse.cz>
9097 * Makeifle-in (ipa-devirt.o): New.
9098 (GTFILES): Add ipa-utils.h and ipa-devirt.c
9099 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
9100 (analyze_functions): Look into possible targets of polymorphic call.
9101 * dumpfile.c (dump_files): Add type-inheritance dump.
9102 * dumpfile.h (TDI_inheritance): New.
9103 * ipa-devirt.c: New file.
9104 * ipa-utils.h (odr_type_d): Forward declare.
9105 (odr_type): New type.
9106 (build_type_inheritance_graph): Declare.
9107 (possible_polymorphic_call_targets): Declare and introduce inline
9108 variant when only edge is pased.
9109 (dump_possible_polymorphic_call_targets): Likewise.
9110 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
9111 * tree.c (type_in_anonymous_namespace_p): Break out from ...
9112 (types_same_for_odr): ... here.
9113 * tree.h (type_in_anonymous_namespace_p): Declare.
9115 2013-08-18 Jakub Jelinek <jakub@redhat.com>
9117 PR tree-optimization/58006
9118 * tree-parloops.c (take_address_of): Don't ICE if get_name
9120 (eliminate_local_variables_stmt): Remove clobber stmts.
9122 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
9124 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
9125 error message is issued for an alias to undefined symbol.
9127 2013-08-18 Jan Hubicka <jh@suse.cz>
9129 * cgraph.c (cgraph_create_indirect_edge): Discover
9130 polymorphic calls and record basic info into indirect_info.
9131 * gimple-fold.c (gimple_fold_call): When doing BINFO based
9132 devirtualization, ignore objc function calls.
9133 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
9134 call with no parm index info.
9135 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
9136 * tree.c (virtual_method_call_p): New function.
9137 * tree.h (virtual_method_call_p): Declare.
9139 2013-08-16 Jan Hubicka <jh@suse.cz>
9142 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
9144 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
9146 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
9149 2013-08-16 David Malcolm <dmalcolm@redhat.com>
9151 * gengtype.c (type_for_name): Add special-case support for
9152 locating types within the "gcc::" namespace.
9153 (open_base_files): Emit a "using namespace gcc" directive.
9155 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9158 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
9159 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
9161 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
9162 array instead of each individual operand as a separate argument.
9163 (emit_fusion_gpr_load): Likewise.
9164 (expand_fusion_gpr_load): Add new function declaration.
9166 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
9167 signature to have the operands passed as an array, instead of as
9168 separate arguments. Allow ZERO_EXTEND to be in the memory
9169 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
9170 depend on the register live/dead flags when peepholes are run.
9171 (expand_fusion_gpr_load): New function to be called from the
9172 peephole2 pass, to change the register that addis sets to be the
9174 (emit_fusion_gpr_load): Change the calling signature to have the
9175 operands passed as an array, instead of as separate arguments.
9176 Allow ZERO_EXTEND to be in the memory address, and also
9177 SIGN_EXTEND if -mpower8-fusion-sign.
9179 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
9181 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
9182 adjust the register addis loads up in the peephole2 pass. Do not
9183 depend on the register live/dead state when the peephole pass is done.
9185 2013-08-16 David Malcolm <dmalcolm@redhat.com>
9187 * gengtype.c (create_user_defined_type): Ensure that the kind
9188 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
9189 declaration is seen before the GTY((user)) marking.
9191 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
9194 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
9196 2013-08-16 Jan Hubicka <jh@suse.cz>
9198 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
9199 arugment expected_type.
9200 (gimple_fold_call): Use it.
9201 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
9202 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
9203 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
9204 (try_make_edge_direct_virtual_call): Likewise.
9205 * tree.c (obj_type_ref_class): New.
9206 * tree.h (obj_type_ref_class): Use it.
9208 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
9210 * sched-vis.c (rtl_slim_pp_initialized): Remove.
9211 (rtl_slim_pp): Likewise.
9212 (init_rtl_slim_pretty_print): Likewise.
9213 (dump_value_slim): Don't call it. Use local pretty printer.
9214 (dump_insn_slim): Likewise.
9215 (dump_rtl_slim): Likewise.
9216 (str_pattern_slim): Likewise.
9217 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
9220 2013-08-16 Jakub Jelinek <jakub@redhat.com>
9222 PR tree-optimization/58164
9223 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
9224 walk gimple_goto_dest of GIMPLE_GOTO.
9226 PR tree-optimization/58165
9227 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
9228 bi_call must be the last stmt in a bb, don't split_block, instead
9229 use fallthru edge from it and give up if there is none.
9230 Release conds vector when returning early.
9232 2013-08-14 Xinliang David Li <davidxl@google.com>
9234 * config/i386/i386.c (ix86_option_override_internal):
9235 Remove unused variable and field.
9237 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9240 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
9241 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
9242 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
9243 and Linux, correct BLKmode alignment when 128-bit alignment is
9244 required and compatibility flag is not set.
9245 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
9246 for zero-size arguments when compatibility flag is not set.
9248 2013-08-14 Jakub Jelinek <jakub@redhat.com>
9250 PR tree-optimization/58145
9251 * tree-sra.c (build_ref_for_offset): If prev_base has
9252 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
9254 2013-08-14 Xinliang David Li <davidxl@google.com>
9256 * config/i386/i386.c (ix86_option_override_internal):
9257 Fix uninitialized variable error.
9259 2013-08-14 Xinliang David Li <davidxl@google.com>
9261 * config/i386/i386.opt: Define two new options.
9262 * config/i386/x86-tune.def: Add arch selector field in macros.
9263 * config/i386/i386.h: Adjust macro definition.
9264 * config/i386/i386.c (ix86_option_override_internal):
9266 (parse_mtune_ctrl_str): New function.
9267 (set_ix86_tune_features): New function.
9268 (ix86_function_specific_restore): Call the new helper function.
9270 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
9272 PR rtl-optimization/57662
9273 * sel-sched.c (code_motion_process_successors): When the current insn
9274 is removed after the recursive traversal, break from the loop.
9275 Add comments and debug printouts.
9277 2013-08-14 Jakub Jelinek <jakub@redhat.com>
9278 Alexandre Oliva <aoliva@redhat.com>
9281 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
9282 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
9283 there also UNSPEC_PLTOFF.
9285 2013-08-14 Marek Polacek <polacek@redhat.com>
9287 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
9288 NUM_CONDITIONS bit positions.
9290 2013-08-13 Cary Coutant <ccoutant@google.com>
9292 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
9293 (attr_checksum): Hash vector contents instead of pointer.
9294 (attr_checksum_ordered): Likewise.
9296 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
9298 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
9299 when Pmode != word_mode. Add length_address attribute.
9300 (sse3_monitor_<mode>): Merge from sse3_monitor and
9301 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
9302 Pmode != word_mode. Update insn length attribute.
9303 * config/i386/i386.c (ix86_option_override_internal): Update
9304 ix86_gen_monitor selection for merged sse3_monitor insn.
9306 2013-08-13 Julian Brown <julian@codesourcery.com>
9308 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
9309 perform invalid legitimization on greater-than-word-size modes for
9312 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
9314 * ira.c (setup_class_translate_array): Use aclass instead of cl
9315 for classes not fully covered by allocno classes.
9317 2013-08-13 Jakub Jelinek <jakub@redhat.com>
9319 PR tree-optimization/57661
9320 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
9321 * tree-inline.c (tree_function_versioning): Initialize it.
9322 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
9323 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
9324 that is not being copied.
9327 * asan.c (instrument_strlen_call): Fix typo in comment.
9328 Use char * type even for the lhs of POINTER_PLUS_EXPR.
9330 2013-08-13 Steve Ellcey <sellcey@mips.com>
9332 * config/mips/mips.md (prefetch): Use lw instead of ld on
9333 loongson in 32bit mode.
9335 2013-08-13 Nick Clifton <nickc@redhat.com>
9337 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
9339 2013-08-13 Jan Hubicka <jh@suse.cz>
9341 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
9342 introduced edge; fix typo in sanity check.
9343 (cgraph_resolve_speculation): Export; improve diagnostic.
9344 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
9345 speculation at type mismatch.
9346 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
9347 (cgraph_resolve_speculation): Declare.
9348 (symtab_can_be_discarded): New function.
9349 * value-prof.c (gimple_ic_transform): Remove actual transform code.
9350 * ipa-inline-transform.c (speculation_removed): New global var.
9351 (clone_inlined_nodes): See if speculation can be removed.
9352 (inline_call): If speculations was removed, we growths may not match.
9353 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
9354 (speculation_useful_p): New function.
9355 (resolve_noninline_speculation): New function.
9356 (inline_small_functions): Resolve useless speculations.
9357 * ipa-inline.h (speculation_useful_p): Declare
9358 * ipa.c (can_replace_by_local_alias): Simplify.
9359 (ipa_profile): Produce speculative calls in non-lto, too;
9360 add simple cost model; produce local aliases.
9362 2013-08-13 David Malcolm <dmalcolm@redhat.com>
9364 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
9367 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
9369 * config/i386/i386.c (ix86_function_versions): Use error + inform.
9371 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
9373 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
9374 iterator instead of X87MODEF.
9376 2013-08-12 Perez Read <netfirewall@gmail.com>
9379 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
9380 operand 0 for intel asm alternative.
9381 (*movabs<mode>_2): Ditto for operand 1.
9383 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
9385 * config/aarch64/arm_none.h
9386 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
9388 2013-08-12 Nick Clifton <nickc@redhat.com>
9390 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
9392 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
9394 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
9395 expand for QI/HImode operand to produce more effictive code for
9396 unsigned char(short) --> float(double) conversion.
9398 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
9400 * doc/invoke.texi: Mention that -ftls-model does not force the final
9403 2013-08-12 Marek Polacek <polacek@redhat.com>
9404 Marc Glisse <marc.glisse@inria.fr>
9406 PR tree-optimization/57980
9407 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
9408 when creating -1 constant.
9410 2013-08-10 Jan Hubicka <jh@suse.cz>
9412 Workaround binutils PR14342.
9413 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
9414 (gimple_init_edge_profiler): Likewise.
9415 (gimple_gen_ic_func_profiler): Likewise.
9417 2013-08-09 Jan Hubicka <jh@suse.cz>
9419 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
9421 2013-08-09 Xinliang David Li <davidxl@google.com>
9423 * config/i386/stringop.def: New file.
9424 * config/i386/stringop.opt: New file.
9425 * config/i386/i386-opts.h: Include stringopt.def.
9426 * config/i386/i386.opt: Include stringopt.opt.
9427 * config/i386/i386.c (ix86_option_override_internal):
9428 Override default size based stringop inline strategies with options.
9429 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
9432 2013-08-09 Jan Hubicka <jh@suse.cz>
9434 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
9436 2013-08-09 Jan Hubicka <jh@suse.cz>
9438 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
9440 (dump_cgraph_node): Dump profile-id.
9441 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
9442 and common_target_probability.
9443 * lto-cgraph.c (lto_output_edge): Stream common targets.
9444 (lto_output_node): Stream profile ids.
9445 (input_node): Stream profile ids.
9446 (input_edge): Stream common targets.
9447 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
9448 * ipa.c: Include value-prof.h
9449 (ipa_profile_generate_summary): Turn indirect call statement histograms
9450 into common targets.
9451 (ipa_profile): Turn common targets into speculative edges.
9453 2013-08-09 Jan Hubicka <jh@suse.cz>
9455 * cgraph.h (cgraph_node): Add profile_id.
9456 * value-prof.c (cgraph_node_map): Turn into pointer_map.
9457 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
9458 (del_node_map): Update.
9459 (find_func_by_funcdef_no): Replace by ...
9460 (find_func_by_profile_id): ... this one.
9461 (gimple_ic_transform): Do not remove useful histograms when
9462 speculation is not done; dump info when indirect call removal
9464 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
9465 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
9466 (__gcov_indirect_call_profiler_v2): .. this one.
9467 * profile.h (init_node_map): Update.
9468 * coverage.c (coverage_compute_profile_id): New function.
9469 * coverage.h (coverage_compute_profile_id): Declare.
9470 * tree-profile.c (init_ic_make_global_vars): Make
9471 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
9472 (gimple_init_edge_profiler): Update prototype of
9473 __gcov_indirect_call_profiler.
9474 (gimple_gen_ic_func_profiler): Simplify.
9475 (tree_profiling): Use init_node_map
9477 2013-08-09 Jan Hubicka <jh@suse.cz>
9479 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
9480 non-speculative refs.
9481 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
9482 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
9483 (cgraph_set_call_stmt): Likewise.
9484 (cgraph_create_edge_1): Fix release checking compilatoin;
9486 (cgraph_free_edge): Free indirect info.
9487 (cgraph_turn_edge_to_speculative): New function.
9488 (cgraph_speculative_call_info): New function.
9489 (cgraph_make_edge_direct): Return direct edge; handle speculation.
9490 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
9491 (dump_cgraph_node): Dump speculation.
9492 (verify_edge_count_and_frequency): Accept speculative edges.
9493 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
9494 (verify_cgraph_node): Handle speculation.
9495 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
9496 (cgraph_set_call_stmt): Update prototype.
9497 (cgraph_make_edge_direct): Update prototype.
9498 (cgraph_speculative_call_info): Declare.
9499 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
9500 to change; update call of ipa_find_references.
9501 * ipa-ref.c (ipa_record_reference): Fix return value; clear
9502 lto_stmt_uid and speculative flags.
9503 (ipa_dump_references): Dump speculation.
9504 (ipa_clone_references): Clone speculative flag.
9505 (ipa_clone_referring): Likewise.
9506 (ipa_clone_ref): New function.
9507 (ipa_find_reference): Look into lto_stmt_uids
9508 (ipa_clear_stmts_in_references): Do not clear speculative calls.
9509 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
9510 (ipa_find_reference): Update declaration.
9511 (ipa_clone_ref): Declare.
9512 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
9513 stream speculative flag.
9514 (lto_output_ref): Stream statements uids and speculation.
9515 (input_ref): Likewise.
9516 (input_edge): Stream speuclation.
9517 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
9518 (cgraph_set_call_stmt_including_clones): Handle speculation.
9519 * ipa-inline.c (heap_edge_removal_hook): New function.
9520 (inline_small_functions): Register it.
9521 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
9522 also initialize refs.
9523 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
9525 (try_make_edge_direct_simple_call): Likewise.
9526 (try_make_edge_direct_simple_call): Likewise.
9527 (update_indirect_edges_after_inlining): Likewise.
9528 (remove_described_reference): Look proper lto_stmt_uid.
9529 (propagate_controlled_uses): Likewise.
9530 (propagate_controlled_uses): Liekwise.
9531 * tree-inline.c (copy_bb): Copy speculative edges.
9532 (redirect_all_calls): New function.
9533 (copy_cfg_body): Do redirection after loop info is updated.
9534 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
9536 2013-08-09 Jan Hubicka <jh@suse.cz>
9538 * lto-streamer-out.c (output_function): Renumber PHIs.
9539 * lto-streamer-in.c (input_function): Likewise.
9541 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
9543 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
9544 (get_lane_unsigned): Likewise.
9545 (dup_lane_scalar): Likewise.
9546 (get_lane): enable for VALL.
9547 * config/aarch64/aarch64-simd.md
9548 (aarch64_dup_lane_scalar<mode>): Remove.
9549 (aarch64_get_lane_signed<mode>): Likewise.
9550 (aarch64_get_lane_unsigned<mode>): Likewise.
9551 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
9552 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
9553 (aarch64_get_lane<mode>): Enable for all vector modes.
9554 (aarch64_get_lanedi): Remove misleading constraints.
9555 * config/aarch64/arm_neon.h
9556 (__aarch64_vget_lane_any): Define.
9557 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
9558 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
9559 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
9560 * config/aarch64/iterators.md (VDQQH): New.
9564 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
9566 * configure.ac: Add GAS check for LEON instructions on SPARC.
9567 * configure: Regenerate.
9568 * config.in: Likewise.
9569 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
9571 * config/sparc/sparc.opt (LEON, LEON3): New masks.
9572 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
9574 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
9575 (AS_LEON_FLAG): New macro.
9576 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
9577 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
9578 Deal with LEON and LEON3 for the memory model.
9579 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
9580 (atomic_compare_and_swap<mode>_1): Likewise.
9581 (*atomic_compare_and_swap<mode>_1): Likewise.
9583 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9585 * config/arm/neon.md (vcond): Fix floating-point vector
9586 comparisons against 0.
9588 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
9590 * lra-constraints.c (emit_spill_move): Remove assert.
9591 (process_alt_operands): Add more debugging
9592 output. Increase reject for spilling into memory. Decrease
9593 reject for reloading scratch.
9594 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
9596 2013-08-08 Steve Ellcey <sellcey@mips.com>
9598 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
9599 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
9600 micromips incompatible. Add nan2008.
9601 (MULTILIB_DIRNAMES): Add nan2008.
9602 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9603 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
9604 and micromips incompatible. Add nan2008.
9605 (MULTILIB_DIRNAMES): Add nan2008.
9606 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9608 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
9610 PR rtl-optimization/58079
9611 * combine.c (combine_simplify_rtx): Avoid using SUBST if
9612 simplify_comparison has widened a comparison with an integer.
9614 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9616 * config/arm/neon.md (movmisalign<mode>): Disable when we
9617 don't allow unaligned accesses.
9618 (*movmisalign<mode>_neon_store): Likewise.
9619 (*movmisalign<mode>_neon_load): Likewise.
9620 (*movmisalign<mode>_neon_store): Likewise.
9621 (*movmisalign<mode>_neon_load): Likewise.
9623 2013-08-08 Jan Hubicka <jh@suse.cz>
9625 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
9626 (make_pass_rebuild_cgraph_edges): Also clear references.
9627 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
9628 * ipa-inline-transform.c (inline_transform): Remove all references
9630 * cgraphunit.c (expand_function): Remove all references after
9632 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
9633 (ipa_find_reference): Rewrite to iterator.
9634 (remove_stmt_references): Likewise.
9635 (ipa_clear_stmts_in_references): New function.
9636 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
9637 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
9639 * ipa-split.c (split_function): Remove references in split function.
9641 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9644 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
9645 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
9647 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9650 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
9651 suggested mode for the assignment isn't compatible with the
9654 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9657 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
9659 2013-08-07 Xinliang David Li <davidxl@google.com>
9661 * config/i386/i386.opt: New option -mtune-ctrl=.
9662 * config/i386/x86-tune.def: New file.
9663 * config/i386/i386.h: include x86-tune.def.
9664 * config/i386/i386.c (ix86_option_override_internal):
9665 Parsing -mtune-ctrl= option and set tune features.
9667 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
9670 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
9671 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
9674 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9676 * rtl.h (update_alignments): Declare.
9677 * final.c (grow_label_align): New function extracted from...
9678 (shorten_branches): ...here. Call it.
9679 (update_alignments): New function.
9680 * reorg.c (sibling_labels): New variable.
9681 (get_label_before): Add SIBLING parameter. If it is non-zero, push
9682 the new label along with it onto the sibling_labels vector.
9683 (fill_simple_delay_slots): Adjust call to get_label_before.
9684 (fill_slots_from_thread): Likewise.
9685 (relax_delay_slots): Likewise.
9686 (make_return_insns): Likewise.
9687 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
9689 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9691 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
9692 document its semantics.
9693 (diagnostic_report_diagnostic): Adjust accordingly.
9695 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9697 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
9698 (sparc_option_override): ...and port to new C++ pass API.
9699 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
9701 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
9703 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
9705 2013-08-06 Caroline Tice <cmtice@google.com>
9707 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
9708 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
9709 * tree-pass.h: Add pass_vtable_verify.
9710 * varasm.c (assemble_variable): Add code to properly set the comdat
9711 section and name for the .vtable_map_vars section.
9712 (assemble_vtyv_preinit_initializer): New function.
9713 (default_sectin_type_flags): Make sure .vtable_map_vars section has
9715 * output.h: Add function decl for assemble_vtv_preinit_initializer.
9716 * vtable-verify.c: New file.
9717 * vtable-verify.h: New file.
9718 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
9719 initialiation levels.
9720 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
9721 * passes.def: Insert pass_vtable_verify.
9722 * aclocal.m4: Reorder includes.
9723 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
9724 -fvtv-counts options.
9725 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
9726 as appropriate, if -fvtable-verify=... is used.
9727 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
9728 -fvtable-verify=... is used.
9729 * Makefile.in (OBJS): Add vtable-verify.o to list.
9730 (vtable-verify.o): Add new build rule.
9731 (GTFILES): Add vtable-verify.c to list.
9732 * common.opt (fvtable-verify=): New flag.
9733 (vtv_priority): Values for fvtable-verify= flag.
9734 (fvtv-counts): New flag.
9735 (fvtv-debug): New flag.
9736 * tree.h (save_vtable_map_decl): New extern function decl.
9738 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9740 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
9741 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
9742 (pass_data_rl78_devirt): ...new pass_data instance and...
9743 (make_pass_rl78_devirt): ...new function.
9744 (rl78_asm_file_start): Port pass registration to new C++ API.
9746 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9748 * coretypes.h (rtl_opt_pass): Add.
9749 (gcc::context): Add.
9750 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
9751 (epiphany_init): Port to new C++ pass API.
9752 (epiphany_optimize_mode_switching): Likewise.
9753 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
9754 (pass_manager::get_pass_mode_switching): New.
9755 (pass_manager::get_pass_peephole2): New.
9756 * mode-switching.c (pass_mode_switching): Add clone method.
9757 * recog.c (pass_peephole2): Add clone method.
9758 (pass_split_all_insns): Add clone method.
9760 2013-08-06 David Malcolm <dmalcolm@redhat.com>
9762 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
9763 (mips_option_override): ...here, porting to new C++ API for passes.
9765 2013-08-06 Jan Hubicka <jh@suse.cz>
9767 * cgraph.c (cgraph_get_body): New function based on lto.c
9769 * cgraph.h (cgraph_get_body): Declare.
9770 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
9772 * cgraphunit.c (expand_function): Get body prior expanding.
9773 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
9774 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
9776 * passes.c (do_per_function_toporder): Get body.
9777 * tree-inline.c (expand_call_inline): Get body prior inlining it.
9778 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
9780 2013-08-06 Martin Jambor <mjambor@suse.cz>
9783 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
9784 is not re-finalized. Rename second parameter to no_collect.
9786 2013-08-06 Martin Jambor <mjambor@suse.cz>
9789 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
9790 MEM_REF has proper alignment information.
9792 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
9795 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
9797 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
9798 rtx (*) (rtx, ...) with insn_gen_fn.
9799 * genoutput.c (output_insn_data): Cast gen_? function pointers to
9800 insn_gen_fn::stored_funcptr. Add initializer braces.
9802 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9804 Rewrite how instances of passes are cloned to remove assumptions
9805 about their sizes (thus allowing pass subclasses to have
9806 additional data fields, albeit non-GC-managed ones at this point).
9808 * passes.c (make_pass_instance): Now that passes have clone
9809 methods, rewrite this function to eliminate XNEW and memcpy
9810 calls that used hardcoded sizes. Since this function no longer
9811 creates pass instances, rename it to...
9812 (add_pass_instance): ...this. Document the old way that passes were
9813 numbered and flagged, and rework this function to continue using it.
9814 (next_pass_1): Add an initial_pass argument for use by
9816 (position_pass): When adding multiple instances of a pass, use
9817 the pass's clone method, rather than relying on the XNEW/memcpy
9818 within the former make_pass_instance (now add_pass_instance).
9819 (pass_manager::pass_manager): When invoking next_pass_1, also supply
9820 the initial instance of the current pass within the pass manager.
9822 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9824 This is the automated part of the conversion of passes from C
9825 structs to C++ classes.
9827 Patch autogenerated by refactor_passes.py from
9828 https://github.com/davidmalcolm/gcc-refactoring-scripts
9829 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
9831 * asan.c (pass_asan): Convert from a global struct to a subclass of
9832 gimple_opt_pass along with...
9833 (pass_data_asan): ...new pass_data instance and...
9834 (make_pass_asan): ...new function.
9835 (pass_asan_O0): Convert from a global struct to a subclass of
9836 gimple_opt_pass along with...
9837 (pass_data_asan_O0): ...new pass_data instance and...
9838 (make_pass_asan_O0): ...new function.
9839 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
9840 subclass of rtl_opt_pass along with...
9841 (pass_data_inc_dec): ...new pass_data instance and...
9842 (make_pass_inc_dec): ...new function.
9843 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
9844 a subclass of rtl_opt_pass along with...
9845 (pass_data_reorder_blocks): ...new pass_data instance and...
9846 (make_pass_reorder_blocks): ...new function.
9847 (pass_duplicate_computed_gotos): Convert from a global struct to a
9848 subclass of rtl_opt_pass along with...
9849 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
9850 (make_pass_duplicate_computed_gotos): ...new function.
9851 (pass_partition_blocks): Convert from a global struct to a subclass of
9852 rtl_opt_pass along with...
9853 (pass_data_partition_blocks): ...new pass_data instance and...
9854 (make_pass_partition_blocks): ...new function.
9855 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
9856 struct to a subclass of rtl_opt_pass along with...
9857 (pass_data_branch_target_load_optimize1): ...new pass_data instance
9859 (make_pass_branch_target_load_optimize1): ...new function.
9860 (pass_branch_target_load_optimize2): Convert from a global struct to a
9861 subclass of rtl_opt_pass along with...
9862 (pass_data_branch_target_load_optimize2): ...new pass_data instance
9864 (make_pass_branch_target_load_optimize2): ...new function.
9865 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
9866 of rtl_opt_pass along with...
9867 (pass_data_jump): ...new pass_data instance and...
9868 (make_pass_jump): ...new function.
9869 (pass_jump2): Convert from a global struct to a subclass of
9870 rtl_opt_pass along with...
9871 (pass_data_jump2): ...new pass_data instance and...
9872 (make_pass_jump2): ...new function.
9873 * cfgexpand.c (pass_expand): Convert from a global struct to a
9874 subclass of rtl_opt_pass along with...
9875 (pass_data_expand): ...new pass_data instance and...
9876 (make_pass_expand): ...new function.
9877 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
9878 of rtl_opt_pass along with...
9879 (pass_data_free_cfg): ...new pass_data instance and...
9880 (make_pass_free_cfg): ...new function.
9881 (pass_into_cfg_layout_mode): Convert from a global struct to a
9882 subclass of rtl_opt_pass along with...
9883 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
9884 (make_pass_into_cfg_layout_mode): ...new function.
9885 (pass_outof_cfg_layout_mode): Convert from a global struct to a
9886 subclass of rtl_opt_pass along with...
9887 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
9888 (make_pass_outof_cfg_layout_mode): ...new function.
9889 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
9890 struct to a subclass of gimple_opt_pass along with...
9891 (pass_data_build_cgraph_edges): ...new pass_data instance and...
9892 (make_pass_build_cgraph_edges): ...new function.
9893 (pass_rebuild_cgraph_edges): Convert from a global struct to a
9894 subclass of gimple_opt_pass along with...
9895 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
9896 (make_pass_rebuild_cgraph_edges): ...new function.
9897 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
9898 subclass of gimple_opt_pass along with...
9899 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
9901 (make_pass_remove_cgraph_callee_edges): ...new function.
9902 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
9903 struct to a subclass of rtl_opt_pass along with...
9904 (pass_data_stack_adjustments): ...new pass_data instance and...
9905 (make_pass_stack_adjustments): ...new function.
9906 * combine.c (pass_combine): Convert from a global struct to a subclass
9907 of rtl_opt_pass along with...
9908 (pass_data_combine): ...new pass_data instance and...
9909 (make_pass_combine): ...new function.
9910 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
9911 global struct to a subclass of rtl_opt_pass along with...
9912 (pass_data_compare_elim_after_reload): ...new pass_data instance
9914 (make_pass_compare_elim_after_reload): ...new function.
9915 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
9916 of rtl_opt_pass along with...
9917 (pass_data_rtl_cprop): ...new pass_data instance and...
9918 (make_pass_rtl_cprop): ...new function.
9919 * cse.c (pass_cse): Convert from a global struct to a subclass of
9920 rtl_opt_pass along with...
9921 (pass_data_cse): ...new pass_data instance and...
9922 (make_pass_cse): ...new function.
9923 (pass_cse2): Convert from a global struct to a subclass of
9924 rtl_opt_pass along with...
9925 (pass_data_cse2): ...new pass_data instance and...
9926 (make_pass_cse2): ...new function.
9927 (pass_cse_after_global_opts): Convert from a global struct to a
9928 subclass of rtl_opt_pass along with...
9929 (pass_data_cse_after_global_opts): ...new pass_data instance and...
9930 (make_pass_cse_after_global_opts): ...new function.
9931 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
9932 of rtl_opt_pass along with...
9933 (pass_data_ud_rtl_dce): ...new pass_data instance and...
9934 (make_pass_ud_rtl_dce): ...new function.
9935 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
9936 rtl_opt_pass along with...
9937 (pass_data_fast_rtl_dce): ...new pass_data instance and...
9938 (make_pass_fast_rtl_dce): ...new function.
9939 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
9940 a subclass of rtl_opt_pass along with...
9941 (pass_data_df_initialize_opt): ...new pass_data instance and...
9942 (make_pass_df_initialize_opt): ...new function.
9943 (pass_df_initialize_no_opt): Convert from a global struct to a
9944 subclass of rtl_opt_pass along with...
9945 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
9946 (make_pass_df_initialize_no_opt): ...new function.
9947 (pass_df_finish): Convert from a global struct to a subclass of
9948 rtl_opt_pass along with...
9949 (pass_data_df_finish): ...new pass_data instance and...
9950 (make_pass_df_finish): ...new function.
9951 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
9952 rtl_opt_pass along with...
9953 (pass_data_rtl_dse1): ...new pass_data instance and...
9954 (make_pass_rtl_dse1): ...new function.
9955 (pass_rtl_dse2): Convert from a global struct to a subclass of
9956 rtl_opt_pass along with...
9957 (pass_data_rtl_dse2): ...new pass_data instance and...
9958 (make_pass_rtl_dse2): ...new function.
9959 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
9960 subclass of rtl_opt_pass along with...
9961 (pass_data_dwarf2_frame): ...new pass_data instance and...
9962 (make_pass_dwarf2_frame): ...new function.
9963 * except.c (pass_set_nothrow_function_flags): Convert from a global
9964 struct to a subclass of rtl_opt_pass along with...
9965 (pass_data_set_nothrow_function_flags): ...new pass_data instance
9967 (make_pass_set_nothrow_function_flags): ...new function.
9968 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
9969 subclass of rtl_opt_pass along with...
9970 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
9972 (make_pass_convert_to_eh_region_ranges): ...new function.
9973 * final.c (pass_compute_alignments): Convert from a global struct to a
9974 subclass of rtl_opt_pass along with...
9975 (pass_data_compute_alignments): ...new pass_data instance and...
9976 (make_pass_compute_alignments): ...new function.
9977 (pass_final): Convert from a global struct to a subclass of
9978 rtl_opt_pass along with...
9979 (pass_data_final): ...new pass_data instance and...
9980 (make_pass_final): ...new function.
9981 (pass_shorten_branches): Convert from a global struct to a subclass of
9982 rtl_opt_pass along with...
9983 (pass_data_shorten_branches): ...new pass_data instance and...
9984 (make_pass_shorten_branches): ...new function.
9985 (pass_clean_state): Convert from a global struct to a subclass of
9986 rtl_opt_pass along with...
9987 (pass_data_clean_state): ...new pass_data instance and...
9988 (make_pass_clean_state): ...new function.
9989 * function.c (pass_instantiate_virtual_regs): Convert from a global
9990 struct to a subclass of rtl_opt_pass along with...
9991 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
9992 (make_pass_instantiate_virtual_regs): ...new function.
9993 (pass_leaf_regs): Convert from a global struct to a subclass of
9994 rtl_opt_pass along with...
9995 (pass_data_leaf_regs): ...new pass_data instance and...
9996 (make_pass_leaf_regs): ...new function.
9997 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
9998 subclass of rtl_opt_pass along with...
9999 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
10001 (make_pass_thread_prologue_and_epilogue): ...new function.
10002 (pass_match_asm_constraints): Convert from a global struct to a
10003 subclass of rtl_opt_pass along with...
10004 (pass_data_match_asm_constraints): ...new pass_data instance and...
10005 (make_pass_match_asm_constraints): ...new function.
10006 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
10007 subclass of rtl_opt_pass along with...
10008 (pass_data_rtl_fwprop): ...new pass_data instance and...
10009 (make_pass_rtl_fwprop): ...new function.
10010 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
10011 rtl_opt_pass along with...
10012 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
10013 (make_pass_rtl_fwprop_addr): ...new function.
10014 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
10015 rtl_opt_pass along with...
10016 (pass_data_rtl_pre): ...new pass_data instance and...
10017 (make_pass_rtl_pre): ...new function.
10018 (pass_rtl_hoist): Convert from a global struct to a subclass of
10019 rtl_opt_pass along with...
10020 (pass_data_rtl_hoist): ...new pass_data instance and...
10021 (make_pass_rtl_hoist): ...new function.
10022 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
10023 subclass of gimple_opt_pass along with...
10024 (pass_data_lower_cf): ...new pass_data instance and...
10025 (make_pass_lower_cf): ...new function.
10026 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
10027 from a global struct to a subclass of gimple_opt_pass along with...
10028 (pass_data_strength_reduction): ...new pass_data instance and...
10029 (make_pass_strength_reduction): ...new function.
10030 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
10031 of rtl_opt_pass along with...
10032 (pass_data_rtl_ifcvt): ...new pass_data instance and...
10033 (make_pass_rtl_ifcvt): ...new function.
10034 (pass_if_after_combine): Convert from a global struct to a subclass of
10035 rtl_opt_pass along with...
10036 (pass_data_if_after_combine): ...new pass_data instance and...
10037 (make_pass_if_after_combine): ...new function.
10038 (pass_if_after_reload): Convert from a global struct to a subclass of
10039 rtl_opt_pass along with...
10040 (pass_data_if_after_reload): ...new pass_data instance and...
10041 (make_pass_if_after_reload): ...new function.
10042 * init-regs.c (pass_initialize_regs): Convert from a global struct to
10043 a subclass of rtl_opt_pass along with...
10044 (pass_data_initialize_regs): ...new pass_data instance and...
10045 (make_pass_initialize_regs): ...new function.
10046 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
10047 of ipa_opt_pass_d along with...
10048 (pass_data_ipa_cp): ...new pass_data instance and...
10049 (make_pass_ipa_cp): ...new function.
10050 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
10051 global struct to a subclass of gimple_opt_pass along with...
10052 (pass_data_inline_parameters): ...new pass_data instance and...
10053 (make_pass_inline_parameters): ...new function.
10054 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
10055 subclass of gimple_opt_pass along with...
10056 (pass_data_early_inline): ...new pass_data instance and...
10057 (make_pass_early_inline): ...new function.
10058 (pass_ipa_inline): Convert from a global struct to a subclass of
10059 ipa_opt_pass_d along with...
10060 (pass_data_ipa_inline): ...new pass_data instance and...
10061 (make_pass_ipa_inline): ...new function.
10062 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
10063 struct to a subclass of gimple_opt_pass along with...
10064 (pass_data_local_pure_const): ...new pass_data instance and...
10065 (make_pass_local_pure_const): ...new function.
10066 (pass_ipa_pure_const): Convert from a global struct to a subclass of
10067 ipa_opt_pass_d along with...
10068 (pass_data_ipa_pure_const): ...new pass_data instance and...
10069 (make_pass_ipa_pure_const): ...new function.
10070 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
10071 to a subclass of ipa_opt_pass_d along with...
10072 (pass_data_ipa_reference): ...new pass_data instance and...
10073 (make_pass_ipa_reference): ...new function.
10074 * ipa-split.c (pass_split_functions): Convert from a global struct to
10075 a subclass of gimple_opt_pass along with...
10076 (pass_data_split_functions): ...new pass_data instance and...
10077 (make_pass_split_functions): ...new function.
10078 (pass_feedback_split_functions): Convert from a global struct to a
10079 subclass of gimple_opt_pass along with...
10080 (pass_data_feedback_split_functions): ...new pass_data instance and...
10081 (make_pass_feedback_split_functions): ...new function.
10082 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
10083 global struct to a subclass of simple_ipa_opt_pass along with...
10084 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
10086 (make_pass_ipa_function_and_variable_visibility): ...new function.
10087 (pass_ipa_free_inline_summary): Convert from a global struct to a
10088 subclass of simple_ipa_opt_pass along with...
10089 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
10090 (make_pass_ipa_free_inline_summary): ...new function.
10091 (pass_ipa_whole_program_visibility): Convert from a global struct to a
10092 subclass of ipa_opt_pass_d along with...
10093 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
10095 (make_pass_ipa_whole_program_visibility): ...new function.
10096 (pass_ipa_profile): Convert from a global struct to a subclass of
10097 ipa_opt_pass_d along with...
10098 (pass_data_ipa_profile): ...new pass_data instance and...
10099 (make_pass_ipa_profile): ...new function.
10100 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
10101 ipa_opt_pass_d along with...
10102 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
10103 (make_pass_ipa_cdtor_merge): ...new function.
10104 * ira.c (pass_ira): Convert from a global struct to a subclass of
10105 rtl_opt_pass along with...
10106 (pass_data_ira): ...new pass_data instance and...
10107 (make_pass_ira): ...new function.
10108 (pass_reload): Convert from a global struct to a subclass of
10109 rtl_opt_pass along with...
10110 (pass_data_reload): ...new pass_data instance and...
10111 (make_pass_reload): ...new function.
10112 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
10113 subclass of rtl_opt_pass along with...
10114 (pass_data_cleanup_barriers): ...new pass_data instance and...
10115 (make_pass_cleanup_barriers): ...new function.
10116 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
10117 of rtl_opt_pass along with...
10118 (pass_data_loop2): ...new pass_data instance and...
10119 (make_pass_loop2): ...new function.
10120 (pass_rtl_loop_init): Convert from a global struct to a subclass of
10121 rtl_opt_pass along with...
10122 (pass_data_rtl_loop_init): ...new pass_data instance and...
10123 (make_pass_rtl_loop_init): ...new function.
10124 (pass_rtl_loop_done): Convert from a global struct to a subclass of
10125 rtl_opt_pass along with...
10126 (pass_data_rtl_loop_done): ...new pass_data instance and...
10127 (make_pass_rtl_loop_done): ...new function.
10128 (pass_rtl_move_loop_invariants): Convert from a global struct to a
10129 subclass of rtl_opt_pass along with...
10130 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
10131 (make_pass_rtl_move_loop_invariants): ...new function.
10132 (pass_rtl_unswitch): Convert from a global struct to a subclass of
10133 rtl_opt_pass along with...
10134 (pass_data_rtl_unswitch): ...new pass_data instance and...
10135 (make_pass_rtl_unswitch): ...new function.
10136 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
10137 subclass of rtl_opt_pass along with...
10138 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
10140 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10141 (pass_rtl_doloop): Convert from a global struct to a subclass of
10142 rtl_opt_pass along with...
10143 (pass_data_rtl_doloop): ...new pass_data instance and...
10144 (make_pass_rtl_doloop): ...new function.
10145 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
10146 a subclass of rtl_opt_pass along with...
10147 (pass_data_lower_subreg): ...new pass_data instance and...
10148 (make_pass_lower_subreg): ...new function.
10149 (pass_lower_subreg2): Convert from a global struct to a subclass of
10150 rtl_opt_pass along with...
10151 (pass_data_lower_subreg2): ...new pass_data instance and...
10152 (make_pass_lower_subreg2): ...new function.
10153 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
10154 struct to a subclass of ipa_opt_pass_d along with...
10155 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
10156 (make_pass_ipa_lto_gimple_out): ...new function.
10157 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
10158 of ipa_opt_pass_d along with...
10159 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
10160 (make_pass_ipa_lto_finish_out): ...new function.
10161 * mode-switching.c (pass_mode_switching): Convert from a global struct
10162 to a subclass of rtl_opt_pass along with...
10163 (pass_data_mode_switching): ...new pass_data instance and...
10164 (make_pass_mode_switching): ...new function.
10165 * modulo-sched.c (pass_sms): Convert from a global struct to a
10166 subclass of rtl_opt_pass along with...
10167 (pass_data_sms): ...new pass_data instance and...
10168 (make_pass_sms): ...new function.
10169 * omp-low.c (pass_expand_omp): Convert from a global struct to a
10170 subclass of gimple_opt_pass along with...
10171 (pass_data_expand_omp): ...new pass_data instance and...
10172 (make_pass_expand_omp): ...new function.
10173 (pass_lower_omp): Convert from a global struct to a subclass of
10174 gimple_opt_pass along with...
10175 (pass_data_lower_omp): ...new pass_data instance and...
10176 (make_pass_lower_omp): ...new function.
10177 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
10178 of gimple_opt_pass along with...
10179 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
10180 (make_pass_diagnose_omp_blocks): ...new function.
10181 * passes.c (pass_early_local_passes): Convert from a global struct to
10182 a subclass of simple_ipa_opt_pass along with...
10183 (pass_data_early_local_passes): ...new pass_data instance and...
10184 (make_pass_early_local_passes): ...new function.
10185 (pass_all_early_optimizations): Convert from a global struct to a
10186 subclass of gimple_opt_pass along with...
10187 (pass_data_all_early_optimizations): ...new pass_data instance and...
10188 (make_pass_all_early_optimizations): ...new function.
10189 (pass_all_optimizations): Convert from a global struct to a subclass
10190 of gimple_opt_pass along with...
10191 (pass_data_all_optimizations): ...new pass_data instance and...
10192 (make_pass_all_optimizations): ...new function.
10193 (pass_all_optimizations_g): Convert from a global struct to a subclass
10194 of gimple_opt_pass along with...
10195 (pass_data_all_optimizations_g): ...new pass_data instance and...
10196 (make_pass_all_optimizations_g): ...new function.
10197 (pass_rest_of_compilation): Convert from a global struct to a subclass
10198 of rtl_opt_pass along with...
10199 (pass_data_rest_of_compilation): ...new pass_data instance and...
10200 (make_pass_rest_of_compilation): ...new function.
10201 (pass_postreload): Convert from a global struct to a subclass of
10202 rtl_opt_pass along with...
10203 (pass_data_postreload): ...new pass_data instance and...
10204 (make_pass_postreload): ...new function.
10205 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
10206 subclass of rtl_opt_pass along with...
10207 (pass_data_gcse2): ...new pass_data instance and...
10208 (make_pass_gcse2): ...new function.
10209 * postreload.c (pass_postreload_cse): Convert from a global struct to
10210 a subclass of rtl_opt_pass along with...
10211 (pass_data_postreload_cse): ...new pass_data instance and...
10212 (make_pass_postreload_cse): ...new function.
10213 * predict.c (pass_profile): Convert from a global struct to a subclass
10214 of gimple_opt_pass along with...
10215 (pass_data_profile): ...new pass_data instance and...
10216 (make_pass_profile): ...new function.
10217 (pass_strip_predict_hints): Convert from a global struct to a subclass
10218 of gimple_opt_pass along with...
10219 (pass_data_strip_predict_hints): ...new pass_data instance and...
10220 (make_pass_strip_predict_hints): ...new function.
10221 * recog.c (pass_peephole2): Convert from a global struct to a subclass
10222 of rtl_opt_pass along with...
10223 (pass_data_peephole2): ...new pass_data instance and...
10224 (make_pass_peephole2): ...new function.
10225 (pass_split_all_insns): Convert from a global struct to a subclass of
10226 rtl_opt_pass along with...
10227 (pass_data_split_all_insns): ...new pass_data instance and...
10228 (make_pass_split_all_insns): ...new function.
10229 (pass_split_after_reload): Convert from a global struct to a subclass
10230 of rtl_opt_pass along with...
10231 (pass_data_split_after_reload): ...new pass_data instance and...
10232 (make_pass_split_after_reload): ...new function.
10233 (pass_split_before_regstack): Convert from a global struct to a
10234 subclass of rtl_opt_pass along with...
10235 (pass_data_split_before_regstack): ...new pass_data instance and...
10236 (make_pass_split_before_regstack): ...new function.
10237 (pass_split_before_sched2): Convert from a global struct to a subclass
10238 of rtl_opt_pass along with...
10239 (pass_data_split_before_sched2): ...new pass_data instance and...
10240 (make_pass_split_before_sched2): ...new function.
10241 (pass_split_for_shorten_branches): Convert from a global struct to a
10242 subclass of rtl_opt_pass along with...
10243 (pass_data_split_for_shorten_branches): ...new pass_data instance
10245 (make_pass_split_for_shorten_branches): ...new function.
10246 * ree.c (pass_ree): Convert from a global struct to a subclass of
10247 rtl_opt_pass along with...
10248 (pass_data_ree): ...new pass_data instance and...
10249 (make_pass_ree): ...new function.
10250 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
10251 subclass of rtl_opt_pass along with...
10252 (pass_data_stack_regs): ...new pass_data instance and...
10253 (make_pass_stack_regs): ...new function.
10254 (pass_stack_regs_run): Convert from a global struct to a subclass of
10255 rtl_opt_pass along with...
10256 (pass_data_stack_regs_run): ...new pass_data instance and...
10257 (make_pass_stack_regs_run): ...new function.
10258 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
10259 subclass of rtl_opt_pass along with...
10260 (pass_data_cprop_hardreg): ...new pass_data instance and...
10261 (make_pass_cprop_hardreg): ...new function.
10262 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
10263 subclass of rtl_opt_pass along with...
10264 (pass_data_reginfo_init): ...new pass_data instance and...
10265 (make_pass_reginfo_init): ...new function.
10266 * regmove.c (pass_regmove): Convert from a global struct to a subclass
10267 of rtl_opt_pass along with...
10268 (pass_data_regmove): ...new pass_data instance and...
10269 (make_pass_regmove): ...new function.
10270 * regrename.c (pass_regrename): Convert from a global struct to a
10271 subclass of rtl_opt_pass along with...
10272 (pass_data_regrename): ...new pass_data instance and...
10273 (make_pass_regrename): ...new function.
10274 * reorg.c (pass_delay_slots): Convert from a global struct to a
10275 subclass of rtl_opt_pass along with...
10276 (pass_data_delay_slots): ...new pass_data instance and...
10277 (make_pass_delay_slots): ...new function.
10278 (pass_machine_reorg): Convert from a global struct to a subclass of
10279 rtl_opt_pass along with...
10280 (pass_data_machine_reorg): ...new pass_data instance and...
10281 (make_pass_machine_reorg): ...new function.
10282 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
10283 of rtl_opt_pass along with...
10284 (pass_data_sched): ...new pass_data instance and...
10285 (make_pass_sched): ...new function.
10286 (pass_sched2): Convert from a global struct to a subclass of
10287 rtl_opt_pass along with...
10288 (pass_data_sched2): ...new pass_data instance and...
10289 (make_pass_sched2): ...new function.
10290 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
10291 to a subclass of rtl_opt_pass along with...
10292 (pass_data_stack_ptr_mod): ...new pass_data instance and...
10293 (make_pass_stack_ptr_mod): ...new function.
10294 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
10295 to a subclass of rtl_opt_pass along with...
10296 (pass_data_rtl_store_motion): ...new pass_data instance and...
10297 (make_pass_rtl_store_motion): ...new function.
10298 * tracer.c (pass_tracer): Convert from a global struct to a subclass
10299 of gimple_opt_pass along with...
10300 (pass_data_tracer): ...new pass_data instance and...
10301 (make_pass_tracer): ...new function.
10302 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
10303 to a subclass of gimple_opt_pass along with...
10304 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
10305 (make_pass_diagnose_tm_blocks): ...new function.
10306 (pass_lower_tm): Convert from a global struct to a subclass of
10307 gimple_opt_pass along with...
10308 (pass_data_lower_tm): ...new pass_data instance and...
10309 (make_pass_lower_tm): ...new function.
10310 (pass_tm_init): Convert from a global struct to a subclass of
10311 gimple_opt_pass along with...
10312 (pass_data_tm_init): ...new pass_data instance and...
10313 (make_pass_tm_init): ...new function.
10314 (pass_tm_mark): Convert from a global struct to a subclass of
10315 gimple_opt_pass along with...
10316 (pass_data_tm_mark): ...new pass_data instance and...
10317 (make_pass_tm_mark): ...new function.
10318 (pass_tm_edges): Convert from a global struct to a subclass of
10319 gimple_opt_pass along with...
10320 (pass_data_tm_edges): ...new pass_data instance and...
10321 (make_pass_tm_edges): ...new function.
10322 (pass_tm_memopt): Convert from a global struct to a subclass of
10323 gimple_opt_pass along with...
10324 (pass_data_tm_memopt): ...new pass_data instance and...
10325 (make_pass_tm_memopt): ...new function.
10326 (pass_ipa_tm): Convert from a global struct to a subclass of
10327 simple_ipa_opt_pass along with...
10328 (pass_data_ipa_tm): ...new pass_data instance and...
10329 (make_pass_ipa_tm): ...new function.
10330 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
10331 subclass of gimple_opt_pass along with...
10332 (pass_data_call_cdce): ...new pass_data instance and...
10333 (make_pass_call_cdce): ...new function.
10334 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
10335 subclass of gimple_opt_pass along with...
10336 (pass_data_build_cfg): ...new pass_data instance and...
10337 (make_pass_build_cfg): ...new function.
10338 (pass_split_crit_edges): Convert from a global struct to a subclass of
10339 gimple_opt_pass along with...
10340 (pass_data_split_crit_edges): ...new pass_data instance and...
10341 (make_pass_split_crit_edges): ...new function.
10342 (pass_warn_function_return): Convert from a global struct to a
10343 subclass of gimple_opt_pass along with...
10344 (pass_data_warn_function_return): ...new pass_data instance and...
10345 (make_pass_warn_function_return): ...new function.
10346 (pass_warn_function_noreturn): Convert from a global struct to a
10347 subclass of gimple_opt_pass along with...
10348 (pass_data_warn_function_noreturn): ...new pass_data instance and...
10349 (make_pass_warn_function_noreturn): ...new function.
10350 (pass_warn_unused_result): Convert from a global struct to a subclass
10351 of gimple_opt_pass along with...
10352 (pass_data_warn_unused_result): ...new pass_data instance and...
10353 (make_pass_warn_unused_result): ...new function.
10354 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
10355 a subclass of gimple_opt_pass along with...
10356 (pass_data_merge_phi): ...new pass_data instance and...
10357 (make_pass_merge_phi): ...new function.
10358 * tree-complex.c (pass_lower_complex): Convert from a global struct to
10359 a subclass of gimple_opt_pass along with...
10360 (pass_data_lower_complex): ...new pass_data instance and...
10361 (make_pass_lower_complex): ...new function.
10362 (pass_lower_complex_O0): Convert from a global struct to a subclass of
10363 gimple_opt_pass along with...
10364 (pass_data_lower_complex_O0): ...new pass_data instance and...
10365 (make_pass_lower_complex_O0): ...new function.
10366 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
10367 subclass of gimple_opt_pass along with...
10368 (pass_data_lower_eh): ...new pass_data instance and...
10369 (make_pass_lower_eh): ...new function.
10370 (pass_refactor_eh): Convert from a global struct to a subclass of
10371 gimple_opt_pass along with...
10372 (pass_data_refactor_eh): ...new pass_data instance and...
10373 (make_pass_refactor_eh): ...new function.
10374 (pass_lower_resx): Convert from a global struct to a subclass of
10375 gimple_opt_pass along with...
10376 (pass_data_lower_resx): ...new pass_data instance and...
10377 (make_pass_lower_resx): ...new function.
10378 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
10379 of gimple_opt_pass along with...
10380 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
10381 (make_pass_lower_eh_dispatch): ...new function.
10382 (pass_cleanup_eh): Convert from a global struct to a subclass of
10383 gimple_opt_pass along with...
10384 (pass_data_cleanup_eh): ...new pass_data instance and...
10385 (make_pass_cleanup_eh): ...new function.
10386 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
10387 to a subclass of simple_ipa_opt_pass along with...
10388 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
10389 (make_pass_ipa_lower_emutls): ...new function.
10390 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
10391 a subclass of gimple_opt_pass along with...
10392 (pass_data_if_conversion): ...new pass_data instance and...
10393 (make_pass_if_conversion): ...new function.
10394 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
10395 subclass of gimple_opt_pass along with...
10396 (pass_data_build_ssa): ...new pass_data instance and...
10397 (make_pass_build_ssa): ...new function.
10398 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
10399 global struct to a subclass of gimple_opt_pass along with...
10400 (pass_data_loop_distribution): ...new pass_data instance and...
10401 (make_pass_loop_distribution): ...new function.
10402 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
10403 subclass of gimple_opt_pass along with...
10404 (pass_data_mudflap_1): ...new pass_data instance and...
10405 (make_pass_mudflap_1): ...new function.
10406 (pass_mudflap_2): Convert from a global struct to a subclass of
10407 gimple_opt_pass along with...
10408 (pass_data_mudflap_2): ...new pass_data instance and...
10409 (make_pass_mudflap_2): ...new function.
10410 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
10411 subclass of gimple_opt_pass along with...
10412 (pass_data_mudflap_1): ...new pass_data instance and...
10413 (make_pass_mudflap_1): ...new function.
10414 (pass_mudflap_2): Convert from a global struct to a subclass of
10415 gimple_opt_pass along with...
10416 (pass_data_mudflap_2): ...new pass_data instance and...
10417 (make_pass_mudflap_2): ...new function.
10418 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
10419 gimple_opt_pass along with...
10420 (pass_data_nrv): ...new pass_data instance and...
10421 (make_pass_nrv): ...new function.
10422 (pass_return_slot): Convert from a global struct to a subclass of
10423 gimple_opt_pass along with...
10424 (pass_data_return_slot): ...new pass_data instance and...
10425 (make_pass_return_slot): ...new function.
10426 * tree-object-size.c (pass_object_sizes): Convert from a global struct
10427 to a subclass of gimple_opt_pass along with...
10428 (pass_data_object_sizes): ...new pass_data instance and...
10429 (make_pass_object_sizes): ...new function.
10430 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
10431 global struct to a subclass of gimple_opt_pass along with...
10432 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
10434 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10435 (pass_fixup_cfg): Convert from a global struct to a subclass of
10436 gimple_opt_pass along with...
10437 (pass_data_fixup_cfg): ...new pass_data instance and...
10438 (make_pass_fixup_cfg): ...new function.
10439 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
10440 (make_pass_mudflap_1): ...new function.
10441 (pass_mudflap_2): Replace declaration with that of...
10442 (make_pass_mudflap_2): ...new function.
10443 (pass_asan): Replace declaration with that of...
10444 (make_pass_asan): ...new function.
10445 (pass_asan_O0): Replace declaration with that of...
10446 (make_pass_asan_O0): ...new function.
10447 (pass_tsan): Replace declaration with that of...
10448 (make_pass_tsan): ...new function.
10449 (pass_tsan_O0): Replace declaration with that of...
10450 (make_pass_tsan_O0): ...new function.
10451 (pass_lower_cf): Replace declaration with that of...
10452 (make_pass_lower_cf): ...new function.
10453 (pass_refactor_eh): Replace declaration with that of...
10454 (make_pass_refactor_eh): ...new function.
10455 (pass_lower_eh): Replace declaration with that of...
10456 (make_pass_lower_eh): ...new function.
10457 (pass_lower_eh_dispatch): Replace declaration with that of...
10458 (make_pass_lower_eh_dispatch): ...new function.
10459 (pass_lower_resx): Replace declaration with that of...
10460 (make_pass_lower_resx): ...new function.
10461 (pass_build_cfg): Replace declaration with that of...
10462 (make_pass_build_cfg): ...new function.
10463 (pass_early_tree_profile): Replace declaration with that of...
10464 (make_pass_early_tree_profile): ...new function.
10465 (pass_cleanup_eh): Replace declaration with that of...
10466 (make_pass_cleanup_eh): ...new function.
10467 (pass_sra): Replace declaration with that of...
10468 (make_pass_sra): ...new function.
10469 (pass_sra_early): Replace declaration with that of...
10470 (make_pass_sra_early): ...new function.
10471 (pass_early_ipa_sra): Replace declaration with that of...
10472 (make_pass_early_ipa_sra): ...new function.
10473 (pass_tail_recursion): Replace declaration with that of...
10474 (make_pass_tail_recursion): ...new function.
10475 (pass_tail_calls): Replace declaration with that of...
10476 (make_pass_tail_calls): ...new function.
10477 (pass_tree_loop): Replace declaration with that of...
10478 (make_pass_tree_loop): ...new function.
10479 (pass_tree_loop_init): Replace declaration with that of...
10480 (make_pass_tree_loop_init): ...new function.
10481 (pass_lim): Replace declaration with that of...
10482 (make_pass_lim): ...new function.
10483 (pass_tree_unswitch): Replace declaration with that of...
10484 (make_pass_tree_unswitch): ...new function.
10485 (pass_predcom): Replace declaration with that of...
10486 (make_pass_predcom): ...new function.
10487 (pass_iv_canon): Replace declaration with that of...
10488 (make_pass_iv_canon): ...new function.
10489 (pass_scev_cprop): Replace declaration with that of...
10490 (make_pass_scev_cprop): ...new function.
10491 (pass_empty_loop): Replace declaration with that of...
10492 (make_pass_empty_loop): ...new function.
10493 (pass_record_bounds): Replace declaration with that of...
10494 (make_pass_record_bounds): ...new function.
10495 (pass_graphite): Replace declaration with that of...
10496 (make_pass_graphite): ...new function.
10497 (pass_graphite_transforms): Replace declaration with that of...
10498 (make_pass_graphite_transforms): ...new function.
10499 (pass_if_conversion): Replace declaration with that of...
10500 (make_pass_if_conversion): ...new function.
10501 (pass_loop_distribution): Replace declaration with that of...
10502 (make_pass_loop_distribution): ...new function.
10503 (pass_vectorize): Replace declaration with that of...
10504 (make_pass_vectorize): ...new function.
10505 (pass_slp_vectorize): Replace declaration with that of...
10506 (make_pass_slp_vectorize): ...new function.
10507 (pass_complete_unroll): Replace declaration with that of...
10508 (make_pass_complete_unroll): ...new function.
10509 (pass_complete_unrolli): Replace declaration with that of...
10510 (make_pass_complete_unrolli): ...new function.
10511 (pass_parallelize_loops): Replace declaration with that of...
10512 (make_pass_parallelize_loops): ...new function.
10513 (pass_loop_prefetch): Replace declaration with that of...
10514 (make_pass_loop_prefetch): ...new function.
10515 (pass_iv_optimize): Replace declaration with that of...
10516 (make_pass_iv_optimize): ...new function.
10517 (pass_tree_loop_done): Replace declaration with that of...
10518 (make_pass_tree_loop_done): ...new function.
10519 (pass_ch): Replace declaration with that of...
10520 (make_pass_ch): ...new function.
10521 (pass_ccp): Replace declaration with that of...
10522 (make_pass_ccp): ...new function.
10523 (pass_phi_only_cprop): Replace declaration with that of...
10524 (make_pass_phi_only_cprop): ...new function.
10525 (pass_build_ssa): Replace declaration with that of...
10526 (make_pass_build_ssa): ...new function.
10527 (pass_build_alias): Replace declaration with that of...
10528 (make_pass_build_alias): ...new function.
10529 (pass_build_ealias): Replace declaration with that of...
10530 (make_pass_build_ealias): ...new function.
10531 (pass_dominator): Replace declaration with that of...
10532 (make_pass_dominator): ...new function.
10533 (pass_dce): Replace declaration with that of...
10534 (make_pass_dce): ...new function.
10535 (pass_dce_loop): Replace declaration with that of...
10536 (make_pass_dce_loop): ...new function.
10537 (pass_cd_dce): Replace declaration with that of...
10538 (make_pass_cd_dce): ...new function.
10539 (pass_call_cdce): Replace declaration with that of...
10540 (make_pass_call_cdce): ...new function.
10541 (pass_merge_phi): Replace declaration with that of...
10542 (make_pass_merge_phi): ...new function.
10543 (pass_split_crit_edges): Replace declaration with that of...
10544 (make_pass_split_crit_edges): ...new function.
10545 (pass_pre): Replace declaration with that of...
10546 (make_pass_pre): ...new function.
10547 (pass_profile): Replace declaration with that of...
10548 (make_pass_profile): ...new function.
10549 (pass_strip_predict_hints): Replace declaration with that of...
10550 (make_pass_strip_predict_hints): ...new function.
10551 (pass_lower_complex_O0): Replace declaration with that of...
10552 (make_pass_lower_complex_O0): ...new function.
10553 (pass_lower_complex): Replace declaration with that of...
10554 (make_pass_lower_complex): ...new function.
10555 (pass_lower_vector): Replace declaration with that of...
10556 (make_pass_lower_vector): ...new function.
10557 (pass_lower_vector_ssa): Replace declaration with that of...
10558 (make_pass_lower_vector_ssa): ...new function.
10559 (pass_lower_omp): Replace declaration with that of...
10560 (make_pass_lower_omp): ...new function.
10561 (pass_diagnose_omp_blocks): Replace declaration with that of...
10562 (make_pass_diagnose_omp_blocks): ...new function.
10563 (pass_expand_omp): Replace declaration with that of...
10564 (make_pass_expand_omp): ...new function.
10565 (pass_expand_omp_ssa): Replace declaration with that of...
10566 (make_pass_expand_omp_ssa): ...new function.
10567 (pass_object_sizes): Replace declaration with that of...
10568 (make_pass_object_sizes): ...new function.
10569 (pass_strlen): Replace declaration with that of...
10570 (make_pass_strlen): ...new function.
10571 (pass_fold_builtins): Replace declaration with that of...
10572 (make_pass_fold_builtins): ...new function.
10573 (pass_stdarg): Replace declaration with that of...
10574 (make_pass_stdarg): ...new function.
10575 (pass_early_warn_uninitialized): Replace declaration with that of...
10576 (make_pass_early_warn_uninitialized): ...new function.
10577 (pass_late_warn_uninitialized): Replace declaration with that of...
10578 (make_pass_late_warn_uninitialized): ...new function.
10579 (pass_cse_reciprocals): Replace declaration with that of...
10580 (make_pass_cse_reciprocals): ...new function.
10581 (pass_cse_sincos): Replace declaration with that of...
10582 (make_pass_cse_sincos): ...new function.
10583 (pass_optimize_bswap): Replace declaration with that of...
10584 (make_pass_optimize_bswap): ...new function.
10585 (pass_optimize_widening_mul): Replace declaration with that of...
10586 (make_pass_optimize_widening_mul): ...new function.
10587 (pass_warn_function_return): Replace declaration with that of...
10588 (make_pass_warn_function_return): ...new function.
10589 (pass_warn_function_noreturn): Replace declaration with that of...
10590 (make_pass_warn_function_noreturn): ...new function.
10591 (pass_cselim): Replace declaration with that of...
10592 (make_pass_cselim): ...new function.
10593 (pass_phiopt): Replace declaration with that of...
10594 (make_pass_phiopt): ...new function.
10595 (pass_forwprop): Replace declaration with that of...
10596 (make_pass_forwprop): ...new function.
10597 (pass_phiprop): Replace declaration with that of...
10598 (make_pass_phiprop): ...new function.
10599 (pass_tree_ifcombine): Replace declaration with that of...
10600 (make_pass_tree_ifcombine): ...new function.
10601 (pass_dse): Replace declaration with that of...
10602 (make_pass_dse): ...new function.
10603 (pass_nrv): Replace declaration with that of...
10604 (make_pass_nrv): ...new function.
10605 (pass_rename_ssa_copies): Replace declaration with that of...
10606 (make_pass_rename_ssa_copies): ...new function.
10607 (pass_sink_code): Replace declaration with that of...
10608 (make_pass_sink_code): ...new function.
10609 (pass_fre): Replace declaration with that of...
10610 (make_pass_fre): ...new function.
10611 (pass_check_data_deps): Replace declaration with that of...
10612 (make_pass_check_data_deps): ...new function.
10613 (pass_copy_prop): Replace declaration with that of...
10614 (make_pass_copy_prop): ...new function.
10615 (pass_vrp): Replace declaration with that of...
10616 (make_pass_vrp): ...new function.
10617 (pass_uncprop): Replace declaration with that of...
10618 (make_pass_uncprop): ...new function.
10619 (pass_return_slot): Replace declaration with that of...
10620 (make_pass_return_slot): ...new function.
10621 (pass_reassoc): Replace declaration with that of...
10622 (make_pass_reassoc): ...new function.
10623 (pass_rebuild_cgraph_edges): Replace declaration with that of...
10624 (make_pass_rebuild_cgraph_edges): ...new function.
10625 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
10626 (make_pass_remove_cgraph_callee_edges): ...new function.
10627 (pass_build_cgraph_edges): Replace declaration with that of...
10628 (make_pass_build_cgraph_edges): ...new function.
10629 (pass_local_pure_const): Replace declaration with that of...
10630 (make_pass_local_pure_const): ...new function.
10631 (pass_tracer): Replace declaration with that of...
10632 (make_pass_tracer): ...new function.
10633 (pass_warn_unused_result): Replace declaration with that of...
10634 (make_pass_warn_unused_result): ...new function.
10635 (pass_diagnose_tm_blocks): Replace declaration with that of...
10636 (make_pass_diagnose_tm_blocks): ...new function.
10637 (pass_lower_tm): Replace declaration with that of...
10638 (make_pass_lower_tm): ...new function.
10639 (pass_tm_init): Replace declaration with that of...
10640 (make_pass_tm_init): ...new function.
10641 (pass_tm_mark): Replace declaration with that of...
10642 (make_pass_tm_mark): ...new function.
10643 (pass_tm_memopt): Replace declaration with that of...
10644 (make_pass_tm_memopt): ...new function.
10645 (pass_tm_edges): Replace declaration with that of...
10646 (make_pass_tm_edges): ...new function.
10647 (pass_split_functions): Replace declaration with that of...
10648 (make_pass_split_functions): ...new function.
10649 (pass_feedback_split_functions): Replace declaration with that of...
10650 (make_pass_feedback_split_functions): ...new function.
10651 (pass_strength_reduction): Replace declaration with that of...
10652 (make_pass_strength_reduction): ...new function.
10653 (pass_ipa_lower_emutls): Replace declaration with that of...
10654 (make_pass_ipa_lower_emutls): ...new function.
10655 (pass_ipa_function_and_variable_visibility): Replace declaration with
10657 (make_pass_ipa_function_and_variable_visibility): ...new function.
10658 (pass_ipa_tree_profile): Replace declaration with that of...
10659 (make_pass_ipa_tree_profile): ...new function.
10660 (pass_early_local_passes): Replace declaration with that of...
10661 (make_pass_early_local_passes): ...new function.
10662 (pass_ipa_whole_program_visibility): Replace declaration with that
10664 (make_pass_ipa_whole_program_visibility): ...new function.
10665 (pass_ipa_lto_gimple_out): Replace declaration with that of...
10666 (make_pass_ipa_lto_gimple_out): ...new function.
10667 (pass_ipa_increase_alignment): Replace declaration with that of...
10668 (make_pass_ipa_increase_alignment): ...new function.
10669 (pass_ipa_inline): Replace declaration with that of...
10670 (make_pass_ipa_inline): ...new function.
10671 (pass_ipa_free_lang_data): Replace declaration with that of...
10672 (make_pass_ipa_free_lang_data): ...new function.
10673 (pass_ipa_free_inline_summary): Replace declaration with that of...
10674 (make_pass_ipa_free_inline_summary): ...new function.
10675 (pass_ipa_cp): Replace declaration with that of...
10676 (make_pass_ipa_cp): ...new function.
10677 (pass_ipa_reference): Replace declaration with that of...
10678 (make_pass_ipa_reference): ...new function.
10679 (pass_ipa_pure_const): Replace declaration with that of...
10680 (make_pass_ipa_pure_const): ...new function.
10681 (pass_ipa_pta): Replace declaration with that of...
10682 (make_pass_ipa_pta): ...new function.
10683 (pass_ipa_lto_finish_out): Replace declaration with that of...
10684 (make_pass_ipa_lto_finish_out): ...new function.
10685 (pass_ipa_tm): Replace declaration with that of...
10686 (make_pass_ipa_tm): ...new function.
10687 (pass_ipa_profile): Replace declaration with that of...
10688 (make_pass_ipa_profile): ...new function.
10689 (pass_ipa_cdtor_merge): Replace declaration with that of...
10690 (make_pass_ipa_cdtor_merge): ...new function.
10691 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
10693 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10694 (pass_init_datastructures): Replace declaration with that of...
10695 (make_pass_init_datastructures): ...new function.
10696 (pass_fixup_cfg): Replace declaration with that of...
10697 (make_pass_fixup_cfg): ...new function.
10698 (pass_expand): Replace declaration with that of...
10699 (make_pass_expand): ...new function.
10700 (pass_instantiate_virtual_regs): Replace declaration with that of...
10701 (make_pass_instantiate_virtual_regs): ...new function.
10702 (pass_rtl_fwprop): Replace declaration with that of...
10703 (make_pass_rtl_fwprop): ...new function.
10704 (pass_rtl_fwprop_addr): Replace declaration with that of...
10705 (make_pass_rtl_fwprop_addr): ...new function.
10706 (pass_jump): Replace declaration with that of...
10707 (make_pass_jump): ...new function.
10708 (pass_jump2): Replace declaration with that of...
10709 (make_pass_jump2): ...new function.
10710 (pass_lower_subreg): Replace declaration with that of...
10711 (make_pass_lower_subreg): ...new function.
10712 (pass_cse): Replace declaration with that of...
10713 (make_pass_cse): ...new function.
10714 (pass_fast_rtl_dce): Replace declaration with that of...
10715 (make_pass_fast_rtl_dce): ...new function.
10716 (pass_ud_rtl_dce): Replace declaration with that of...
10717 (make_pass_ud_rtl_dce): ...new function.
10718 (pass_rtl_dce): Replace declaration with that of...
10719 (make_pass_rtl_dce): ...new function.
10720 (pass_rtl_dse1): Replace declaration with that of...
10721 (make_pass_rtl_dse1): ...new function.
10722 (pass_rtl_dse2): Replace declaration with that of...
10723 (make_pass_rtl_dse2): ...new function.
10724 (pass_rtl_dse3): Replace declaration with that of...
10725 (make_pass_rtl_dse3): ...new function.
10726 (pass_rtl_cprop): Replace declaration with that of...
10727 (make_pass_rtl_cprop): ...new function.
10728 (pass_rtl_pre): Replace declaration with that of...
10729 (make_pass_rtl_pre): ...new function.
10730 (pass_rtl_hoist): Replace declaration with that of...
10731 (make_pass_rtl_hoist): ...new function.
10732 (pass_rtl_store_motion): Replace declaration with that of...
10733 (make_pass_rtl_store_motion): ...new function.
10734 (pass_cse_after_global_opts): Replace declaration with that of...
10735 (make_pass_cse_after_global_opts): ...new function.
10736 (pass_rtl_ifcvt): Replace declaration with that of...
10737 (make_pass_rtl_ifcvt): ...new function.
10738 (pass_into_cfg_layout_mode): Replace declaration with that of...
10739 (make_pass_into_cfg_layout_mode): ...new function.
10740 (pass_outof_cfg_layout_mode): Replace declaration with that of...
10741 (make_pass_outof_cfg_layout_mode): ...new function.
10742 (pass_loop2): Replace declaration with that of...
10743 (make_pass_loop2): ...new function.
10744 (pass_rtl_loop_init): Replace declaration with that of...
10745 (make_pass_rtl_loop_init): ...new function.
10746 (pass_rtl_move_loop_invariants): Replace declaration with that of...
10747 (make_pass_rtl_move_loop_invariants): ...new function.
10748 (pass_rtl_unswitch): Replace declaration with that of...
10749 (make_pass_rtl_unswitch): ...new function.
10750 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
10751 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10752 (pass_rtl_doloop): Replace declaration with that of...
10753 (make_pass_rtl_doloop): ...new function.
10754 (pass_rtl_loop_done): Replace declaration with that of...
10755 (make_pass_rtl_loop_done): ...new function.
10756 (pass_web): Replace declaration with that of...
10757 (make_pass_web): ...new function.
10758 (pass_cse2): Replace declaration with that of...
10759 (make_pass_cse2): ...new function.
10760 (pass_df_initialize_opt): Replace declaration with that of...
10761 (make_pass_df_initialize_opt): ...new function.
10762 (pass_df_initialize_no_opt): Replace declaration with that of...
10763 (make_pass_df_initialize_no_opt): ...new function.
10764 (pass_reginfo_init): Replace declaration with that of...
10765 (make_pass_reginfo_init): ...new function.
10766 (pass_inc_dec): Replace declaration with that of...
10767 (make_pass_inc_dec): ...new function.
10768 (pass_stack_ptr_mod): Replace declaration with that of...
10769 (make_pass_stack_ptr_mod): ...new function.
10770 (pass_initialize_regs): Replace declaration with that of...
10771 (make_pass_initialize_regs): ...new function.
10772 (pass_combine): Replace declaration with that of...
10773 (make_pass_combine): ...new function.
10774 (pass_if_after_combine): Replace declaration with that of...
10775 (make_pass_if_after_combine): ...new function.
10776 (pass_ree): Replace declaration with that of...
10777 (make_pass_ree): ...new function.
10778 (pass_partition_blocks): Replace declaration with that of...
10779 (make_pass_partition_blocks): ...new function.
10780 (pass_match_asm_constraints): Replace declaration with that of...
10781 (make_pass_match_asm_constraints): ...new function.
10782 (pass_regmove): Replace declaration with that of...
10783 (make_pass_regmove): ...new function.
10784 (pass_split_all_insns): Replace declaration with that of...
10785 (make_pass_split_all_insns): ...new function.
10786 (pass_fast_rtl_byte_dce): Replace declaration with that of...
10787 (make_pass_fast_rtl_byte_dce): ...new function.
10788 (pass_lower_subreg2): Replace declaration with that of...
10789 (make_pass_lower_subreg2): ...new function.
10790 (pass_mode_switching): Replace declaration with that of...
10791 (make_pass_mode_switching): ...new function.
10792 (pass_sms): Replace declaration with that of...
10793 (make_pass_sms): ...new function.
10794 (pass_sched): Replace declaration with that of...
10795 (make_pass_sched): ...new function.
10796 (pass_ira): Replace declaration with that of...
10797 (make_pass_ira): ...new function.
10798 (pass_reload): Replace declaration with that of...
10799 (make_pass_reload): ...new function.
10800 (pass_clean_state): Replace declaration with that of...
10801 (make_pass_clean_state): ...new function.
10802 (pass_branch_prob): Replace declaration with that of...
10803 (make_pass_branch_prob): ...new function.
10804 (pass_value_profile_transformations): Replace declaration with that
10806 (make_pass_value_profile_transformations): ...new function.
10807 (pass_postreload_cse): Replace declaration with that of...
10808 (make_pass_postreload_cse): ...new function.
10809 (pass_gcse2): Replace declaration with that of...
10810 (make_pass_gcse2): ...new function.
10811 (pass_split_after_reload): Replace declaration with that of...
10812 (make_pass_split_after_reload): ...new function.
10813 (pass_branch_target_load_optimize1): Replace declaration with that
10815 (make_pass_branch_target_load_optimize1): ...new function.
10816 (pass_thread_prologue_and_epilogue): Replace declaration with that
10818 (make_pass_thread_prologue_and_epilogue): ...new function.
10819 (pass_stack_adjustments): Replace declaration with that of...
10820 (make_pass_stack_adjustments): ...new function.
10821 (pass_peephole2): Replace declaration with that of...
10822 (make_pass_peephole2): ...new function.
10823 (pass_if_after_reload): Replace declaration with that of...
10824 (make_pass_if_after_reload): ...new function.
10825 (pass_regrename): Replace declaration with that of...
10826 (make_pass_regrename): ...new function.
10827 (pass_cprop_hardreg): Replace declaration with that of...
10828 (make_pass_cprop_hardreg): ...new function.
10829 (pass_reorder_blocks): Replace declaration with that of...
10830 (make_pass_reorder_blocks): ...new function.
10831 (pass_branch_target_load_optimize2): Replace declaration with that
10833 (make_pass_branch_target_load_optimize2): ...new function.
10834 (pass_leaf_regs): Replace declaration with that of...
10835 (make_pass_leaf_regs): ...new function.
10836 (pass_split_before_sched2): Replace declaration with that of...
10837 (make_pass_split_before_sched2): ...new function.
10838 (pass_compare_elim_after_reload): Replace declaration with that of...
10839 (make_pass_compare_elim_after_reload): ...new function.
10840 (pass_sched2): Replace declaration with that of...
10841 (make_pass_sched2): ...new function.
10842 (pass_stack_regs): Replace declaration with that of...
10843 (make_pass_stack_regs): ...new function.
10844 (pass_stack_regs_run): Replace declaration with that of...
10845 (make_pass_stack_regs_run): ...new function.
10846 (pass_df_finish): Replace declaration with that of...
10847 (make_pass_df_finish): ...new function.
10848 (pass_compute_alignments): Replace declaration with that of...
10849 (make_pass_compute_alignments): ...new function.
10850 (pass_duplicate_computed_gotos): Replace declaration with that of...
10851 (make_pass_duplicate_computed_gotos): ...new function.
10852 (pass_variable_tracking): Replace declaration with that of...
10853 (make_pass_variable_tracking): ...new function.
10854 (pass_free_cfg): Replace declaration with that of...
10855 (make_pass_free_cfg): ...new function.
10856 (pass_machine_reorg): Replace declaration with that of...
10857 (make_pass_machine_reorg): ...new function.
10858 (pass_cleanup_barriers): Replace declaration with that of...
10859 (make_pass_cleanup_barriers): ...new function.
10860 (pass_delay_slots): Replace declaration with that of...
10861 (make_pass_delay_slots): ...new function.
10862 (pass_split_for_shorten_branches): Replace declaration with that of...
10863 (make_pass_split_for_shorten_branches): ...new function.
10864 (pass_split_before_regstack): Replace declaration with that of...
10865 (make_pass_split_before_regstack): ...new function.
10866 (pass_convert_to_eh_region_ranges): Replace declaration with that
10868 (make_pass_convert_to_eh_region_ranges): ...new function.
10869 (pass_shorten_branches): Replace declaration with that of...
10870 (make_pass_shorten_branches): ...new function.
10871 (pass_set_nothrow_function_flags): Replace declaration with that of...
10872 (make_pass_set_nothrow_function_flags): ...new function.
10873 (pass_dwarf2_frame): Replace declaration with that of...
10874 (make_pass_dwarf2_frame): ...new function.
10875 (pass_final): Replace declaration with that of...
10876 (make_pass_final): ...new function.
10877 (pass_rtl_seqabstr): Replace declaration with that of...
10878 (make_pass_rtl_seqabstr): ...new function.
10879 (pass_release_ssa_names): Replace declaration with that of...
10880 (make_pass_release_ssa_names): ...new function.
10881 (pass_early_inline): Replace declaration with that of...
10882 (make_pass_early_inline): ...new function.
10883 (pass_inline_parameters): Replace declaration with that of...
10884 (make_pass_inline_parameters): ...new function.
10885 (pass_update_address_taken): Replace declaration with that of...
10886 (make_pass_update_address_taken): ...new function.
10887 (pass_convert_switch): Replace declaration with that of...
10888 (make_pass_convert_switch): ...new function.
10889 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
10890 to a subclass of simple_ipa_opt_pass along with...
10891 (pass_data_ipa_tree_profile): ...new pass_data instance and...
10892 (make_pass_ipa_tree_profile): ...new function.
10893 * tree-sra.c (pass_sra_early): Convert from a global struct to a
10894 subclass of gimple_opt_pass along with...
10895 (pass_data_sra_early): ...new pass_data instance and...
10896 (make_pass_sra_early): ...new function.
10897 (pass_sra): Convert from a global struct to a subclass of
10898 gimple_opt_pass along with...
10899 (pass_data_sra): ...new pass_data instance and...
10900 (make_pass_sra): ...new function.
10901 (pass_early_ipa_sra): Convert from a global struct to a subclass of
10902 gimple_opt_pass along with...
10903 (pass_data_early_ipa_sra): ...new pass_data instance and...
10904 (make_pass_early_ipa_sra): ...new function.
10905 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
10906 subclass of gimple_opt_pass along with...
10907 (pass_data_ccp): ...new pass_data instance and...
10908 (make_pass_ccp): ...new function.
10909 (pass_fold_builtins): Convert from a global struct to a subclass of
10910 gimple_opt_pass along with...
10911 (pass_data_fold_builtins): ...new pass_data instance and...
10912 (make_pass_fold_builtins): ...new function.
10913 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
10914 subclass of gimple_opt_pass along with...
10915 (pass_data_copy_prop): ...new pass_data instance and...
10916 (make_pass_copy_prop): ...new function.
10917 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
10918 global struct to a subclass of gimple_opt_pass along with...
10919 (pass_data_rename_ssa_copies): ...new pass_data instance and...
10920 (make_pass_rename_ssa_copies): ...new function.
10921 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
10922 subclass of gimple_opt_pass along with...
10923 (pass_data_dce): ...new pass_data instance and...
10924 (make_pass_dce): ...new function.
10925 (pass_dce_loop): Convert from a global struct to a subclass of
10926 gimple_opt_pass along with...
10927 (pass_data_dce_loop): ...new pass_data instance and...
10928 (make_pass_dce_loop): ...new function.
10929 (pass_cd_dce): Convert from a global struct to a subclass of
10930 gimple_opt_pass along with...
10931 (pass_data_cd_dce): ...new pass_data instance and...
10932 (make_pass_cd_dce): ...new function.
10933 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
10934 subclass of gimple_opt_pass along with...
10935 (pass_data_dominator): ...new pass_data instance and...
10936 (make_pass_dominator): ...new function.
10937 (pass_phi_only_cprop): Convert from a global struct to a subclass of
10938 gimple_opt_pass along with...
10939 (pass_data_phi_only_cprop): ...new pass_data instance and...
10940 (make_pass_phi_only_cprop): ...new function.
10941 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
10942 subclass of gimple_opt_pass along with...
10943 (pass_data_dse): ...new pass_data instance and...
10944 (make_pass_dse): ...new function.
10945 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
10946 a subclass of gimple_opt_pass along with...
10947 (pass_data_forwprop): ...new pass_data instance and...
10948 (make_pass_forwprop): ...new function.
10949 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
10950 struct to a subclass of gimple_opt_pass along with...
10951 (pass_data_tree_ifcombine): ...new pass_data instance and...
10952 (make_pass_tree_ifcombine): ...new function.
10953 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
10954 subclass of gimple_opt_pass along with...
10955 (pass_data_ch): ...new pass_data instance and...
10956 (make_pass_ch): ...new function.
10957 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
10958 subclass of gimple_opt_pass along with...
10959 (pass_data_tree_loop): ...new pass_data instance and...
10960 (make_pass_tree_loop): ...new function.
10961 (pass_tree_loop_init): Convert from a global struct to a subclass of
10962 gimple_opt_pass along with...
10963 (pass_data_tree_loop_init): ...new pass_data instance and...
10964 (make_pass_tree_loop_init): ...new function.
10965 (pass_lim): Convert from a global struct to a subclass of
10966 gimple_opt_pass along with...
10967 (pass_data_lim): ...new pass_data instance and...
10968 (make_pass_lim): ...new function.
10969 (pass_tree_unswitch): Convert from a global struct to a subclass of
10970 gimple_opt_pass along with...
10971 (pass_data_tree_unswitch): ...new pass_data instance and...
10972 (make_pass_tree_unswitch): ...new function.
10973 (pass_predcom): Convert from a global struct to a subclass of
10974 gimple_opt_pass along with...
10975 (pass_data_predcom): ...new pass_data instance and...
10976 (make_pass_predcom): ...new function.
10977 (pass_vectorize): Convert from a global struct to a subclass of
10978 gimple_opt_pass along with...
10979 (pass_data_vectorize): ...new pass_data instance and...
10980 (make_pass_vectorize): ...new function.
10981 (pass_graphite): Convert from a global struct to a subclass of
10982 gimple_opt_pass along with...
10983 (pass_data_graphite): ...new pass_data instance and...
10984 (make_pass_graphite): ...new function.
10985 (pass_graphite_transforms): Convert from a global struct to a subclass
10986 of gimple_opt_pass along with...
10987 (pass_data_graphite_transforms): ...new pass_data instance and...
10988 (make_pass_graphite_transforms): ...new function.
10989 (pass_check_data_deps): Convert from a global struct to a subclass of
10990 gimple_opt_pass along with...
10991 (pass_data_check_data_deps): ...new pass_data instance and...
10992 (make_pass_check_data_deps): ...new function.
10993 (pass_iv_canon): Convert from a global struct to a subclass of
10994 gimple_opt_pass along with...
10995 (pass_data_iv_canon): ...new pass_data instance and...
10996 (make_pass_iv_canon): ...new function.
10997 (pass_scev_cprop): Convert from a global struct to a subclass of
10998 gimple_opt_pass along with...
10999 (pass_data_scev_cprop): ...new pass_data instance and...
11000 (make_pass_scev_cprop): ...new function.
11001 (pass_record_bounds): Convert from a global struct to a subclass of
11002 gimple_opt_pass along with...
11003 (pass_data_record_bounds): ...new pass_data instance and...
11004 (make_pass_record_bounds): ...new function.
11005 (pass_complete_unroll): Convert from a global struct to a subclass of
11006 gimple_opt_pass along with...
11007 (pass_data_complete_unroll): ...new pass_data instance and...
11008 (make_pass_complete_unroll): ...new function.
11009 (pass_complete_unrolli): Convert from a global struct to a subclass of
11010 gimple_opt_pass along with...
11011 (pass_data_complete_unrolli): ...new pass_data instance and...
11012 (make_pass_complete_unrolli): ...new function.
11013 (pass_parallelize_loops): Convert from a global struct to a subclass
11014 of gimple_opt_pass along with...
11015 (pass_data_parallelize_loops): ...new pass_data instance and...
11016 (make_pass_parallelize_loops): ...new function.
11017 (pass_loop_prefetch): Convert from a global struct to a subclass of
11018 gimple_opt_pass along with...
11019 (pass_data_loop_prefetch): ...new pass_data instance and...
11020 (make_pass_loop_prefetch): ...new function.
11021 (pass_iv_optimize): Convert from a global struct to a subclass of
11022 gimple_opt_pass along with...
11023 (pass_data_iv_optimize): ...new pass_data instance and...
11024 (make_pass_iv_optimize): ...new function.
11025 (pass_tree_loop_done): Convert from a global struct to a subclass of
11026 gimple_opt_pass along with...
11027 (pass_data_tree_loop_done): ...new pass_data instance and...
11028 (make_pass_tree_loop_done): ...new function.
11029 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
11030 struct to a subclass of gimple_opt_pass along with...
11031 (pass_data_cse_reciprocals): ...new pass_data instance and...
11032 (make_pass_cse_reciprocals): ...new function.
11033 (pass_cse_sincos): Convert from a global struct to a subclass of
11034 gimple_opt_pass along with...
11035 (pass_data_cse_sincos): ...new pass_data instance and...
11036 (make_pass_cse_sincos): ...new function.
11037 (pass_optimize_bswap): Convert from a global struct to a subclass of
11038 gimple_opt_pass along with...
11039 (pass_data_optimize_bswap): ...new pass_data instance and...
11040 (make_pass_optimize_bswap): ...new function.
11041 (pass_optimize_widening_mul): Convert from a global struct to a
11042 subclass of gimple_opt_pass along with...
11043 (pass_data_optimize_widening_mul): ...new pass_data instance and...
11044 (make_pass_optimize_widening_mul): ...new function.
11045 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
11046 subclass of gimple_opt_pass along with...
11047 (pass_data_phiopt): ...new pass_data instance and...
11048 (make_pass_phiopt): ...new function.
11049 (pass_cselim): Convert from a global struct to a subclass of
11050 gimple_opt_pass along with...
11051 (pass_data_cselim): ...new pass_data instance and...
11052 (make_pass_cselim): ...new function.
11053 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
11054 subclass of gimple_opt_pass along with...
11055 (pass_data_phiprop): ...new pass_data instance and...
11056 (make_pass_phiprop): ...new function.
11057 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
11058 subclass of gimple_opt_pass along with...
11059 (pass_data_pre): ...new pass_data instance and...
11060 (make_pass_pre): ...new function.
11061 (pass_fre): Convert from a global struct to a subclass of
11062 gimple_opt_pass along with...
11063 (pass_data_fre): ...new pass_data instance and...
11064 (make_pass_fre): ...new function.
11065 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
11066 subclass of gimple_opt_pass along with...
11067 (pass_data_reassoc): ...new pass_data instance and...
11068 (make_pass_reassoc): ...new function.
11069 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
11070 subclass of gimple_opt_pass along with...
11071 (pass_data_sink_code): ...new pass_data instance and...
11072 (make_pass_sink_code): ...new function.
11073 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
11074 subclass of gimple_opt_pass along with...
11075 (pass_data_strlen): ...new pass_data instance and...
11076 (make_pass_strlen): ...new function.
11077 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
11078 struct to a subclass of gimple_opt_pass along with...
11079 (pass_data_build_alias): ...new pass_data instance and...
11080 (make_pass_build_alias): ...new function.
11081 (pass_build_ealias): Convert from a global struct to a subclass of
11082 gimple_opt_pass along with...
11083 (pass_data_build_ealias): ...new pass_data instance and...
11084 (make_pass_build_ealias): ...new function.
11085 (pass_ipa_pta): Convert from a global struct to a subclass of
11086 simple_ipa_opt_pass along with...
11087 (pass_data_ipa_pta): ...new pass_data instance and...
11088 (make_pass_ipa_pta): ...new function.
11089 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
11090 subclass of gimple_opt_pass along with...
11091 (pass_data_uncprop): ...new pass_data instance and...
11092 (make_pass_uncprop): ...new function.
11093 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
11094 global struct to a subclass of gimple_opt_pass along with...
11095 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
11096 (make_pass_late_warn_uninitialized): ...new function.
11097 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
11098 to a subclass of gimple_opt_pass along with...
11099 (pass_data_init_datastructures): ...new pass_data instance and...
11100 (make_pass_init_datastructures): ...new function.
11101 (pass_early_warn_uninitialized): Convert from a global struct to a
11102 subclass of gimple_opt_pass along with...
11103 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
11104 (make_pass_early_warn_uninitialized): ...new function.
11105 (pass_update_address_taken): Convert from a global struct to a
11106 subclass of gimple_opt_pass along with...
11107 (pass_data_update_address_taken): ...new pass_data instance and...
11108 (make_pass_update_address_taken): ...new function.
11109 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
11110 struct to a subclass of gimple_opt_pass along with...
11111 (pass_data_release_ssa_names): ...new pass_data instance and...
11112 (make_pass_release_ssa_names): ...new function.
11113 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
11114 subclass of gimple_opt_pass along with...
11115 (pass_data_stdarg): ...new pass_data instance and...
11116 (make_pass_stdarg): ...new function.
11117 * tree-switch-conversion.c (pass_convert_switch): Convert from a
11118 global struct to a subclass of gimple_opt_pass along with...
11119 (pass_data_convert_switch): ...new pass_data instance and...
11120 (make_pass_convert_switch): ...new function.
11121 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
11122 to a subclass of gimple_opt_pass along with...
11123 (pass_data_tail_recursion): ...new pass_data instance and...
11124 (make_pass_tail_recursion): ...new function.
11125 (pass_tail_calls): Convert from a global struct to a subclass of
11126 gimple_opt_pass along with...
11127 (pass_data_tail_calls): ...new pass_data instance and...
11128 (make_pass_tail_calls): ...new function.
11129 * tree-vect-generic.c (pass_lower_vector): Convert from a global
11130 struct to a subclass of gimple_opt_pass along with...
11131 (pass_data_lower_vector): ...new pass_data instance and...
11132 (make_pass_lower_vector): ...new function.
11133 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
11134 gimple_opt_pass along with...
11135 (pass_data_lower_vector_ssa): ...new pass_data instance and...
11136 (make_pass_lower_vector_ssa): ...new function.
11137 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
11138 to a subclass of gimple_opt_pass along with...
11139 (pass_data_slp_vectorize): ...new pass_data instance and...
11140 (make_pass_slp_vectorize): ...new function.
11141 (pass_ipa_increase_alignment): Convert from a global struct to a
11142 subclass of simple_ipa_opt_pass along with...
11143 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
11144 (make_pass_ipa_increase_alignment): ...new function.
11145 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
11146 gimple_opt_pass along with...
11147 (pass_data_vrp): ...new pass_data instance and...
11148 (make_pass_vrp): ...new function.
11149 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
11150 subclass of simple_ipa_opt_pass along with...
11151 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
11152 (make_pass_ipa_free_lang_data): ...new function.
11153 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
11154 gimple_opt_pass along with...
11155 (pass_data_tsan): ...new pass_data instance and...
11156 (make_pass_tsan): ...new function.
11157 (pass_tsan_O0): Convert from a global struct to a subclass of
11158 gimple_opt_pass along with...
11159 (pass_data_tsan_O0): ...new pass_data instance and...
11160 (make_pass_tsan_O0): ...new function.
11161 * var-tracking.c (pass_variable_tracking): Convert from a global
11162 struct to a subclass of rtl_opt_pass along with...
11163 (pass_data_variable_tracking): ...new pass_data instance and...
11164 (make_pass_variable_tracking): ...new function.
11165 * web.c (pass_web): Convert from a global struct to a subclass of
11166 rtl_opt_pass along with...
11167 (pass_data_web): ...new pass_data instance and...
11168 (make_pass_web): ...new function.
11169 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
11170 declaration with that of...
11171 (make_pass_mode_switch_use): ...new function.
11172 (pass_resolve_sw_modes): Replace declaration with that of...
11173 (make_pass_resolve_sw_modes): ...new function.
11174 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
11175 from a global struct to a subclass of rtl_opt_pass along with...
11176 (pass_data_mode_switch_use): ...new pass_data instance and...
11177 (make_pass_mode_switch_use): ...new function.
11178 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
11179 from a global struct to a subclass of rtl_opt_pass along with...
11180 (pass_data_resolve_sw_modes): ...new pass_data instance and...
11181 (make_pass_resolve_sw_modes): ...new function.
11182 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
11183 struct to a subclass of rtl_opt_pass along with...
11184 (pass_data_insert_vzeroupper): ...new pass_data instance and...
11185 (make_pass_insert_vzeroupper): ...new function.
11186 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
11187 global struct to a subclass of rtl_opt_pass along with...
11188 (pass_data_work_around_errata): ...new pass_data instance and...
11189 (make_pass_work_around_errata): ...new function.
11190 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
11191 struct to a subclass of rtl_opt_pass along with...
11192 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
11193 (make_pass_mips_machine_reorg2): ...new function.
11195 2013-08-05 David Malcolm <dmalcolm@redhat.com>
11197 * passes.c (pass_manager::operator new): New.
11199 2013-08-05 David Malcolm <dmalcolm@redhat.com>
11201 Handwritten part of conversion of passes to C++ classes.
11203 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
11204 (toplev.o): Add dep on PASS_MANAGER_H.
11205 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
11206 of early local pases to reflect this moving from a global to a
11207 member of gcc::pass_manager.
11208 (cgraph_add_new_function): Likewise.
11209 * lto-cgraph.c (lto_output_node): Update for conversion of
11210 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
11211 * passes.c (opt_pass::clone): New.
11212 (opt_pass::gate): New.
11213 (opt_pass::execute): New.
11214 (opt_pass::opt_pass): New.
11215 (pass_manager::execute_early_local_passes): New.
11216 (pass_manager::execute_pass_mode_switching): new.
11217 (finish_optimization_passes): Convert to...
11218 (pass_manager::finish_optimization_passes): ...this.
11219 (finish_optimization_passes): Update for conversion of passes to
11221 (register_dump_files_1): Use has_gate since we cannot portably
11222 check a vtable entry against NULL.
11223 (dump_one_pass): Likewise.
11224 (ipa_write_summaries_2): Likewise.
11225 (ipa_write_optimization_summaries_1): Likewise.
11226 (ipa_read_summaries_1): Likewise.
11227 (ipa_read_optimization_summaries_1): Likewise.
11228 (execute_ipa_stmt_fixups): Likewise.
11229 (pass_manager::pass_manager): Rewrite pass-creation, invoking
11230 pass-creation functions rather than wiring up globals, and
11231 storing the results in fields of pass_manager generated using
11232 pass-instances.def.
11233 (pass_manager::dump_profile_report): Update for conversion of
11234 passes to C++ classes.
11235 (pass_manager::execute_ipa_summary_passes): Likewise.
11236 (execute_one_ipa_transform_pass): Likewise.
11237 (execute_one_pass): Use has_gate and has_execute since we cannot
11238 portably check a vtable entry against NULL.
11239 * pass_manager.h (pass_manager::finish_optimization_passes): New.
11240 (pass_manager): Use pass-instances.def to add fields for the
11241 various pass instances.
11242 * toplev.c (finalize): Update for move of
11243 finish_optimization_passes to a method of gcc::pass_manager.
11244 * toplev.h (finish_optimization_passes): Move to method of class
11246 * tree-pass.h (struct pass_data): New.
11247 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
11248 (opt_pass::gate): Convert to virtual function.
11249 (opt_pass::~opt_pass): New.
11250 (opt_pass::clone): New.
11251 (opt_pass::execute): Convert to virtual function.
11252 (opt_pass::opt_pass): New.
11253 (opt_pass::ctxt_): new.
11254 (gimple_opt_pass): Convert to subclass of opt_pass.
11255 (gimple_opt_pass::gimple_opt_pass): New.
11256 (rtl_opt_pass): Convert to subclass of opt_pass.
11257 (rtl_opt_pass::rtl_opt_pass): New.
11258 (ipa_opt_pass_d): Convert to subclass of opt_pass.
11259 (ipa_opt_pass_d::ipa_opt_pass_d): New.
11260 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
11261 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
11262 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
11263 invocation of pass_mode_switching to reflect this moving from a
11264 global to a member of gcc::pass_manager.
11265 (ix86_option_override): Rework how pass_insert_vzeroupper is
11266 added to the pass_manager to reflect autogenerated changes.
11267 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
11269 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
11271 PR rtl-optimization/57708
11272 * recog.c (peep2_find_free_register): Validate all regs in a
11275 2013-08-05 Jan Hubicka <jh@suse.cz>
11278 * cgraph.c (verify_cgraph_node): Accept local flags from other
11280 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
11281 (function_and_variable_visibility): Likewise.
11282 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
11284 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11286 * graph.c (init_graph_slim_pretty_print): Remove.
11287 (print_graph_cfg): Do not call it. Use local pretty printer.
11288 (start_graph_dump): Likewise.
11290 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11292 * gimple-pretty-print.c (buffer): Remove.
11293 (initialized): Likewise.
11294 (maybe_init_pretty_print): Likewise.
11295 (print_gimple_stmt): Do not call it. Use non-static local
11296 pretty_printer variable.
11297 (print_gimple_expr): Likewise.
11298 (print_gimple_seq): Likewise.
11299 (gimple_dump_bb): Likewise.
11301 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11303 * asan.c (asan_pp): Remove.
11304 (asan_pp_initialized): Likewise.
11305 (asan_pp_initialize): Likewise.
11306 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
11307 (asan_emit_stack_protection): Tidy. Use local pretty printer.
11308 (asan_add_global): Likewise.
11310 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
11312 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
11313 * diagnostic.h (diagnostic_flush_buffer): Adjust.
11314 * pretty-print.c (pp_formatted_text_data): Likewise.
11315 (pp_indent): Rename from pp_base_indent.
11316 (pp_format): Rename from pp_base_format.
11317 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
11318 (pp_format_verbatim): Rename from pp_base_format_verbatim.
11319 (pp_flush): Rename from pp_base_flush.
11320 (pp_set_line_maximum_length): Rename from
11321 pp_base_set_line_maximum_length.
11322 (pp_clear_output_area): Rename from pp_base_clear_output_area.
11323 (pp_set_prefix): Rename from pp_base_set_prefix.
11324 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
11325 (pp_emit_prefix): Rename from pp_base_emit_prefix.
11326 (pp_append_text): Rename from pp_base_append_text.
11327 (pp_formatted_text): Rename from pp_base_formatted_text.
11328 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
11329 (pp_remaining_character_count_for_line): Rename from
11330 pp_base_remaining_character_count_for_line.
11331 (pp_newline): Rename from pp_base_newline.
11332 (pp_character): Rename from pp_base_character.
11333 (pp_string): Rename from pp_base_string.
11334 (pp_maybe_space): Rename from pp_base_maybe_space.
11335 * asan.c (asan_pp_string): Adjust.
11336 (asan_emit_stack_protection): Likewise.
11337 (asan_add_global): Likewise.
11338 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
11339 * tree-mudflap.c (mf_varname_tree): Likewise.
11340 * tree-pretty-print.c (pp_tree_identifier): Rename from
11341 pp_base_tree_identifier.
11342 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
11343 Declare as function.
11345 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11347 * pretty-print.h (pp_bar_bar): New.
11348 (pp_ampersand_ampersand): Likewise.
11349 (pp_less_equal): Likewise.
11350 (pp_greater_equal): Likewise.
11351 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
11352 printer functions instead of pp_string or operators and punctuators.
11353 (dump_gimple_call): Likewise.
11354 (dump_gimple_omp_for): Likewise.
11355 (dump_gimple_transaction): Likewise.
11356 (dump_gimple_phi): Likewise.
11357 (pp_gimple_stmt_1): Likewise.
11358 * sched-vis.c (print_insn): Likewise.
11359 * tree-mudflap.c (mf_varname_tree): Likewise.
11360 * tree-pretty-print.c (dump_block_node): Likewise.
11361 (dump_generic_node): Likewise.
11363 2013-08-02 Jan Hubicka <jh@suse.cz>
11365 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
11367 * lto-streamer-out.c (tree_is_indexable): Results decls and
11368 parm decls are not indexable.
11369 (DFS_write_tree_body): Do not follow args and results.
11370 (hash_tree): Likewise.
11371 (output_functions): Rearrange so struct function is needed
11372 only when real body is output; be able to also ouptut abstract
11373 functions; output DECL_ARGUMENTS and DECL_RESULT.
11374 (lto_output): When not in WPA, ale store abstract functions.
11375 (write_symbol): Do not care about RESULT_DECL.
11376 (output_symbol_p): Handle correctly sbtract decls.
11377 * lto-streamer-in.c (input_function): Rearrange so struct
11378 function can be NULL at entry; allow streaming of
11379 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
11380 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
11381 sanity check during LTO.
11382 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
11383 RESULT_DECl and DECL_ARGUMENTS.
11384 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
11387 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11389 * pretty-print.h (pp_underscore): New.
11391 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
11392 printer functions instead of pp_character.
11393 (dump_binary_rhs): Likewise.
11394 (dump_ternary_rhs): Likewise.
11395 (dump_gimple_call_args): Likewise.
11396 (pp_points_to_solution): Likewise.
11397 (dump_gimple_call): Likewise.
11398 (dump_gimple_switch): Likewise.
11399 (dump_gimple_cond): Likewise.
11400 (dump_gimple_bind): Likewise.
11401 (dump_gimple_try): Likewise.
11402 (dump_gimple_omp_for): Likewise.
11403 (dump_gimple_omp_continue): Likewise.
11404 (dump_gimple_omp_single): Likewise.
11405 (dump_gimple_omp_sections): Likewise.
11406 (dump_gimple_omp_block): Likewise.
11407 (dump_gimple_omp_critical): Likewise.
11408 (dump_gimple_transaction): Likewise.
11409 (dump_gimple_asm): Likewise.
11410 (dump_gimple_phi): Likewise.
11411 (dump_gimple_omp_parallel): Likewise.
11412 (dump_gimple_omp_task): Likewise.
11413 (dump_gimple_omp_atomic_load): Likewise.
11414 (dump_gimple_omp_atomic_store): Likewise.
11415 (dump_gimple_mem_ops): Likewise.
11416 (pp_gimple_stmt_1): Likewise.
11417 (pp_cfg_jump): Likewise.
11418 (dump_implicit_edges): Likewise.
11419 (gimple_dump_bb_for_graph): Likewise.
11420 * graph.c (draw_cfg_node): Likewise.
11421 * langhooks.c (lhd_print_error_function): Likewise.
11422 * sched-vis.c (print_exp): Likewise.
11423 (print_value): Likewise.
11424 (print_pattern): Likewise.
11425 (print_insn): Likewise.
11426 (rtl_dump_bb_for_graph): Likewise.
11427 * tree-pretty-print.c (dump_function_declaration): Likewise.
11428 (dump_array_domain): Likewise.
11429 (dump_omp_clause): Likewise.
11430 (dump_location): Likewise.
11431 (dump_generic_node): Likewise.
11432 (print_struct_decl): Likewise.
11433 * diagnostic.c (diagnostic_show_locus): Use pp_space.
11435 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11437 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
11438 candidate table when replacing a candidate statement.
11439 (replace_rhs_if_not_dup): Likewise.
11440 (replace_one_candidate): Likewise.
11442 2013-08-02 Jan Hubicka <jh@suse.cz>
11443 Martin Liska <marxin.liska@gmail.com>
11445 * cgraphunit.c (add_new_function): Fix logic when adding from
11447 (assemble_thunk): Rename to ...
11448 (expand_thunk); .. this one; export; get it working with
11449 general functions; make produced gimple valid.
11450 * cgraph.h (expand_thunk): Declare.
11452 2013-08-02 Jan Hubicka <jh@suse.cz>
11454 * ipa-cp.c (gather_context_independent_values): Use
11455 ipa_get_param_move_cost.
11456 (get_replacement_map): Remove PARAM; move parameter folding
11458 (create_specialized_node): Update.
11459 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
11460 assert that we have gimple body; update move_cost.
11461 (count_formal_params): Assert that we have gimple body.
11462 (ipa_dump_param): New function.
11463 (ipa_alloc_node_params): Break out from ...
11464 (ipa_initialize_node_params): ... here.
11465 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
11466 (ipa_write_node_info): Stream move costs.
11467 (ipa_read_node_info): Read move costs.
11468 (ipa_update_after_lto_read): Do not recompute node params.
11469 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
11470 (ipa_get_param): Check we are not in WPA.
11471 (ipa_get_param_move_cost): New.
11472 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
11473 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
11474 parm numbers to be present.
11476 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11478 PR rtl-optimization/58048
11479 * lra-constraints.c (process_alt_operands): Don't check asm
11480 operand on register.
11482 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
11484 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
11485 the implied StoreLoad barrier for atomic operations if before.
11487 2013-08-02 Jan Hubicka <jh@suse.cz>
11488 Martin Liska <marxin.liska@gmail.com>
11490 * cgraph.c (cgraph_function_body_availability): Do not check
11492 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
11493 symtab_node_availability): Declare.
11494 * ipa.c (can_replace_by_local_alias): New.
11495 (function_and_variable_visibility): Use it.
11496 * symtab.c (symtab_for_node_and_aliases,
11497 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
11499 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11501 PR rtl-optimization/57963
11502 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
11503 (lra_constraints): Use them.
11505 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
11507 * config/arm/types.md (define_attr "type"): Add "load_acq"
11509 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
11511 (cortex_a53_store1): Likewise.
11513 2013-08-01 Jan Hubicka <jh@suse.cz>
11515 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
11516 partitions are not needed.
11518 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
11520 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
11521 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
11522 MAYBE_NON_Q_CLASS_P where appropriate.
11524 2013-08-01 Jan Hubicka <jh@suse.cz>
11526 * cgraph.h (release_function_body): Declare.
11527 * tree.c (free_lang_data_in_decl): Free, parameters and return values
11528 of unused delcarations.
11530 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11532 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
11533 RTL form when subtracting a constant.
11535 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11537 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
11538 Generate canonical plus rtx with negated immediate instead of minus
11540 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
11542 2013-08-01 Jan Hubicka <jh@suse.cz>
11544 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
11545 (cgraph_release_function_body): Likewise.
11546 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
11547 * cgraph.h (cgrpah_node): Rename abstract_and_needed
11548 to used_as_abstract_origin.
11549 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
11550 symbols used as abstract origins.
11551 * cgraphunit.c (analyze_functions): Update.
11552 * ipa.c (symtab_remove_unreachable_nodes): Recompute
11553 used_as_abstract_origin.
11554 * tree-inline.c (tree_function_versioning): Update
11555 used_as_abstract_origin; be ready for DECL_RESULT and
11556 DECL_ARGUMENTS to be NULL.
11558 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
11559 for abstract functions.
11560 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
11563 2013-08-01 Jan Hubicka <jh@suse.cz>
11565 * profile.c (compute_value_histograms): Fix thinko.
11567 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11569 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
11570 aarch-common-protos.h to extra_headers.
11571 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11572 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
11573 * config/aarch64/t-aarch64 (aarch-common.o): Define.
11575 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11577 * config/aarch64/aarch64.md (define_attr "type"): Delete.
11578 Include "../arm/types.md". Define "type" attribute for all patterns.
11579 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
11582 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
11584 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
11585 to support power8 load fusion.
11586 (fusion_gpr_mem_load): Likewise.
11588 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
11590 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
11591 declarations for power8 load fusion.
11592 (emit_fusion_gpr_load): Likewise.
11594 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
11595 tuning for power8, turn on fusion mode by default. Turn on sign
11596 extending fusion mode if normal fusion mode is on, and we are at
11598 (fusion_gpr_load_p): New function, return true if we can fuse an
11599 addis instruction with a dependent load to a GPR.
11600 (emit_fusion_gpr_load): Emit the instructions for power8 load
11603 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
11604 (VSX load fusion peepholes): New peepholes to fuse together an
11605 addi instruction with a VSX load instruction.
11607 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
11608 peepholes to fuse an addis instruction with a load to a GPR base
11609 register. If we are supporting sign extending fusions, convert
11610 sign extending loads to zero extending loads and add an explicit
11613 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11615 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
11616 aarch-common-protos.h to extra_headers.
11617 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11618 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
11619 (arm_early_store_addr_dep): Likewise.
11620 (arm_no_early_alu_shift_dep): Likewise.
11621 (arm_no_early_alu_shift_value_dep): Likewise.
11622 (arm_no_early_mul_dep): Likewise.
11623 (arm_no_early_store_addr_dep): Likewise.
11624 (arm_mac_accumulator_is_mul_result): Likewise.
11625 (arm_mac_accumulator_is_result): Likewise.
11626 * config/arm/aarch-common.c: ... here. New file.
11627 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
11629 (arm_early_store_addr_dep): Likewise.
11630 (arm_no_early_alu_shift_dep): Likewise.
11631 (arm_no_early_alu_shift_value_dep): Likewise.
11632 (arm_no_early_mul_dep): Likewise.
11633 (arm_no_early_store_addr_dep): Likewise.
11634 (arm_mac_accumulator_is_mul_result): Likewise.
11635 (arm_mac_accumulator_is_result): Likewise.
11636 * config/arm/aarch-common-protos.h: ... here. New file.
11637 * config/arm/t-arm (aarch-common.o): Define.
11639 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11641 * config/arm/arm.md: Include new file "types.md".
11642 (define_attr "type"): Move from here to ...
11643 (define_attr "mul32"): Likewise.
11644 (define_attr "mul64"): Likewise.
11645 * config/arm/types.md: ... here. New file.
11647 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
11649 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
11650 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
11652 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11654 * gen-pass-instances.awk: Fix offset of substr().
11656 2013-07-31 David Malcolm <dmalcolm@redhat.com>
11658 * Makefile.in (pass-instances.def): New.
11659 (passes.o): Replace dependency on passes.def with one on
11662 * gen-pass-instances.awk: New.
11664 * passes.c (pass_manager::pass_manager): Use pass-instances.def
11665 rather than passes.def, updating local definition of NEXT_PASS
11666 macro to add an extra NUM parameter (currently unused).
11668 2013-07-30 David Malcolm <dmalcolm@redhat.com>
11670 * Makefile.in (PASS_MANAGER_H): New.
11671 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
11672 (passes.o): Likewise.
11673 (statistics.o): Likewise.
11674 (cgraphunit.o): Likewise.
11675 (context.o): Depend on PASS_MANAGER_H.
11677 * pass_manager.h: New.
11679 * cgraphunit.c (cgraph_add_new_function): Update for moves
11680 of globals to fields of pass_manager.
11681 (analyze_function): Likewise.
11682 (expand_function): Likewise.
11683 (ipa_passes): Likewise.
11684 (compile): Likewise.
11686 * context.c (context::context): New.
11687 * context.h (context::context): New.
11688 (context::get_passes): New.
11689 (context::passes_): New.
11691 * lto-cgraph.c (input_node): Update for moves of globals to
11692 fields of pass_manager.
11694 * passes.c (all_passes): Remove, in favor of a field of the
11695 same name within the new class pass_manager.
11696 (all_small_ipa_passes): Likewise.
11697 (all_lowering_passes): Likewise.
11698 (all_regular_ipa_passes): Likewise.
11699 (all_late_ipa_passes): Likewise.
11700 (all_lto_gen_passes): Likewise.
11701 (passes_by_id): Likewise.
11702 (passes_by_id_size): Likewise.
11703 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11704 the new class pass_manager.
11705 (set_pass_for_id): Convert to...
11706 (pass_manager::set_pass_for_id): ...method.
11707 (get_pass_for_id): Convert to...
11708 (pass_manager::get_pass_for_id): ...method.
11709 (register_one_dump_file): Move body of implementation into...
11710 (pass_manager::register_one_dump_file): ...here.
11711 (register_dump_files_1): Convert to...
11712 (pass_manager::register_dump_files_1): ...method.
11713 (register_dump_files): Convert to...
11714 (pass_manager::register_dump_files): ...method.
11715 (create_pass_tab): Update for moves of globals to fields of
11717 (dump_passes): Move body of implementation into...
11718 (pass_manager::dump_passes): ...here.
11719 (register_pass): Move body of implementation into...
11720 (pass_manager::register_pass): ...here.
11721 (init_optimization_passes): Convert into...
11722 (pass_manager::pass_manager): ...constructor for new
11723 pass_manager class, and initialize the pass_lists array.
11724 (check_profile_consistency): Update for moves of globals to
11725 fields of pass_manager.
11726 (dump_profile_report): Move body of implementation into...
11727 (pass_manager::dump_profile_report): ...here.
11728 (ipa_write_summaries_1): Update for moves of pass lists from
11729 being globals to fields of pass_manager.
11730 (ipa_write_optimization_summaries): Likewise.
11731 (ipa_read_summaries): Likewise.
11732 (ipa_read_optimization_summaries): Likewise.
11733 (execute_all_ipa_stmt_fixups): Likewise.
11735 * statistics.c (statistics_fini): Update for moves of globals to
11736 fields of pass_manager.
11738 * toplev.c (general_init): Replace call to
11739 init_optimization_passes with construction of the pass_manager
11742 * tree-pass.h (all_passes): Remove, in favor of a field of the
11743 same name within the new class pass_manager.
11744 (all_small_ipa_passes): Likewise.
11745 (all_lowering_passes): Likewise.
11746 (all_regular_ipa_passes): Likewise.
11747 (all_lto_gen_passes): Likewise.
11748 (all_late_ipa_passes): Likewise.
11749 (passes_by_id): Likewise.
11750 (passes_by_id_size): Likewise.
11751 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11752 the new class pass_manager.
11753 (get_pass_for_id): Remove.
11755 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11757 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
11760 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11762 * arm.md (mulhi3): New expand pattern.
11764 2013-07-30 Jan Hubicka <jh@suse.cz>
11765 Martin Liska <marxin.liska@gmail.com>
11767 * profile.c (compute_value_histograms): Do not ICE when
11768 there is mismatch only on some counters.
11770 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11772 PR rtl-optimization/57637
11773 * function.c (move_insn_for_shrink_wrap): Also check the
11774 GEN set of the LIVE problem for the liveness analysis
11775 if it exists, otherwise give up.
11777 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11779 PR tree-optimization/57993
11780 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
11781 replaced statement in the candidate table.
11782 (phi_add_costs): Return infinite cost when the hidden basis does
11783 not dominate all phis on which the candidate is dependent.
11784 (replace_one_candidate): Record replaced statement in the
11787 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
11789 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
11790 (ashlv2si3): New expander.
11791 (*ashlv2si3_i): New define_insn_and_split.
11792 * predicates.md (float_operation): Allow patterns with three
11793 basic sub-patterns.
11795 PR rtl-optimization/58021
11796 * mode-switching.c (create_pre_exit): Always split off preceding
11797 insns if we are not at the basic block head.
11799 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
11801 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
11802 (UCLIBC_DYNAMIC_LINKER): New macro.
11803 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
11805 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
11806 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
11808 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
11809 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
11810 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
11811 for SF and DF modes. Use ieee_quad_format for TF mode.
11812 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
11813 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
11814 (mips_option_override): Handle `-mnan=legacy'.
11815 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
11816 `-mabs=2008' and `-mnan=2008'.
11817 (OPTION_DEFAULT_SPECS): Add "nan" default.
11818 (ASM_SPEC): Handle `-mnan='.
11819 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
11820 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
11821 comment accordingly.
11822 (neg<mode>2): Likewise.
11823 * config/mips/mips.opt (mabs, mnan): New options.
11824 * doc/install.texi (Configuration): Document `--with-nan=' option.
11825 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
11827 (MIPS Options): Document them.
11828 * config.gcc <mips*-*-*>: Handle `--with-nan='.
11829 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
11830 * configure: Regenerate.
11831 * config.in: Regenerate.
11833 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11835 * config/i386/i386.md (float post-reload splitters): Do not check
11836 for subregs of SSE registers.
11838 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11839 H.J. Lu <hongjiu.lu@intel.com>
11843 * config/i386/i386.md (post-reload splitter
11844 to avoid partial SSE reg dependency stalls): New pattern.
11846 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
11848 * config/s390/s390.md ("movcc"): Swap load and store instructions.
11850 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
11852 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11853 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
11855 2013-07-26 Cary Coutant <ccoutant@google.com>
11857 * dwarf2out.c (die_checksum_ordered): Don't include template
11858 instantiations in signature.
11859 (is_template_parameter): New function.
11860 (is_template_instantiation): New function.
11861 (generate_skeleton_bottom_up): Don't include template instantiations
11863 (generate_skeleton): Likewise.
11864 (break_out_comdat_types): Move recursive call to break out nested
11866 (prune_unused_types_mark_generic_parms_dies): Call
11867 is_template_parameter.
11869 2013-07-26 Ian Bolton <ian.bolton@arm.com>
11871 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
11872 uses vector registers.
11873 * config/aarch64/iterators.md: Add attributes rtn and vas.
11875 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11876 Richard Earnshaw <richard.earnshaw@arm.com>
11878 * combine.c (simplify_comparison): Re-canonicalize operands
11880 * config/arm/arm.md (movcond_addsi): New splitter.
11882 2013-07-25 Sterling Augustine <saugustine@google.com>
11884 * dwarf2out.c (size_of_pubnames): Move code to...
11885 (include_pubname_in_output): ...here. New.
11886 (want_pubnames): Rearrange.
11887 (output_pubnames): Call include_pubname_in_output. Move assertion.
11889 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11891 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
11893 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11896 * doc/extend.texi: Remove obsolete builtins. Fix
11897 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
11899 2013-07-25 Jan Hubicka <jh@suse.cz>
11901 * cgraph.c (release_function_body): Break out from ...
11902 (cgraph_release_function_body): ... this one; also release DECL_RESULT
11903 and DECL_ARGUMENTS.
11904 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
11905 old_tree in the map.
11906 (create_specialized_node): Update.
11907 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
11909 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
11910 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
11911 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
11912 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
11915 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11917 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
11918 addsi3_carryin_alt2_<optab>): Correct output template.
11920 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11922 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
11923 Adjust for arm_restrict_it.
11924 Remove trailing whitespace.
11926 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
11928 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
11929 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
11931 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
11933 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
11935 PR rtl-optimization/57960
11936 * lra-constraints.c (process_alt_operands): Use the right mode
11937 when checking strict_low.
11939 2013-07-25 Jan Hubicka <jh@suse.cz>
11941 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
11942 * cgraph.c (cgraph_remove_node): Do not release function body
11943 when in cgraph streaming.
11944 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
11945 in other partitions are not considered reachable; fix handling of
11948 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11950 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
11952 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11957 * config/arm/arm.md ("*sibcall_insn): Replace use of
11958 Ss with US. Adjust output for v5 and v4t.
11959 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
11961 * config/arm/constraints.md ("Ss"): Rename to US.
11963 2013-07-25 Terry Guo <terry.guo@arm.com>
11965 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
11966 shift_add/shift_sub0/shift_sub1 RTXs.
11968 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
11969 Anton Blanchard <anton@au1.ibm.com>
11971 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
11972 (altivec_vpks<VI_char>ss): Likewise.
11973 (altivec_vpks<VI_char>us): Likewise.
11974 (altivec_vpku<VI_char>us): Likewise.
11975 (altivec_vpku<VI_char>um): Likewise.
11977 2013-07-24 David Malcolm <dmalcolm@redhat.com>
11979 Introduce context class.
11981 * Makefile.in (CONTEXT_H): New.
11982 (OBJS): Add context.o.
11983 (toplev.o): Add CONTEXT_H to dependencies.
11986 * toplev.c (general_init): Create the singleton gcc::context instance.
11992 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
11994 PR rtl-optimization/57968
11995 * mode-switching.c (create_pre_exit): Allow instructions that
11996 don't set a return register to need a non-exit mode.
11998 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11999 Anton Blanchard <anton@au1.ibm.com>
12001 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
12002 operands to vperm for little endian.
12003 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
12004 of lvsl to create the control mask for a vperm for little endian.
12006 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12007 Anton Blanchard <anton@au1.ibm.com>
12009 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
12010 two operands for little-endian.
12012 2013-07-23 Steve Ellcey <sellcey@mips.com>
12014 * config/mips/mips.c (mips_case_values_threshold): New.
12015 (TARGET_CASE_VALUES_THRESHOLD): Define.
12017 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12018 Anton Blanchard <anton@au1.ibm.com>
12020 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
12021 selection of field for vector splat in little endian mode.
12023 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
12025 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
12026 expanders to rs6000.md.
12027 (ior<mode>3): Likewise.
12028 (and<mode>3): Likewise.
12029 (one_cmpl<mode>2): Likewise.
12030 (nor<mode>3): Likewise.
12031 (andc<mode>3): Likewise.
12032 (eqv<mode>3): Likewise.
12033 (nand<mode>3): Likewise.
12034 (orc<mode>3): Likewise.
12036 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
12039 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
12040 to split multi-word logical operations.
12041 (rs6000_split_logical_di): Likewise.
12042 (rs6000_split_logical): Likewise.
12044 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
12045 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
12046 and allow TImode operations in 32-bit.
12047 (vsx_and<mode>3_64bit): Likewise.
12048 (vsx_ior<mode>3_32bit): Likewise.
12049 (vsx_ior<mode>3_64bit): Likewise.
12050 (vsx_xor<mode>3_32bit): Likewise.
12051 (vsx_xor<mode>3_64bit): Likewise.
12052 (vsx_one_cmpl<mode>2_32bit): Likewise.
12053 (vsx_one_cmpl<mode>2_64bit): Likewise.
12054 (vsx_nor<mode>3_32bit): Likewise.
12055 (vsx_nor<mode>3_64bit): Likewise.
12056 (vsx_andc<mode>3_32bit): Likewise.
12057 (vsx_andc<mode>3_64bit): Likewise.
12058 (vsx_eqv<mode>3_32bit): Likewise.
12059 (vsx_eqv<mode>3_64bit): Likewise.
12060 (vsx_nand<mode>3_32bit): Likewise.
12061 (vsx_nand<mode>3_64bit): Likewise.
12062 (vsx_orc<mode>3_32bit): Likewise.
12063 (vsx_orc<mode>3_64bit): Likewise.
12065 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
12066 logical types in GPRs.
12068 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
12069 logical insns to rs6000.md, and allow TImode operations in
12071 (altivec_ior<mode>3): Likewise.
12072 (altivec_xor<mode>3): Likewise.
12073 (altivec_one_cmpl<mode>2): Likewise.
12074 (altivec_nor<mode>3): Likewise.
12075 (altivec_andc<mode>3): Likewise.
12077 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
12078 attributes for moving the 128-bit logical operations into
12080 (BOOL_REGS_OUTPUT): Likewise.
12081 (BOOL_REGS_OP1): Likewise.
12082 (BOOL_REGS_OP2): Likewise.
12083 (BOOL_REGS_UNARY): Likewise.
12084 (BOOL_REGS_AND_CR0): Likewise.
12085 (one_cmpl<mode>2): Add support for DI logical operations on
12086 32-bit, splitting the operations to 32-bit.
12087 (anddi3): Likewise.
12088 (iordi3): Likewise.
12089 (xordi3): Likewise.
12090 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
12091 changes to combine the 32/64-bit code, allow logical operations on
12092 TI mode in 32-bit, and to use similar match_operator patterns like
12093 scalar mode uses. Combine the Altivec and VSX code for logical
12094 operations, and move it here.
12095 (ior<mode>3, 128-bit types): Likewise.
12096 (xor<mode>3, 128-bit types): Likewise.
12097 (one_cmpl<mode>3, 128-bit types): Likewise.
12098 (nor<mode>3, 128-bit types): Likewise.
12099 (andc<mode>3, 128-bit types): Likewise.
12100 (eqv<mode>3, 128-bit types): Likewise.
12101 (nand<mode>3, 128-bit types): Likewise.
12102 (orc<mode>3, 128-bit types): Likewise.
12103 (and<mode>3_internal): Likewise.
12104 (bool<mode>3_internal): Likewise.
12105 (boolc<mode>3_internal1): Likewise.
12106 (boolc<mode>3_internal2): Likewise.
12107 (boolcc<mode>3_internal1): Likewise.
12108 (boolcc<mode>3_internal2): Likewise.
12109 (eqv<mode>3_internal1): Likewise.
12110 (eqv<mode>3_internal2): Likewise.
12111 (one_cmpl1<mode>3_internal): Likewise.
12113 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
12115 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12116 Rename flag_stack_usage to flag_stack_usage_info.
12118 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
12120 * config/microblaze/sync.md: New file.
12121 * config/microblaze/microblaze.md: Include sync.md
12122 * config/microblaze/microblaze.c: Add print_operand 'y'.
12123 * config/microblaze/constraints.md: Add memory_contraint
12124 'Q' which is a single register.
12126 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
12128 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
12130 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
12132 * reload.c (find_reloads): Exit loop once we find this operand
12133 cannot be reloaded somehow for this alternative.
12135 * reload.c (find_reloads): Exit loop once we find a hard register.
12137 * rtlanal.c (computed_jump_p): Exit loop once we find label
12140 * i386.c (ix86_pad_returns): Exit loop after setting replace.
12142 * cfgloopmanip.c (remove_path): Exit loop after setting
12145 * gensupport.c (subst_dup): Avoid loop if code is not
12146 MATCH_DUP nor MATCH_OP_DUP.
12148 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
12150 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
12152 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12154 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
12155 true for SP_REGNUM if mode == ptr_mode.
12156 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
12157 with value R0_REGNUM + 31.
12159 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12161 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
12162 pad pointer-typed argument downward.
12164 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12166 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
12167 and __ILP32__ when the ILP32 model is in use.
12169 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12171 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
12172 (aarch64_load_symref_appropriately): In the case of
12173 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
12174 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
12175 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
12176 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
12177 if 'mode' doesn't equal to 'ptr_mode'.
12178 (aarch64_output_mi_thunk): Add an assertion on the alignment of
12179 'vcall_offset'; change to call aarch64_emit_move differently depending
12180 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
12181 to calculate the upper bound of 'vcall_offset'.
12182 (aarch64_cannot_force_const_mem): Change to also return true if
12184 (aarch64_legitimize_reload_address): In the case of large
12185 displacements, add new local variable 'xmode' and an assertion
12186 based on it; change to use 'xmode' to generate the new rtx and
12188 (aarch64_asm_trampoline_template): Change to generate the template
12189 differently depending on TARGET_ILP32 or not; change to use
12190 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
12191 (aarch64_trampoline_size): Removed.
12192 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
12193 and replace immediate literals with it. Change to use 'ptr_mode'
12194 instead of 'DImode' and call convert_memory_address if the mode
12195 of 'fnaddr' doesn't equal to 'ptr_mode'.
12196 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
12198 (aarch64_elf_asm_destructor): Likewise.
12199 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
12200 on TARGET_ILP32 instead of aarch64_trampoline_size.
12201 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
12202 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
12203 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
12204 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
12205 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
12206 (storewb_pair<GPI:mode>_<P:mode>): ... this.
12207 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
12208 depending on the value of 'mode'.
12209 (add_losym_<mode>): New.
12210 (ldr_got_small_<mode>): New, based on ldr_got_small.
12211 (ldr_got_small): Remove.
12212 (ldr_got_small_sidi): New.
12213 * config/aarch64/iterators.md (P): New.
12214 (PTR): Change to 'ptr_mode' in the condition.
12216 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12218 * config.gcc (aarch64*-*-*): Support --with-abi.
12219 (aarch64*-*-elf): Support --with-multilib-list.
12220 (aarch64*-*-linux*): Likewise.
12221 (supported_defaults): Add abi to aarch64*-*-*.
12222 * configure.ac: Mention AArch64 for --with-multilib-list.
12223 * configure: Re-generated.
12224 * config/aarch64/biarchilp32.h: New file.
12225 * config/aarch64/biarchlp64.h: New file.
12226 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
12228 (MULTILIB_DEFAULTS): Ditto.
12229 (DRIVER_SELF_SPECS): Ditto.
12230 (ASM_SPEC): Update to also substitute -mabi.
12231 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
12232 file whose name depends on -mabi= and -mbig-endian.
12233 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
12235 (POINTER_SIZE): New define.
12236 (POINTERS_EXTEND_UNSIGNED): Ditto.
12237 (enum aarch64_abi_type): New enumeration tag.
12238 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
12239 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
12240 (TARGET_ILP32): New define.
12241 * config/aarch64/aarch64.opt (mabi): New.
12242 (aarch64_abi): New.
12243 (ilp32, lp64): New values for -mabi.
12244 * config/aarch64/t-aarch64 (comma): New define.
12245 (MULTILIB_OPTIONS): Ditto.
12246 (MULTILIB_DIRNAMES): Ditto.
12247 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
12248 * doc/invoke.texi: Document -mabi for AArch64.
12250 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
12252 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
12254 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12255 Anton Blanchard <anton@au1.ibm.com>
12257 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
12258 endianness when selecting field to splat.
12260 2013-07-22 Eric Christopher <echristo@gmail.com>
12262 * dwarf2out.c (die_odr_checksum): New function to use
12263 CHECKSUM_ macros and ULEB128 for DIE tag.
12264 (generate_type_signature): Use.
12266 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
12268 * config.gcc (sparc*-*-*): Accept leon3 processor.
12269 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
12270 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
12271 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
12272 * config/sparc/sparc.opt (enum processor_type): Add leon3.
12273 (mfix-ut699): Adjust comment.
12274 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
12275 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
12276 (CPP_CPU_SPEC): Likewise.
12277 (ASM_CPU_SPEC): Likewise.
12278 * config/sparc/sparc.c (leon3_cost): New constant.
12279 (sparc_option_override): Add leon3 support.
12280 (mem_ref): New function.
12281 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
12282 (sparc_do_work_around_errata): Look into the instruction in the delay
12283 slot and adjust accordingly. Add fix for the data cache nullify issues
12284 of the UT699. Change insertion position for the NOP.
12285 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
12286 (leon3_load): New reservation.
12287 (leon_store): Bump latency to 2.
12288 (grfpu): New automaton.
12289 (grfpu_alu): New unit.
12290 (grfpu_ds): Likewise.
12291 (leon_fp_alu): Adjust.
12292 (leon_fp_mult): Delete.
12293 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
12294 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
12295 * config/sparc/sparc.md (cpu): Add leon3.
12296 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
12297 (swapsi): Likewise.
12298 (atomic_test_and_set): Likewise.
12299 (ldstub): Likewise.
12301 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
12303 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
12304 default for R5900 targets.
12305 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
12306 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
12307 * config/mips/mips.c (mips_option_override): Report an error for
12308 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
12309 for -march=r5900 -mhard-float.
12311 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
12313 * df-problems.c (can_move_insns_across): Exit loop once we
12314 find a non-fixed, non-global register.
12316 * ipa-pure-const.c (propagate_nothrow): Exit loop after
12319 * omega.c (omega_eliminate_red): Break after setting red_found.
12320 (omega_problem_has_red_equations): Similarly after setting found.
12321 (omega_query_variable): Similarly after setting coupled.
12323 2013-07-22 Marek Polacek <polacek@redhat.com>
12325 * gimplify.c: Don't include gimple.h twice.
12327 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12329 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
12330 instead of TARGET_THUMB1.
12331 (Pz): New constraint.
12332 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
12334 (compare_negsi_si): Likewise.
12335 (compare_addsi2_op0): Likewise.
12336 (compare_addsi2_op1): Likewise.
12337 (addsi3_carryin_<optab>): Likewise.
12338 (addsi3_carryin_alt2_<optab>): Likewise.
12339 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
12340 for arm_restrict_it.
12341 (subsi3_carryin): Likewise.
12342 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
12343 (minmax_arithsi): Disable for arm_restrict_it.
12344 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
12345 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
12346 (satsi_<SAT:code>_shift): Likewise.
12347 (arm_shiftsi3): Add alternative for 16-bit encoding.
12348 (arm32_movhf): Disable for arm_restrict_it.
12349 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
12350 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
12352 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
12354 * config/arm/arm.md (attribute "insn"): Delete.
12355 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
12356 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
12357 (not_shiftsi): Update for attribute change.
12358 (not_shiftsi_compare0): Likewise.
12359 (not_shiftsi_compare0_scratch): Likewise.
12360 (arm_one_cmplsi2): Likewise.
12361 (thumb1_one_cmplsi2): Likewise.
12362 (notsi_compare0): Likewise.
12363 (notsi_compare0_scratch): Likewise.
12364 (thumb1_movdi_insn): Likewise.
12365 (arm_movsi_insn): Likewise.
12366 (movhi_insn_arch4): Likewise.
12367 (movhi_bytes): Likewise.
12368 (arm_movqi_insn): Likewise.
12369 (thumb1_movqi_insn): Likewise.
12370 (arm32_movhf): Likewise.
12371 (thumb1_movhf): Likewise.
12372 (arm_movsf_soft_insn): Likewise.
12373 (thumb1_movsf_insn): Likewise.
12374 (thumb_movdf_insn): Likewise.
12375 (movsicc_insn): Likewise.
12376 (movsfcc_soft_insn): Likewise.
12377 (and_scc): Likewise.
12378 (cond_move): Likewise.
12379 (if_move_not): Likewise.
12380 (if_not_move): Likewise.
12381 (if_shift_move): Likewise.
12382 (if_move_shift): Likewise.
12383 (if_shift_shift): Likewise.
12384 (if_not_arith): Likewise.
12385 (if_arith_not): Likewise.
12386 (cond_move_not): Likewise.
12387 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
12388 (neon_mov<mode>): Likewise.
12389 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
12390 (thumb2_movsi_vfp): Likewise.
12391 (movsf_vfp): Likewise.
12392 (thumb2_movsf_vfp): Likewise.
12393 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12395 (cortexa7_older_only): Likewise.
12396 (cortexa7_younger): Likewise.
12397 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
12398 (1020alu_shift_op): Likewise.
12399 (1020alu_shift_reg_op): Likewise.
12400 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
12401 (alu_shift_op): Likewise.
12402 (alu_shift_reg_op): Likewise.
12403 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
12404 (11_alu_shift_op): Likewise.
12405 (11_alu_shift_reg_op): Likewise.
12406 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
12407 (9_alu_shift_reg_op): Likewise.
12408 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12410 (cortex_a15_alu_shift): Likewise.
12411 (cortex_a15_alu_shift_reg): Likewise.
12412 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
12413 (cortex_a5_alu_shift): Likewise.
12414 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
12416 (cortex_a53_alu_shift): Likewise.
12417 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12419 (cortex_a7_alu_reg): Likewise.
12420 (cortex_a7_alu_shift): Likewise.
12421 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12422 (cortex_a8_alu_shift): Likewise.
12423 (cortex_a8_alu_shift_reg): Likewise.
12424 (cortex_a8_mov): Likewise.
12425 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
12426 (cortex_a9_dp_shift): Likewise.
12427 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
12428 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
12429 (cortex_r4_mov): Likewise.
12430 (cortex_r4_alu_shift): Likewise.
12431 (cortex_r4_alu_shift_reg): Likewise.
12432 * config/arm/fa526.md (526_alu_op): Update for attribute change.
12433 (526_alu_shift_op): Likewise.
12434 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
12435 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
12436 (626te_alu_shift_op): Likewise.
12437 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
12438 (726te_alu_op): Likewise.
12439 (726te_alu_shift_op): Likewise.
12440 (726te_alu_shift_reg_op): Likewise.
12441 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
12442 (mp626_alu_shift_op): Likewise.
12443 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
12444 (pj4_alu_e1_conds): Likewise.
12445 (pj4_alu): Likewise.
12446 (pj4_alu_conds): Likewise.
12447 (pj4_shift): Likewise.
12448 (pj4_shift_conds): Likewise.
12449 (pj4_alu_shift): Likewise.
12450 (pj4_alu_shift_conds): Likewise.
12452 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12454 * config/arm/predicates.md (shiftable_operator_strict_it):
12456 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
12457 Disable cond_exec version for arm_restrict_it.
12458 (thumb2_smaxsi3): Convert to generate cond_exec.
12459 (thumb2_sminsi3): Likewise.
12460 (thumb32_umaxsi3): Likewise.
12461 (thumb2_uminsi3): Likewise.
12462 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
12463 (thumb2_neg_abssi2): Likewise.
12464 (thumb2_mov_scc): Add alternative for 16-bit encoding.
12465 (thumb2_movsicc_insn): Adjust alternatives.
12466 (thumb2_mov_negscc): Disable for arm_restrict_it.
12467 (thumb2_mov_negscc_strict_it): New pattern.
12468 (thumb2_mov_notscc_strict_it): New pattern.
12469 (thumb2_mov_notscc): Disable for arm_restrict_it.
12470 (thumb2_ior_scc): Likewise.
12471 (thumb2_ior_scc_strict_it): New pattern.
12472 (thumb2_cond_move): Adjust for arm_restrict_it.
12473 (thumb2_cond_arith): Disable for arm_restrict_it.
12474 (thumb2_cond_arith_strict_it): New pattern.
12475 (thumb2_cond_sub): Adjust for arm_restrict_it.
12476 (thumb2_movcond): Likewise.
12477 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
12478 (thumb2_zero_extendhisi2_v6): Likewise.
12479 (thumb2_zero_extendqisi2_v6): Likewise.
12480 (orsi_notsi_si): Likewise.
12481 (orsi_not_shiftsi_si): Likewise.
12483 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
12485 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
12486 instruction sequence is 1 byte shorter.
12488 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
12490 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
12491 it is not needed after split.
12493 2013-07-20 Iain Sandoe <iain@codesourcery.com>
12496 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
12497 second label for nonlocal goto receivers. Don't output pic base labels
12498 unless we're producing PIC; mark that action unreachable().
12499 (ix86_save_reg): If the function contains a nonlocal label, save the
12501 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
12502 * config/darwin.c (emitted_pic_label_num): New GTY.
12503 (update_pic_label_number_if_needed): New.
12504 (machopic_output_function_base_name): Adjust for nonlocal receiver
12506 (machopic_should_output_picbase_label): New.
12507 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
12508 (nonlocal_goto_receiver): New insn and split.
12510 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
12512 * config/aarch64/aarch64-builtins.c
12513 (aarch64_fold_builtin): Fold abs in all modes.
12514 * config/aarch64/aarch64-simd-builtins.def
12515 (abs): Enable for all modes.
12516 * config/aarch64/arm_neon.h
12517 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
12518 (vabs_f64): Add missing intrinsic.
12520 2013-07-19 Ian Bolton <ian.bolton@arm.com>
12522 * config/aarch64/arm_neon.h (vabs_s64): New function
12524 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
12527 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
12528 * config/avr/avr.md (adjust_len): Add `round'.
12529 * config/avr/avr-protos.h (avr_out_round): New prototype.
12530 (avr_out_plus): Add `out_label' argument.
12531 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
12532 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
12533 Handle the case where `insn' is just a pattern.
12534 (avr_out_bitop): Handle the case where `insn' is just a pattern.
12535 (avr_out_round): New function.
12536 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
12538 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
12540 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12541 Add check for flag_stack_usage to handle -fstack-usage support
12543 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
12545 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
12546 interaction for new Power8 flags and VSX.
12548 2013-07-18 Sriraman Tallam <tmsriram@google.com>
12550 PR middle-end/57698
12551 * tree-inline.c (expand_call_inline): Emit errors during
12552 early_inlining only if optimization is not turned on.
12554 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12558 * passes.c (init_optimization_passes): Move the construction of
12559 the pass hierarchy into a new passes.def file.
12561 * Makefile.in (passes.o): Add dependency on passes.def.
12563 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12565 * passes.c (init_optimization_passes): Introduce macros for
12566 constructing the tree of passes (INSERT_PASSES_AFTER,
12567 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
12568 TERMINATE_PASS_LIST).
12570 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
12571 Wei Mi <wmi@google.com>
12573 PR rtl-optimization/57878
12574 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
12576 (reload_pseudo_compare_func): Check nregs first for reload
12579 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12581 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
12583 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
12585 * read-rtl.c (validate_const_int): Once an invalid character is
12586 seen, quit the loop.
12588 * gengtype.c (write_roots): Similarly once we find the "deletable"
12589 or "if_marked" option.
12591 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12593 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
12594 "xtab" and "sat". Move value "clz" from here to ...
12595 (attriubte "type"): ... here.
12596 (satsi_<SAT:code>): Delete "insn" attribute.
12597 (satsi_<SAT:code>_shift): Likewise.
12598 (arm_zero_extendqisi2addsi): Likewise.
12599 (arm_extendqisi2addsi): Likewise.
12600 (clzsi2): Update for attribute changes.
12601 (rbitsi2): Likewise.
12602 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
12604 (arm_usatsihi): Likewise.
12605 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12607 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12609 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
12610 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
12611 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
12612 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
12613 in alphabetical order.
12614 (attribute "core_cycles"): Update for attribute changes.
12615 (arm_addsi3): Likewise.
12616 (addsi3_compare0): Likewise.
12617 (addsi3_compare0_scratch): Likewise.
12618 (addsi3_compare_op1): Likewise.
12619 (addsi3_compare_op2): Likewise.
12620 (compare_addsi2_op0): Likewise.
12621 (compare_addsi2_op1): Likewise.
12622 (addsi3_carryin_shift_<optab>): Likewise.
12623 (subsi3_carryin_shift): Likewise.
12624 (rsbsi3_carryin_shift): Likewise.
12625 (arm_subsi3_insn): Likewise.
12626 (subsi3_compare0): Likewise.
12627 (subsi3_compare): Likewise.
12628 (arm_andsi3_insn): Likewise.
12629 (thumb1_andsi3_insn): Likewise.
12630 (andsi3_compare0): Likewise.
12631 (andsi3_compare0_scratch): Likewise.
12632 (zeroextractsi_compare0_scratch
12633 (andsi_not_shiftsi_si): Likewise.
12634 (iorsi3_insn): Likewise.
12635 (iorsi3_compare0): Likewise.
12636 (iorsi3_compare0_scratch): Likewise.
12637 (arm_xorsi3): Likewise.
12638 (thumb1_xorsi3_insn): Likewise.
12639 (xorsi3_compare0): Likewise.
12640 (xorsi3_compare0_scratch): Likewise.
12641 (satsi_<SAT:code>_shift): Likewise.
12643 (arm_shiftsi3): Likewise.
12644 (shiftsi3_compare0): Likewise.
12645 (not_shiftsi): Likewise.
12646 (not_shiftsi_compare0): Likewise.
12647 (not_shiftsi_compare0_scratch): Likewise.
12648 (arm_one_cmplsi2): Likewise.
12649 (thumb_one_complsi2): Likewise.
12650 (notsi_compare0): Likewise.
12651 (notsi_compare0_scratch): Likewise.
12652 (thumb1_zero_extendhisi2): Likewise.
12653 (arm_zero_extendhisi2): Likewise.
12654 (arm_zero_extendhisi2_v6): Likewise.
12655 (arm_zero_extendhisi2addsi): Likewise.
12656 (thumb1_zero_extendqisi2): Likewise.
12657 (thumb1_zero_extendqisi2_v6): Likewise.
12658 (arm_zero_extendqisi2): Likewise.
12659 (arm_zero_extendqisi2_v6): Likewise.
12660 (arm_zero_extendqisi2addsi): Likewise.
12661 (thumb1_extendhisi2): Likewise.
12662 (arm_extendhisi2): Likewise.
12663 (arm_extendhisi2_v6): Likewise.
12664 (arm_extendqisi): Likewise.
12665 (arm_extendqisi_v6): Likewise.
12666 (arm_extendqisi2addsi): Likewise.
12667 (thumb1_extendqisi2): Likewise.
12668 (thumb1_movdi_insn): Likewise.
12669 (arm_movsi_insn): Likewise.
12670 (movsi_compare0): Likewise.
12671 (movhi_insn_arch4): Likewise.
12672 (movhi_bytes): Likewise.
12673 (arm_movqi_insn): Likewise.
12674 (thumb1_movqi_insn): Likewise.
12675 (arm32_movhf): Likewise.
12676 (thumb1_movhf): Likewise.
12677 (arm_movsf_soft_insn): Likewise.
12678 (thumb1_movsf_insn): Likewise.
12679 (movdf_soft_insn): Likewise.
12680 (thumb_movdf_insn): Likewise.
12681 (arm_cmpsi_insn): Likewise.
12682 (cmpsi_shiftsi): Likewise.
12683 (cmpsi_shiftsi_swp): Likewise.
12684 (arm_cmpsi_negshiftsi_si): Likewise.
12685 (movsicc_insn): Likewise.
12686 (movsfcc_soft_insn): Likewise.
12687 (arith_shiftsi): Likewise.
12688 (arith_shiftsi_compare0
12689 (arith_shiftsi_compare0_scratch
12690 (sub_shiftsi): Likewise.
12691 (sub_shiftsi_compare0
12692 (sub_shiftsi_compare0_scratch
12693 (and_scc): Likewise.
12694 (cond_move): Likewise.
12695 (if_plus_move): Likewise.
12696 (if_move_plus): Likewise.
12697 (if_move_not): Likewise.
12698 (if_not_move): Likewise.
12699 (if_shift_move): Likewise.
12700 (if_move_shift): Likewise.
12701 (if_shift_shift): Likewise.
12702 (if_not_arith): Likewise.
12703 (if_arith_not): Likewise.
12704 (cond_move_not): Likewise.
12705 (thumb1_ashlsi3): Set type attribute.
12706 (thumb1_ashrsi3): Likewise.
12707 (thumb1_lshrsi3): Likewise.
12708 (thumb1_rotrsi3): Likewise.
12709 (shiftsi3_compare0_scratch): Likewise.
12710 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
12711 (neon_mov<mode>): Likewise.
12712 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
12714 (thumb2_movsi_insn): Likewise.
12715 (thumb2_cmpsi_neg_shiftsi): Likewise.
12716 (thumb2_extendqisi_v6): Likewise.
12717 (thumb2_zero_extendhisi2_v6): Likewise.
12718 (thumb2_zero_extendqisi2_v6): Likewise.
12719 (thumb2_shiftsi3_short): Likewise.
12720 (thumb2_addsi3_compare0_scratch): Likewise.
12721 (orsi_not_shiftsi_si): Likewise.
12722 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
12723 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
12725 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
12726 (1020alu_shift_op): Likewise.
12727 (1020alu_shift_reg_op): Likewise.
12728 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
12729 (alu_shift_op): Likewise.
12730 (alu_shift_reg_op): Likewise.
12731 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
12732 (11_alu_shift_op): Likewise.
12733 (11_alu_shift_reg_op): Likewise.
12734 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
12735 (9_alu_shift_reg_op): Likewise.
12736 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12738 (cortex_a15_alu_shift): Likewise.
12739 (cortex_a15_alu_shift_reg): Likewise.
12740 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
12742 (cortex_a5_alu_shift): Likewise.
12743 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
12745 (cortex_a53_alu_shift): Likewise.
12746 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12748 (cortex_a7_alu_reg): Likewise.
12749 (cortex_a7_alu_shift): Likewise.
12750 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
12752 (cortex_a8_alu_shift): Likewise.
12753 (cortex_a8_alu_shift_reg): Likewise.
12754 (cortex_a8_mov): Likewise.
12755 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
12756 (cortex_a9_dp_shift): Likewise.
12757 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
12759 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
12761 (cortex_r4_mov): Likewise.
12762 (cortex_r4_alu_shift): Likewise.
12763 (cortex_r4_alu_shift_reg): Likewise.
12764 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
12765 (526_alu_shift_op): Likewise.
12766 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
12767 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
12768 (626te_alu_shift_op): Likewise.
12769 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
12770 (726te_alu_op): Likewise.
12771 (726te_alu_shift_op): Likewise.
12772 (726te_alu_shift_reg_op): Likewise.
12773 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
12774 (mp626_alu_shift_op): Likewise.
12775 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
12776 (pj4_alu_e1_conds): Likewise.
12777 (pj4_alu): Likewise.
12778 (pj4_alu_conds): Likewise.
12779 (pj4_shift): Likewise.
12780 (pj4_shift_conds): Likewise.
12781 (pj4_alu_shift): Likewise.
12782 (pj4_alu_shift_conds): Likewise.
12783 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12785 (cortexa7_older_only): Likewise.
12786 (cortexa7_younger): Likewise.
12788 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12790 * ipa-pure-const.c (generate_summary): Rename to...
12791 (pure_const_generate_summary): ... this.
12793 2013-07-17 Iain Sandoe <iain@codesourcery.com>
12795 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
12797 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
12800 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
12803 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12805 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
12806 enabled without -march=zEC12.
12807 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
12810 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12812 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
12813 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
12814 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
12815 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
12816 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
12817 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
12818 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
12820 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
12821 and ISA_HAS_NMADD3_NMSUB3.
12822 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
12823 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
12824 (nmsub4<mode>, nmsub3<mode>): Likewise.
12825 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
12827 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12829 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
12830 TARGET_MIPS5400 checking.
12832 2013-07-16 Jakub Jelinek <jakub@redhat.com>
12833 Peter Bergner <bergner@vnet.ibm.com>
12835 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
12836 registers in the comment.
12837 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
12838 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
12839 rather than FIRST_PSEUDO_REGISTERS.
12841 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
12843 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
12844 enable extra ISA flags with TARGET_HTM.
12846 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12848 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
12851 2013-07-15 Cong Hou <congh@google.com>
12853 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
12854 in compare function for sorting.
12856 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
12858 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
12859 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
12860 * config/rs6000/rs6000.opt: Add -mhtm option.
12861 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
12862 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
12863 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12864 __HTM__ if the HTM instructions are available.
12865 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
12866 htm_spr_reg_operand): New define_predicates.
12867 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
12868 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
12870 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
12871 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
12872 HTM builtin functions.
12873 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
12874 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
12875 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
12876 (rs6000_builtin_mask_calculate): Likewise.
12877 (rs6000_option_override_internal): Likewise.
12878 (bdesc_htm): Add new HTM builtin support.
12879 (htm_spr_num): New function.
12880 (htm_spr_regno): Likewise.
12881 (rs6000_htm_spr_icode): Likewise.
12882 (htm_expand_builtin): Likewise.
12883 (htm_init_builtins): Likewise.
12884 (rs6000_expand_builtin): Add support for HTM builtin functions.
12885 (rs6000_init_builtins): Likewise.
12886 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
12888 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
12889 (TARGET_HTM, MASK_HTM): Define macros.
12890 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
12891 (FIXED_REGISTERS): Likewise.
12892 (CALL_USED_REGISTERS): Likewise.
12893 (CALL_REALLY_USED_REGISTERS): Likewise.
12894 (REG_ALLOC_ORDER): Likewise.
12895 (enum reg_class): Likewise.
12896 (REG_CLASS_NAMES): Likewise.
12897 (REG_CLASS_CONTENTS): Likewise.
12898 (REGISTER_NAMES): Likewise.
12899 (ADDITIONAL_REGISTER_NAMES): Likewise.
12900 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
12901 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
12902 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
12903 * config/rs6000/htm.md: New file.
12904 * config/rs6000/htmintrin.h: New file.
12905 * config/rs6000/htmxlintrin.h: New file.
12907 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
12909 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
12910 Define SYMBOL_TINY_GOT, update comment.
12911 * config/aarch64/aarch64.c
12912 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
12913 (aarch64_expand_mov_immediate): Likewise.
12914 (aarch64_print_operand): Likewise.
12915 (aarch64_classify_symbol): Likewise.
12916 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
12917 (ldr_got_tiny): Define.
12919 2013-07-13 Tobias Grosser <tobias@grosser.es>
12921 PR tree-optimization/54094
12922 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
12923 scheduling dimension for the parallelism check from the polyhedral
12924 information in the AST.
12925 * graphite-dependences.c (carries_deps): Do not assume the schedule is
12928 2013-07-13 Jason Merrill <jason@redhat.com>
12930 * print-tree.c (debug_vec_tree): Use debug_raw.
12931 (debug_raw (vec<tree, va_gc> &)): New.
12932 (debug_raw (vec<tree, va_gc> *)): New.
12933 * tree.h: Declare them.
12935 2013-07-13 Bin Cheng <bin.cheng@arm.com>
12937 * ifcvt.c (ifcvt_after_combine): New static variable.
12938 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
12940 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
12941 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
12942 rest_of_handle_if_after_reload): Pass new argument for if_convert.
12944 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
12946 * config/mips/mips.c (mips_expand_call): Remove empty statement.
12948 2013-07-12 Michael Matz <matz@suse.de>
12950 PR middle-end/55771
12951 * convert.c (convert_to_real): Reject non-float inner types.
12953 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
12955 * config/aarch64/aarch64-protos.h
12956 (aarch64_simd_immediate_valid_for_move): Remove.
12957 * config/aarch64/aarch64.c (simd_immediate_info): New member.
12958 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
12960 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
12962 2013-07-11 Steve Ellcey <sellcey@mips.com>
12964 * config/mips/mips.c (mips_conditional_register_usage): Do not
12965 use t[0-7] registers in MIPS16 mode when optimizing for size.
12967 2013-07-11 Sriraman Tallam <tmsriram@google.com>
12969 * config/i386/i386.c (dispatch_function_versions): Fix array
12970 indexing of function_version_info to match actual_versions.
12972 2013-07-11 Teresa Johnson <tejohnson@google.com>
12974 * vec.h (struct va_gc): Move release out-of-line.
12975 (va_gc::release): Call ggc_free on released vec.
12977 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12979 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
12980 Require GOT register as additional operand in UNSPEC.
12981 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
12982 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12983 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12984 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
12985 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
12986 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
12987 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
12989 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
12992 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
12993 name seen by assembler/linker rather if available.
12995 2013-07-11 Andreas Schwab <schwab@suse.de>
12997 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
12999 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
13001 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
13003 2013-07-10 Joseph Myers <joseph@codesourcery.com>
13005 * doc/tm.texi.in: Move hook documentation to ....
13006 * target.def: ... here.
13008 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
13009 text on @hook line.
13010 * doc/tm.texi: Regenerate.
13012 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
13015 * doc/invoke.texi: Document -Wconditionally-supported.
13017 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
13020 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
13023 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
13026 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
13027 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
13028 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
13029 Remove duplicate devices.
13030 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
13031 * config/avr/t-multilib: Regenerate.
13032 * config/avr/avr-tables.opt: Regenerate.
13033 * doc/avr-mmcu.texi: Regenerate.
13035 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
13038 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
13040 2013-07-10 Graham Stott <graham.stott@btinternet.com>
13042 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
13043 the cost of MULT when optimizing for size.
13045 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
13047 * config/cr16/cr16-protos.h: Don't include target.h.
13049 2013-07-09 Joseph Myers <joseph@codesourcery.com>
13051 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
13052 adjust register size for TDmode and TFmode for VSX registers.
13054 2013-07-08 Kai Tietz <ktietz@redhat.com>
13057 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
13058 hook_bool_const_tree_true.
13060 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13062 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
13063 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
13064 * config/s390/s390.md: Define FPR*_REGNUM constants.
13065 Fix FPR2_REGNUM constant (18 -> 17).
13066 ("*trunc<BFP:mode><DFP_ALL:mode>2")
13067 ("*trunc<DFP_ALL:mode><BFP:mode>2")
13068 ("trunc<BFP:mode><DFP_ALL:mode>2")
13069 ("trunc<DFP_ALL:mode><BFP:mode>2")
13070 ("*extend<BFP:mode><DFP_ALL:mode>2")
13071 ("*extend<DFP_ALL:mode><BFP:mode>2")
13072 ("extend<BFP:mode><DFP_ALL:mode>2")
13073 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
13076 2013-07-08 Graham Stott <graham.stott@btinternet.com>
13078 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
13080 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13082 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
13083 and cfun_fpr_bit_p to cfun_fpr_save_p.
13084 (s390_frame_area, s390_register_info, s390_frame_info)
13085 (s390_emit_prologue, s390_emit_epilogue)
13086 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
13088 * config/s390/s390.h: Define *_REGNUM macros for floating point
13091 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
13093 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
13095 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
13097 PR rtl-optimization/57786
13098 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
13099 and break out of the loop when it is set to false.
13101 2013-07-08 Jakub Jelinek <jakub@redhat.com>
13104 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
13105 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
13106 (const_int 63)) 0)).
13107 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
13108 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
13109 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
13111 PR rtl-optimization/57829
13112 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
13113 mask bits outside of mode are just sign-extension from mode to HWI.
13115 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
13117 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
13118 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
13119 adjust_address instead of change_address to keep info about alignment.
13120 (emit_strmov): Remove.
13121 (emit_memmov): New function.
13122 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
13123 (expand_movmem_epilogue): Likewise and return updated rtx for
13125 (expand_constant_movmem_prologue): Likewise and return updated rtx for
13126 destination and source.
13127 (decide_alignment): Refactor, handle vector_loop.
13128 (ix86_expand_movmem): Likewise.
13129 (ix86_expand_setmem): Likewise.
13130 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
13132 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
13134 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
13135 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
13137 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
13139 * config/i386/sse.md (sse_movlhps): Change alternative 3
13140 of operand 2 to "m".
13142 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
13145 * config/i386/sse.md (iptr): New mode attribute.
13146 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
13147 (<sse>_vm<plusminus_insn><mode>3): Ditto.
13148 (<sse>_vmmul<mode>3): Ditto.
13149 (<sse>_vmdiv<mode>3): Ditto.
13150 (sse_vmrcpv4sf2): Ditto.
13151 (<sse>_vmsqrt<mode>2): Ditto.
13152 (sse_vmrsqrtv4sf2): Ditto.
13153 (<sse>_vm<code><mode>3): Ditto.
13154 (avx_vmcmp<mode>3): Ditto.
13155 (<sse>_vmmaskcmp<mode>3): Ditto.
13156 (<sse>_comi): Ditto.
13157 (<sse>_ucomi): Ditto.
13158 (*xop_vmfrcz_<mode>): Ditto.
13159 (*fmai_fmadd_<mode>): Ditto.
13160 (*fmai_fmsub_<mode>): Ditto.
13161 (*fmai_fnmadd_<mode>): Ditto.
13162 (*fmai_fnmsub_<mode>): Ditto.
13163 (*fma4i_vmfmadd_<mode>): Ditto.
13164 (*fma4i_vmfmsub_<mode>): Ditto.
13165 (*fma4i_vmfnmadd_<mode>): Ditto.
13166 (*fma4i_vmfnmsub_<mode>): Ditto.
13167 (*xop_vmfrcz_<mode>): Ditto.
13168 (sse_cvtps2pi): Ditto.
13169 (sse_cvttps2pi): Ditto.
13170 (sse_cvtss2si): Ditto.
13171 (sse_cvtss2si_2): Ditto.
13172 (sse_cvtss2siq_2): Ditto.
13173 (sse_cvttss2si): Ditto.
13174 (sse_cvttss2siq): Ditto.
13175 (sse_cvtsd2si): Ditto.
13176 (sse_cvtsd2si_2): Ditto.
13177 (sse_cvtsd2siq_2): Ditto.
13178 (sse_cvttsd2si): Ditto.
13179 (sse_cvttsd2siq): Ditto.
13180 (sse_cvtsd2ss): Ditto.
13181 (sse_cvtss2sd): Ditto.
13182 (avx2_pbroadcast<mode>): Ditto.
13183 (avx2_pbroadcast<mode>_1): Ditto.
13184 (*avx_vperm_broadcast_v4sf): Ditto.
13186 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
13187 (sse_movlhps): Ditto.
13188 (sse_storehps): Ditto.
13189 (sse_loadhps): Ditto.
13190 (sse_storelps): Ditto.
13191 (sse_loadlps): Ditto.
13192 (*vec_concatv4sf): Ditto.
13193 (*vec_interleave_highv2df): Ditto.
13194 (*vec_interleave_lowv2df): Ditto.
13195 (*vec_extractv2df_1_sse): Ditto.
13196 (*vec_extractv2df_0_sse): Ditto.
13197 (sse2_storelpd): Ditto.
13198 (sse2_loadlpd): Ditto.
13199 (sse2_movsd): Ditto.
13200 (*vec_concatv4si): Ditto.
13201 (vec_concatv2di): Ditto.
13203 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
13204 for Intel asm dialect.
13205 (mmx_punpcklwd): Ditto.
13206 (mmx_punpckldq): Ditto.
13208 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
13209 for intel assembler dialect.
13211 2013-07-06 Jakub Jelinek <jakub@redhat.com>
13214 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
13215 for BUILT_IN_C{LZ,LRSB}*.
13216 * tree.h (CASE_INT_FN): Add FN##IMAX case.
13217 * tree-vrp.c (extract_range_basic): Handle
13218 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
13219 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
13220 fall thru to code calling set_value*.
13221 * builtins.c (expand_builtin): Remove *IMAX cases.
13222 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
13223 if width is bigger than 2*HWI.
13225 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
13227 PR rtl-optimization/55342
13228 * lra-int.h (lra_subreg_reload_pseudos): New.
13229 * lra.c: Add undoing optional reloads to the block diagram.
13230 (lra_subreg_reload_pseudos): New.
13231 (lra_optional_reload_pseudos): Change comments.
13232 (lra): Init and clear lra_subreg_reload_pseudos. Clear
13233 lra_optional_reload_pseudos after undo transformations.
13234 * lra-assigns.c (pseudo_prefix_title): New.
13235 (lra_setup_reg_renumber): Use it.
13236 (spill_for): Ditto. Check subreg reload pseudos too.
13237 (assign_by_spills): Consider subreg reload pseudos too.
13238 * lra-constraints.c (simplify_operand_subreg): Use
13239 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
13240 (curr_insn_transform): Recognize and do optional reloads.
13241 (undo_optional_reloads): New.
13242 (lra_undo_inheritance): Call undo_optional_reloads.
13244 2013-07-05 Thomas Quinot <quinot@adacore.com>
13246 * tree-complex.c (expand_complex_operations_1): Fix typo.
13248 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
13250 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
13251 (tune_params): New member 'const vec_costs'.
13252 * config/aarch64/aarch64.c (generic_vector_cost): New.
13253 (generic_tunings): New member 'generic_vector_cost'.
13254 (aarch64_builtin_vectorization_cost): New.
13255 (aarch64_add_stmt_cost): New.
13256 (TARGET_VECTORIZE_ADD_STMT_COST): New.
13257 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
13259 2013-07-03 Jakub Jelinek <jakub@redhat.com>
13262 * config/i386/predicates.md (vsib_address_operand): Disallow
13263 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
13265 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
13267 PR middle-end/55030
13268 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
13269 expand_builtin_setjmp_receiver.
13270 (expand_label): Adjust, call expand_builtin_setjmp_receiver
13271 with NULL for the label parameter.
13272 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
13273 the frame-pointer. Adjust comments.
13274 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
13275 only if LABEL is non-NULL.
13277 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
13279 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
13280 (ARM_ABI_AAPCS64): Ditto.
13282 (ARM_DEFAULT_ABI): Ditto.
13284 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
13286 * config/aarch64/aarch64-builtins.c
13287 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
13288 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
13290 * config/aarch64/aarch64-simd.md
13291 (aarch64_ld1<VALL:mode>): New.
13292 (aarch64_st1<VALL:mode>): Likewise.
13293 * config/aarch64/arm_neon.h
13294 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
13296 2013-07-02 Sriraman Tallam <tmsriram@google.com>
13298 * config/i386/i386.c (gate_insert_vzeroupper): Check if
13300 (ix86_option_override_internal):Turn on all -mavx target flags by
13301 default as they are dependent on AVX anyway.
13303 2013-07-02 Cary Coutant <ccoutant@google.com>
13305 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
13306 deterministic hash.
13307 (loc_checksum_ordered): Likewise.
13308 (hash_loc_operands): Remove inline keyword.
13310 2013-07-02 Jakub Jelinek <jakub@redhat.com>
13312 PR tree-optimization/57741
13313 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
13314 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
13315 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
13316 Allow REAL_CST step_exprs if flag_associative_math.
13317 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
13319 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13321 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
13323 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13325 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
13327 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13329 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
13331 (iorsi3_insn): Likewise.
13332 (arm_xorsi3): Likewise.
13334 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
13336 * arm.md (attribute "wtype"): Delete. Move attribute values from here
13338 (attribute "type"): ... here, and prefix with "wmmx_".
13339 (attribute "core_cycles"): Update for attribute changes.
13340 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
13341 (tbcstv4hi): Likewise.
13342 (tbcstv2si): Likewise.
13343 (iwmmxt_iordi3): Likewise.
13344 (iwmmxt_xordi3): Likewise.
13345 (iwmmxt_anddi3): Likewise.
13346 (iwmmxt_nanddi3): Likewise.
13347 (iwmmxt_arm_movdi): Likewise.
13348 (iwmmxt_movsi_insn): Likewise.
13349 (mov<mode>_internal): Likewise.
13350 (and<mode>3_iwmmxt): Likewise.
13351 (ior<mode>3_iwmmxt): Likewise.
13352 (xor<mode>3_iwmmxt): Likewise.
13353 (add<mode>3_iwmmxt): Likewise.
13354 (ssaddv8qi3): Likewise.
13355 (ssaddv4hi3): Likewise.
13356 (ssaddv2si3): Likewise.
13357 (usaddv8qi3): Likewise.
13358 (usaddv4hi3): Likewise.
13359 (usaddv2si3): Likewise.
13360 (sub<mode>3_iwmmxt): Likewise.
13361 (sssubv8qi3): Likewise.
13362 (sssubv4hi3): Likewise.
13363 (sssubv2si3): Likewise.
13364 (ussubv8qi3): Likewise.
13365 (ussubv4hi3): Likewise.
13366 (ussubv2si3): Likewise.
13367 (mulv4hi3_iwmmxt): Likewise.
13368 (smulv4hi3_highpart): Likewise.
13369 (umulv4hi3_highpart): Likewise.
13370 (iwmmxt_wmacs): Likewise.
13371 (iwmmxt_wmacsz): Likewise.
13372 (iwmmxt_wmacu): Likewise.
13373 (iwmmxt_wmacuz): Likewise.
13374 (iwmmxt_clrdi): Likewise.
13375 (iwmmxt_clrv8qi): Likewise.
13376 (iwmmxt_clr4hi): Likewise.
13377 (iwmmxt_clr2si): Likewise.
13378 (iwmmxt_uavgrndv8qi3): Likewise.
13379 (iwmmxt_uavgrndv4hi3): Likewise.
13380 (iwmmxt_uavgv8qi3): Likewise.
13381 (iwmmxt_uavgv4hi3): Likewise.
13382 (iwmmxt_tinsrb): Likewise.
13383 (iwmmxt_tinsrh): Likewise.
13384 (iwmmxt_tinsrw): Likewise.
13385 (iwmmxt_textrmub): Likewise.
13386 (iwmmxt_textrmsb): Likewise.
13387 (iwmmxt_textrmuh): Likewise.
13388 (iwmmxt_textrmsh): Likewise.
13389 (iwmmxt_textrmw): Likewise.
13390 (iwmxxt_wshufh): Likewise.
13391 (eqv8qi3): Likewise.
13392 (eqv4hi3): Likewise.
13393 (eqv2si3): Likewise.
13394 (gtuv8qi3): Likewise.
13395 (gtuv4hi3): Likewise.
13396 (gtuv2si3): Likewise.
13397 (gtv8qi3): Likewise.
13398 (gtv4hi3): Likewise.
13399 (gtv2si3): Likewise.
13400 (smax<mode>3_iwmmxt): Likewise.
13401 (umax<mode>3_iwmmxt): Likewise.
13402 (smin<mode>3_iwmmxt): Likewise.
13403 (umin<mode>3_iwmmxt): Likewise.
13404 (iwmmxt_wpackhss): Likewise.
13405 (iwmmxt_wpackwss): Likewise.
13406 (iwmmxt_wpackdss): Likewise.
13407 (iwmmxt_wpackhus): Likewise.
13408 (iwmmxt_wpackwus): Likewise.
13409 (iwmmxt_wpackdus): Likewise.
13410 (iwmmxt_wunpckihb): Likewise.
13411 (iwmmxt_wunpckihh): Likewise.
13412 (iwmmxt_wunpckihw): Likewise.
13413 (iwmmxt_wunpckilb): Likewise.
13414 (iwmmxt_wunpckilh): Likewise.
13415 (iwmmxt_wunpckilw): Likewise.
13416 (iwmmxt_wunpckehub): Likewise.
13417 (iwmmxt_wunpckehuh): Likewise.
13418 (iwmmxt_wunpckehuw): Likewise.
13419 (iwmmxt_wunpckehsb): Likewise.
13420 (iwmmxt_wunpckehsh): Likewise.
13421 (iwmmxt_wunpckehsw): Likewise.
13422 (iwmmxt_wunpckelub): Likewise.
13423 (iwmmxt_wunpckeluh): Likewise.
13424 (iwmmxt_wunpckeluw): Likewise.
13425 (iwmmxt_wunpckelsb): Likewise.
13426 (iwmmxt_wunpckelsh): Likewise.
13427 (iwmmxt_wunpckelsw): Likewise.
13428 (ror<mode>3): Likewise.
13429 (ashr<mode>3_iwmmxt): Likewise.
13430 (lshr<mode>3_iwmmxt): Likewise.
13431 (ashl<mode>3_iwmmxt): Likewise.
13432 (ror<mode>3_di): Likewise.
13433 (ashr<mode>3_di): Likewise.
13434 (lshr<mode>3_di): Likewise.
13435 (ashl<mode>3_di): Likewise.
13436 (iwmmxt_wmadds): Likewise.
13437 (iwmmxt_wmaddu): Likewise.
13438 (iwmmxt_tmia): Likewise.
13439 (iwmmxt_tmiaph): Likewise.
13440 (iwmmxt_tmiabb): Likewise.
13441 (iwmmxt_tmiatb): Likewise.
13442 (iwmmxt_tmiabt): Likewise.
13443 (iwmmxt_tmiatt): Likewise.
13444 (iwmmxt_tmovmskb): Likewise.
13445 (iwmmxt_tmovmskh): Likewise.
13446 (iwmmxt_tmovmskw): Likewise.
13447 (iwmmxt_waccb): Likewise.
13448 (iwmmxt_wacch): Likewise.
13449 (iwmmxt_waccw): Likewise.
13450 (iwmmxt_waligni): Likewise.
13451 (iwmmxt_walignr): Likewise.
13452 (iwmmxt_walignr0): Likewise.
13453 (iwmmxt_walignr1): Likewise.
13454 (iwmmxt_walignr2): Likewise.
13455 (iwmmxt_walignr3): Likewise.
13456 (iwmmxt_wsadb): Likewise.
13457 (iwmmxt_wsadh): Likewise.
13458 (iwmmxt_wsadbz): Likewise.
13459 (iwmmxt_wsadhz): Likewise.
13460 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
13461 (iwmmxt_wabsdiffb): Likewise.
13462 (iwmmxt_wabsdiffh): Likewise.
13463 (iwmmxt_wabsdiffw): Likewise.
13464 (iwmmxt_waddsubhx): Likewise
13465 (iwmmxt_wsubaddhx): Likewise.
13466 (addc<mode>3): Likewise.
13467 (iwmmxt_avg4): Likewise.
13468 (iwmmxt_avg4r): Likewise.
13469 (iwmmxt_wmaddsx): Likewise.
13470 (iwmmxt_wmaddux): Likewise.
13471 (iwmmxt_wmaddsn): Likewise.
13472 (iwmmxt_wmaddun): Likewise.
13473 (iwmmxt_wmulwsm): Likewise.
13474 (iwmmxt_wmulwum): Likewise.
13475 (iwmmxt_wmulsmr): Likewise.
13476 (iwmmxt_wmulumr): Likewise.
13477 (iwmmxt_wmulwsmr): Likewise.
13478 (iwmmxt_wmulwumr): Likewise.
13479 (iwmmxt_wmulwl): Likewise.
13480 (iwmmxt_wqmulm): Likewise.
13481 (iwmmxt_wqmulwm): Likewise.
13482 (iwmmxt_wqmulmr): Likewise.
13483 (iwmmxt_wqmulwmr): Likewise.
13484 (iwmmxt_waddbhusm): Likewise.
13485 (iwmmxt_waddbhusl): Likewise.
13486 (iwmmxt_wqmiabb): Likewise.
13487 (iwmmxt_wqmiabt): Likewise.
13488 (iwmmxt_wqmiatb): Likewise.
13489 (iwmmxt_wqmiatt): Likewise.
13490 (iwmmxt_wqmiabbn): Likewise.
13491 (iwmmxt_wqmiabtn): Likewise.
13492 (iwmmxt_wqmiatbn): Likewise.
13493 (iwmmxt_wqmiattn): Likewise.
13494 (iwmmxt_wmiabb): Likewise.
13495 (iwmmxt_wmiabt): Likewise.
13496 (iwmmxt_wmiatb): Likewise.
13497 (iwmmxt_wmiatt): Likewise.
13498 (iwmmxt_wmiabbn): Likewise.
13499 (iwmmxt_wmiabtn): Likewise.
13500 (iwmmxt_wmiatbn): Likewise.
13501 (iwmmxt_wmiattn): Likewise.
13502 (iwmmxt_wmiawbb): Likewise.
13503 (iwmmxt_wmiawbt): Likewise.
13504 (iwmmxt_wmiawtb): Likewise.
13505 (iwmmxt_wmiawtt): Likewise.
13506 (iwmmxt_wmiawbbn): Likewise.
13507 (iwmmxt_wmiawbtn): Likewise.
13508 (iwmmxt_wmiawtbn): Likewise.
13509 (iwmmxt_wmiawttn): Likewise.
13510 (iwmmxt_wmerge): Likewise.
13511 (iwmmxt_tandc<mode>3): Likewise.
13512 (iwmmxt_torc<mode>3): Likewise.
13513 (iwmmxt_torvsc<mode>3): Likewise.
13514 (iwmmxt_textrc<mode>3): Likewise.
13515 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
13516 (wmmxt_pack): Likewise.
13517 (wmmxt_mult_c1): Likewise.
13518 (wmmxt_mult_c2): Likewise.
13519 (wmmxt_alu_c1): Likewise.
13520 (wmmxt_alu_c2): Likewise.
13521 (wmmxt_alu_c3): Likewise.
13522 (wmmxt_transfer_c1): Likewise.
13523 (wmmxt_transfer_c2): Likewise.
13524 (wmmxt_transfer_c3): Likewise.
13525 (marvell_f_iwmmxt_wstr): Likewise.
13526 (marvell_f_iwmmxt_wldr): Likewise.
13528 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
13530 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
13532 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13535 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13536 * lra-constraints.c (need_for_split_p): Check call used hard regs
13537 living through calls.
13539 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
13540 call used regs for call insn.
13542 2013-06-28 Jakub Jelinek <jakub@redhat.com>
13545 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
13546 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
13548 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
13550 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
13552 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13554 * lra-constraints.c (need_for_split_p): Check call used hard regs
13555 living through calls.
13557 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
13560 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
13561 to tie with any other modes. Eliminate Altivec vector mode tests,
13562 since these are a subset of ALTIVEC or VSX vector modes. Simplify
13563 code, to return 0 if testing MODE2 for a condition, if we've
13564 already tested MODE1 for the same condition.
13566 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13568 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
13571 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13573 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
13574 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
13576 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13578 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
13580 (aarch64_symbolic_constant_p): Remove.
13581 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
13582 static. Fix line length and white space.
13583 (aarch64_symbolic_constant_p): Remove.
13584 * config/aarch64/predicates.md (aarch64_valid_symref):
13585 Use aarch64_classify_symbol_expression.
13587 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13589 * config/arm/constraints.md (Ts): New constraint.
13590 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
13592 (compare_scc): Use "Ts" constraint for operand 0.
13593 (ior_scc_scc): Likewise.
13594 (and_scc_scc): Likewise.
13595 (and_scc_scc_nodom): Likewise.
13596 (ior_scc_scc_cmp): Likewise for operand 7.
13597 (and_scc_scc_cmp): Likewise.
13598 * config/arm/thumb2.md (thumb2_movsi_insn):
13599 Add alternatives for 16-bit encodings.
13600 (thumb2_movhi_insn): Likewise.
13601 (thumb2_movsicc_insn): Likewise.
13602 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
13603 (thumb2_negscc): Use "Ts" constraint.
13604 Move mvn instruction outside cond_exec block.
13605 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
13606 for 16-bit encodings.
13608 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13610 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
13612 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
13613 (mulsi3subsi): Likewise.
13614 (mulsidi3adddi): Likewise.
13615 (mulsidi3_v6): Likewise.
13616 (umulsidi3_v6): Likewise.
13617 (umulsidi3adddi_v6): Likewise.
13618 (smulsi3_highpart_v6): Likewise.
13619 (umulsi3_highpart_v6): Likewise.
13620 (mulhisi3tb): Likewise.
13621 (mulhisi3bt): Likewise.
13622 (mulhisi3tt): Likewise.
13623 (maddhisi4): Likewise.
13624 (maddhisi4tb): Likewise.
13625 (maddhisi4tt): Likewise.
13626 (maddhidi4): Likewise.
13627 (maddhidi4tb): Likewise.
13628 (maddhidi4tt): Likewise.
13629 (zeroextractsi_compare0_scratch): Likewise.
13630 (insv_zero): Likewise.
13631 (insv_t2): Likewise.
13632 (anddi_notzesidi_di): Likewise.
13633 (anddi_notsesidi_di): Likewise.
13634 (andsi_notsi_si): Likewise.
13635 (iordi_zesidi_di): Likewise.
13636 (xordi_zesidi_di): Likewise.
13637 (andsi_iorsi3_notsi): Likewise.
13638 (smax_0): Likewise.
13639 (smax_m1): Likewise.
13640 (smin_0): Likewise.
13641 (not_shiftsi): Likewise.
13642 (unaligned_loadsi): Likewise.
13643 (unaligned_loadhis): Likewise.
13644 (unaligned_loadhiu): Likewise.
13645 (unaligned_storesi): Likewise.
13646 (unaligned_storehi): Likewise.
13647 (extv_reg): Likewise.
13648 (extzv_t2): Likewise.
13649 (divsi3): Likewise.
13650 (udivsi3): Likewise.
13651 (arm_zero_extendhisi2addsi): Likewise.
13652 (arm_zero_extendqisi2addsi): Likewise.
13653 (compareqi_eq0): Likewise.
13654 (arm_extendhisi2_v6): Likewise.
13655 (arm_extendqisi2addsi): Likewise.
13656 (arm_movt): Likewise.
13657 (thumb2_ldrd): Likewise.
13658 (thumb2_ldrd_base): Likewise.
13659 (thumb2_ldrd_base_neg): Likewise.
13660 (thumb2_strd): Likewise.
13661 (thumb2_strd_base): Likewise.
13662 (thumb2_strd_base_neg): Likewise.
13663 (arm_negsi2): Add alternative for 16-bit encoding.
13664 (arm_one_cmplsi2): Likewise.
13666 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13668 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
13669 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
13670 (movdfcc): Likewise.
13671 * config/arm/vfp.md (*thumb2_movsf_vfp):
13672 Disable predication for arm_restrict_it.
13673 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
13674 (*thumb2_movdfcc_vfp): Likewise.
13675 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
13676 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
13677 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
13678 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
13679 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
13680 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
13681 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
13682 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
13683 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
13684 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
13685 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
13686 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
13687 Disable predication for arm_restrict_it.
13689 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
13691 * config/i386/bmiintrin.h (_bextr_u32): New.
13692 (_bextr_u64): Ditto.
13694 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
13696 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
13697 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
13698 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
13699 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
13700 * config/mips/n32-elf.h: ...this new file.
13702 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
13705 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
13706 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
13708 2013-06-27 Catherine Moore <clm@codesourcery.com>
13710 * config/mips/mips-tables.opt: Regenerate.
13711 * config/mips/mips-cpus.def: Add m14ke and m14kec.
13712 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
13713 * doc/invoke.texi: Add -m14kc.
13715 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13718 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
13719 constraints of operand 1 and 2.
13722 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
13723 to match RTL canonicalization. Swap predicates and
13724 constraints of operand 1 and 2.
13726 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
13728 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
13729 Process OP_INOUT regs for splitting too.
13731 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13733 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
13734 decl before the loop, initialize to NULL.
13735 (vectorizable_load): Initialize ptr_incr to NULL.
13737 2013-06-27 Martin Jambor <mjambor@suse.cz>
13740 * ipa-ref.h (ipa_maybe_record_reference): Declare.
13741 * ipa-ref.c (ipa_maybe_record_reference): New function.
13742 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
13743 * ipa-cp.c (create_specialized_node): Record potential references from
13745 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
13747 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13749 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
13750 parameter 'mode' of type 'enum machine_mode mode'; change to pass
13751 'mode' to force_reg.
13752 (aarch64_add_offset): Update calls to aarch64_force_temporary.
13753 (aarch64_expand_mov_immediate): Likewise.
13755 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13757 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
13758 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
13760 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13762 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
13763 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13764 (struct machine_function): Add tbegin_p.
13765 (s390_canonicalize_comparison): Fold CC mode compares to
13766 conditional jump if possible.
13767 (s390_emit_jump): Return the emitted jump.
13768 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
13769 Handle CCRAWmode compares.
13770 (s390_option_override): Default to -mhtm if available.
13771 (s390_reg_clobbered_rtx): Handle floating point regs as well.
13772 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
13773 FPRs instead of df_regs_ever_live_p.
13774 (s390_optimize_nonescaping_tx): New function.
13775 (s390_init_frame_layout): Extend clobbered_regs array to cover
13777 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
13778 (s390_expand_tbegin): New function.
13779 (enum s390_builtin): New enum definition.
13780 (code_for_builtin): New array definition.
13781 (s390_init_builtins): New function.
13782 (s390_expand_builtin): New function.
13783 (TARGET_INIT_BUILTINS): Define.
13784 (TARGET_EXPAND_BUILTIN): Define.
13785 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
13786 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
13787 (s390_alc_comparison): Likewise.
13788 * config/s390/s390-modes.def: Add CCRAWmode.
13789 * config/s390/s390.h (processor_flags): Add PF_TX.
13790 (TARGET_CPU_HTM): Define macro.
13791 (TARGET_HTM): Define macro.
13792 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
13793 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
13794 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13795 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
13796 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
13797 (TBEGIN_MASK, TBEGINC_MASK): New constants.
13798 ("*cc_to_int"): Move up.
13799 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
13800 constants other than 0.
13801 ("*ccraw_to_int"): New insn and splitter definition.
13802 ("tbegin", "tbegin_nofloat", "tbegin_retry")
13803 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
13804 ("tx_assist"): New expander.
13805 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
13806 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
13807 * config/s390/s390.opt: Add -mhtm option.
13808 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
13809 * config/s390/htmxlintrin.h: New file.
13810 * config/s390/htmintrin.h: New file.
13811 * config/s390/s390intrin.h: New file.
13812 * doc/extend.texi: Document htm builtins.
13813 * config.gcc: Add the new header files to extra_headers.
13815 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
13817 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
13818 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
13820 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
13821 Pat Haugen <pthaugen@us.ibm.com>
13822 Peter Bergner <bergner@vnet.ibm.com>
13824 * config/rs6000/power8.md: New.
13825 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
13826 setting for power8 entry.
13827 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
13828 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
13829 test for Power4/Power5 only.
13830 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
13832 (force_new_group): Adjust comment.
13833 * config/rs6000/rs6000.md: Include power8.md.
13835 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
13837 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
13838 * config/arm/arm-protos.h (arm_max_conditional_execute): New
13840 (tune_params): Update comment.
13841 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
13842 (arm_max_conditional_execute): New function.
13843 (thumb2_final_prescan_insn): Use max_insn_skipped and
13844 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
13846 2013-06-25 Jakub Jelinek <jakub@redhat.com>
13848 PR tree-optimization/57705
13849 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
13850 SSA_NAME step, provided that it is not defined inside the loop.
13851 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
13852 (get_initial_def_for_induction): Handle SSA_NAME IV step.
13854 2013-06-25 Martin Jambor <mjambor@suse.cz>
13856 PR middle-end/57670
13857 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
13858 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
13860 (ipa_note_param_call): Initialize member_ptr flag.
13861 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
13862 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
13863 (ipa_write_indirect_edge_info): Stream member_ptr flag.
13864 (ipa_read_indirect_edge_info): Likewise.
13866 2013-06-25 Richard Biener <rguenther@suse.de>
13868 PR middle-end/56977
13869 * passes.c (init_optimization_passes): Move pass_fold_builtins
13870 and pass_dce earlier with -Og.
13872 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
13874 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
13875 <BIT_FIELD_REF>: Remove trailing TAB.
13876 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
13879 2013-06-24 Martin Jambor <mjambor@suse.cz>
13881 PR tree-optimization/57358
13882 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
13883 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
13884 (ipa_analyze_params_uses): Generate pessimistic info when true.
13886 2013-06-24 Martin Jambor <mjambor@suse.cz>
13888 PR tree-optimization/57539
13889 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
13890 global.inlined_to of the new node to it. All callers changed.
13891 * ipa-inline-transform.c (clone_inlined_nodes): New variable
13892 inlining_into, pass it to cgraph_clone_node.
13893 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
13894 ipa_free_edge_args_substructures.
13895 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
13896 rdesc linked list. Do not assert rdesc edges have inlined caller.
13897 Assert we have found an rdesc in the rdesc list.
13899 2013-06-24 Richard Biener <rguenther@suse.de>
13901 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
13902 (pointer_set_lookup): Declare.
13903 (class pointer_map): New template class implementing a
13904 generic pointer to T map.
13905 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
13906 pointer_map<T>::contains, pointer_map<T>::insert,
13907 pointer_map<T>::traverse): New functions.
13908 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
13909 (pointer_set_lookup): New function.
13910 (pointer_set_contains): Use pointer_set_lookup.
13911 (pointer_set_insert): Likewise.
13912 (insert_aux): Remove.
13913 (struct pointer_map_t): Embed a pointer_set_t.
13914 (pointer_map_create): Adjust.
13915 (pointer_map_destroy): Likewise.
13916 (pointer_map_contains): Likewise.
13917 (pointer_map_insert): Likewise.
13918 (pointer_map_traverse): Likewise.
13919 * tree-streamer.h (struct streamer_tree_cache_d): Use a
13920 pointer_map<unsigned> instead of a pointer_map_t.
13921 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
13922 (streamer_tree_cache_lookup): Likewise.
13923 (streamer_tree_cache_create): Likewise.
13924 (streamer_tree_cache_delete): Likewise.
13925 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
13926 pointer_map<unsigned> instead of a pointer_map_t.
13927 (lto_init_tree_ref_encoder): Adjust.
13928 (lto_destroy_tree_ref_encoder): Likewise.
13929 * lto-section-out.c (lto_output_decl_index): Likewise.
13930 (lto_record_function_out_decl_state): Likewise.
13931 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
13933 2013-06-24 Richard Biener <rguenther@suse.de>
13935 PR tree-optimization/57488
13936 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
13938 2013-06-24 Alan Modra <amodra@gmail.com>
13940 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
13941 (gen_easy_altivec_constant): Likewise.
13942 * config/rs6000/predicates.md (easy_vector_constant_add_self,
13943 easy_vector_constant_msb): Likewise.
13945 2013-06-23 Jakub Jelinek <jakub@redhat.com>
13948 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
13949 add missing return true.
13951 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
13954 * config/sh/predicates.md (general_extend_operand): Invoke
13955 general_movsrc_operand for memory operands.
13956 (general_movsrc_operand): Allow reg+reg addressing, do not use
13957 general_operand for memory operands.
13959 2013-06-23 Sriraman Tallam <tmsriram@google.com>
13961 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
13962 when current target options does not apply.
13963 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
13964 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
13965 * config/i386/bmiintrin.h: Pass appropriate target
13966 attributes to header.
13967 * config/i386/mmintrin.h: Ditto.
13968 * config/i386/nmmintrin.h: Ditto.
13969 * config/i386/avx2intrin.h: Ditto.
13970 * config/i386/fxsrintrin.h: Ditto.
13971 * config/i386/tbmintrin.h: Ditto.
13972 * config/i386/xsaveintrin.h: Ditto.
13973 * config/i386/f16cintrin.h: Ditto.
13974 * config/i386/xtestintrin.h: Ditto.
13975 * config/i386/xsaveoptintrin.h: Ditto.
13976 * config/i386/bmi2intrin.h: Ditto.
13977 * config/i386/lzcntintrin.h: Ditto.
13978 * config/i386/smmintrin.h: Ditto.
13979 * config/i386/wmmintrin.h: Ditto.
13980 * config/i386/x86intrin.h: Remove all header include guards.
13981 * config/i386/prfchwintrin.h: Ditto.
13982 * config/i386/pmmintrin.h: Ditto.
13983 * config/i386/tmmintrin.h: Ditto.
13984 * config/i386/xmmintrin.h: Ditto.
13985 * config/i386/popcntintrin.h: Ditto.
13986 * config/i386/rdseedintrin.h: Ditto.
13987 * config/i386/ammintrin.h: Ditto.
13988 * config/i386/emmintrin.h: Ditto.
13989 * config/i386/immintrin.h: Remove all header include guards.
13990 * config/i386/fma4intrin.h: Ditto.
13991 * config/i386/lwpintrin.h: Ditto.
13992 * config/i386/xopintrin.h: Ditto.
13993 * config/i386/ia32intrin.h: Ditto.
13994 * config/i386/avxintrin.h: Ditto.
13995 * config/i386/rtmintrin.h: Ditto.
13996 * config/i386/fmaintrin.h: Ditto.
13997 * config/i386/mm3dnow.h: Ditto.
13999 2013-06-22 Sriraman Tallam <tmsriram@google.com>
14001 * common/config/i386/i386-common.c: Handle LZCNT.
14003 2013-06-22 Andi Kleen <ak@linux.intel.com>
14005 * doc/extend.texi: Use __atomic_store_n instead of
14006 __atomic_store in HLE example.
14008 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
14010 * config/sh/sh.c: Remove <cstdlib> workaround.
14012 2013-06-21 Andi Kleen <ak@linux.intel.com>
14014 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
14016 2013-06-21 Andi Kleen <ak@linux.intel.com>
14018 * doc/extend.texi: Document that __atomic_clear and
14019 __atomic_test_and_set should only be used with bool.
14021 2013-06-20 Jan Hubicka <jh@suse.cz>
14023 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
14024 types_same_for_odr.
14025 * tree.c (decls_same_for_odr): New function.
14026 (same_for_edr): New function.
14027 (types_same_for_odr): New function.
14028 (get_binfo_at_offset): Use it.
14029 * tree.h (types_same_for_odr): Declare.
14031 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
14032 Jason Merrill <jason@redhat.com>
14034 * system.h: Include <cstdlib> as well as <stdlib.h>.
14036 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
14039 * config/i386/i386.c (construct_container): Report error if
14040 long double is used with disabled x87 float returns.
14042 2013-06-20 Jan Hubicka <jh@suse.cz>
14044 * lto-cgraph.c (input_symtab): Do not set cgraph state.
14046 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
14048 PR rtl-optimization/57425
14049 PR rtl-optimization/57569
14050 * alias.c (write_dependence_p): Remove parameters mem_mode and
14051 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
14052 Changed all callers.
14053 (canon_anti_dependence): Get comments and semantics in sync.
14054 Add parameter mem_canonicalized. Changed all callers.
14055 * rtl.h (canon_anti_dependence): Update prototype.
14057 2013-06-20 Richard Biener <rguenther@suse.de>
14059 * data-streamer-in.c (streamer_read_uhwi): Optimize single
14060 byte case, inline streamer_read_uchar and defer section
14063 2013-06-20 Richard Biener <rguenther@suse.de>
14065 PR tree-optimization/57584
14066 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
14067 SSA names into the expanded expression that take part in
14068 abnormal coalescing.
14070 2013-06-19 Sharad Singhai <singhai@google.com>
14072 * gcov.c (print_usage): Handle new option.
14073 (process_args): Ditto.
14074 (get_gcov_intermediate_filename): New function.
14075 (output_intermediate_file): New function.
14076 (output_gcov_file): New function
14077 (generate_results): Handle new option.
14078 (release_function): Relase demangled name.
14079 (read_graph_file): Handle demangled name.
14080 (output_lines): Ditto.
14081 * doc/gcov.texi: Document gcov intermediate format.
14083 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
14086 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
14087 (lra_emit_add): Use the functions. Add comment about Y as an
14090 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
14093 * collect2.c (collect_atexit): New.
14094 (collect_exit): Delete.
14095 (main): Register collect_atexit with atexit.
14096 (collect_wait): Change collect_exit to exit.
14098 * collect2.h (collect_exit): Delete.
14099 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
14101 2013-06-19 Wei Mi <wmi@google.com>
14103 PR rtl-optimization/57518
14104 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
14105 if regno is used in paradoxical subreg.
14106 (update_equiv_regs): Check pdx_subregs[regno] before
14107 set a reg to be equivalent with a mem.
14109 2013-06-19 Matthias Klose <doko@ubuntu.com>
14112 * file-find.h (find_a_file): Add a mode parameter.
14113 * file-find.c (find_a_file): Likewise.
14114 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
14115 with X_OK for the executables.
14116 * collect2.c (main): Call find_a_file with X_OK.
14118 2013-06-19 Steve Ellcey <sellcey@mips.com>
14121 * config/mips/mips.md (casesi_internal_mips16_<mode>):
14122 Use NEXT_INSN instead of next_real_insn.
14124 2013-06-19 Jan Hubicka <jh@suse.cz>
14126 * cgraph.h (const_value_known_p): Replace by ...
14127 (ctor_for_folding): .. this one.
14128 * cgraphunit.c (process_function_and_variable_attributes): Use it.
14129 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
14130 * expr.c (expand_expr_real_1): Likewise.
14131 (string_constant): Likewise.
14132 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
14133 * ipa.c (process_references): Likewise.
14134 (symtab_remove_unreachable_nodes): Likewise.
14135 * ipa-inline-analysis.c (param_change_prob): Likewise.
14136 * gimple-fold.c (canonicalize_constructor_val): Likewise.
14137 (get_base_constructor): Likwise.
14138 * varpool.c (varpool_remove_node): Likewise.
14139 (varpool_remove_initializer): LIkewise.
14140 (dump_varpool_node): LIkwise.
14141 (const_value_known_p): Rewrite to ...
14142 (ctor_for_folding): ... this one.
14144 2013-06-19 Jakub Jelinek <jakub@redhat.com>
14147 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
14148 PERSONALITY in $PATH derived prefixes.
14150 2013-06-19 Jeff Law <law@redhat.com>
14152 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
14155 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
14156 (simplify_bitwise_binary): Use it to simpify certain binary ops on
14159 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
14161 * config/arm/vfp.md: Move VFP instruction classification documentation
14163 * config/arm/arm.md: ... here. Update instruction classification
14166 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
14168 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
14169 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
14170 pattern. Use more efficient sequences on ARMv5 and Thumb2.
14172 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
14175 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14176 with NEXT_INSN. Use tablejump_p to check for jump table data
14179 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
14182 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
14183 that now in C++ the value is correct per the C++ standards.
14185 2013-06-19 Richard Biener <rguenther@suse.de>
14187 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
14188 for global context.
14190 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14193 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14196 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14197 with next_active_insn.
14199 2013-06-18 Sriraman Tallam <tmsriram@google.com>
14201 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
14202 functions are inlined during failures to flag an error.
14203 * tree-inline.c (expand_call_inline): Allow the error to be flagged
14204 in early inline pass.
14206 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
14208 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
14211 2013-06-18 Julian Brown <julian@codesourcery.com>
14213 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
14214 Permit virtual register pre-reload if !strict.
14215 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
14217 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
14219 * config/arm/neon.md (movmisalign<mode>): Use
14220 neon_perm_struct_or_reg_operand instead of
14221 neon_struct_or_register_operand.
14222 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
14223 neon_permissive_struct_operand instead of neon_struct_operand.
14224 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
14225 neon_vector_mem_operand.
14226 * config/arm/predicates.md (neon_struct_operand): Adjust call to
14227 neon_vector_mem_operand.
14228 (neon_permissive_struct_operand): New.
14229 (neon_struct_or_register_operand): Rename to...
14230 (neon_perm_struct_or_reg_operand): This. Adjust call to
14231 neon_vector_mem_operand.
14233 2013-06-18 Richard Biener <rguenther@suse.de>
14235 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
14236 * lto-streamer.h: Include pointer-set.h.
14237 (struct lto_decl_slot): Remove.
14238 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
14239 Remove next_index entry.
14240 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14241 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14242 (lto_init_tree_ref_encoder): Adjust.
14243 (lto_destroy_tree_ref_encoder): Likewise.
14244 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14245 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14246 (lto_output_decl_index): Adjust.
14247 (lto_new_out_decl_state): Likewise.
14248 (lto_record_function_out_decl_state): Likewise.
14249 * lto-streamer-out.c (copy_function): Likewise.
14251 2013-06-18 Richard Biener <rguenther@suse.de>
14253 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
14254 * cgraphunit.c: Include cfgloop.h.
14255 (init_lowered_empty_function): Initialize the loop tree.
14256 (assemble_thunk): Insert new BBs into loops.
14258 2013-06-18 Richard Biener <rguenther@suse.de>
14260 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
14261 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
14262 the map from cache entry to cache index optional.
14263 (streamer_tree_cache_replace_tree): Adjust accordingly.
14264 (streamer_tree_cache_append): Likewise.
14265 (streamer_tree_cache_delete): Likewise.
14266 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
14267 streamer cache map from cache entry to cache index.
14268 * lto-streamer-out.c (create_output_block): Adjust.
14270 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
14272 * config/arm/arm.md (attribute "insn"): Move multiplication and
14273 division attributes to...
14274 (attribute "type"): ... here. Remove mult.
14275 (attribute "mul32"): New attribute.
14276 (attribute "mul64"): Add umaal.
14277 (*arm_mulsi3): Update attributes.
14278 (*arm_mulsi3_v6): Likewise.
14279 (*thumb_mulsi3): Likewise.
14280 (*thumb_mulsi3_v6): Likewise.
14281 (*mulsi3_compare0): Likewise.
14282 (*mulsi3_compare0_v6): Likewise.
14283 (*mulsi_compare0_scratch): Likewise.
14284 (*mulsi_compare0_scratch_v6): Likewise.
14285 (*mulsi3addsi): Likewise.
14286 (*mulsi3addsi_v6): Likewise.
14287 (*mulsi3addsi_compare0): Likewise.
14288 (*mulsi3addsi_compare0_v6): Likewise.
14289 (*mulsi3addsi_compare0_scratch): Likewise.
14290 (*mulsi3addsi_compare0_scratch_v6): Likewise.
14291 (*mulsi3subsi): Likewise.
14292 (*mulsidi3adddi): Likewise.
14293 (*mulsi3addsi_v6): Likewise.
14294 (*mulsidi3adddi_v6): Likewise.
14295 (*mulsidi3_nov6): Likewise.
14296 (*mulsidi3_v6): Likewise.
14297 (*umulsidi3_nov6): Likewise.
14298 (*umulsidi3_v6): Likewise.
14299 (*umulsidi3adddi): Likewise.
14300 (*umulsidi3adddi_v6): Likewise.
14301 (*smulsi3_highpart_nov6): Likewise.
14302 (*smulsi3_highpart_v6): Likewise.
14303 (*umulsi3_highpart_nov6): Likewise.
14304 (*umulsi3_highpart_v6): Likewise.
14305 (mulhisi3): Likewise.
14306 (*mulhisi3tb): Likewise.
14307 (*mulhisi3bt): Likewise.
14308 (*mulhisi3tt): Likewise.
14309 (maddhisi4): Likewise.
14310 (*maddhisi4tb): Likewise.
14311 (*maddhisi4tt): Likewise.
14312 (maddhidi4): Likewise.
14313 (*maddhidi4tb): Likewise.
14314 (*maddhidi4tt): Likewise.
14315 (divsi3): Likewise.
14316 (udivsi3): Likewise.
14317 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
14318 (thumb2_mulsi_short_compare0): Likewise.
14319 (thumb2_mulsi_short_compare0_scratch): Likewise.
14320 * config/arm/arm1020e.md (1020mult1): Update attribute change.
14321 (1020mult2): Likewise.
14322 (1020mult3): Likewise.
14323 (1020mult4): Likewise.
14324 (1020mult5): Likewise.
14325 (1020mult6): Likewise.
14326 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
14328 (cortex_a15_mult64): Likewise.
14329 (cortex_a15_sdiv): Likewise.
14330 (cortex_a15_udiv): Likewise.
14331 * config/arm/arm1026ejs.md (mult1): Update attribute change.
14337 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
14338 (pj4_ir_div): Likewise.
14339 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
14340 (11_mult2): Likewise.
14341 (11_mult3): Likewise.
14342 (11_mult4): Likewise.
14343 (11_mult5): Likewise.
14344 (11_mult6): Likewise.
14345 (11_mult7): Likewise.
14346 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
14347 (cortex_a8_mla): Likewise.
14348 (cortex_a8_mull): Likewise.
14349 (cortex_a8_smulwy): Likewise.
14350 (cortex_a8_smlald): Likewise.
14351 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
14352 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
14353 (cortex_r4_mul_3): Likewise.
14354 (cortex_r4_mla_4): Likewise.
14355 (cortex_r4_mla_3): Likewise.
14356 (cortex_r4_smlald): Likewise.
14357 (cortex_r4_mull): Likewise.
14358 (cortex_r4_sdiv): Likewise.
14359 (cortex_r4_udiv): Likewise.
14360 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
14361 (cortex_a7_idiv): Likewise.
14362 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
14363 (9_mult2): Likewise.
14364 (9_mult3): Likewise.
14365 (9_mult4): Likewise.
14366 (9_mult5): Likewise.
14367 (9_mult6): Likewise.
14368 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
14369 (cortex_a53_sdiv): Likewise.
14370 (cortex_a53_udiv): Likewise.
14371 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
14372 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
14373 (mp626_mult2): Likewise.
14374 (mp626_mult3): Likewise.
14375 (mp626_mult4): Likewise.
14376 * config/arm/fa526.md (526_mult1): Update attribute change.
14377 (526_mult2): Likewise.
14378 * config/arm/arm-generic.md (mult): Update attribute change.
14379 (mult_ldsched_strongarm): Likewise.
14380 (mult_ldsched): Likewise.
14381 (multi_cycle): Likewise.
14382 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
14383 * config/arm/fa606te.md (606te_mult1): Update attribute change.
14384 (606te_mult2): Likewise.
14385 (606te_mult3): Likewise.
14386 (606te_mult4): Likewise.
14387 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
14388 (cortex_a9_mac16): Likewise.
14389 (cortex_a9_multiply): Likewise.
14390 (cortex_a9_mac): Likewise.
14391 (cortex_a9_multiply_long): Likewise.
14392 * config/arm/fa626te.md (626te_mult1): Update attribute change.
14393 (626te_mult2): Likewise.
14394 (626te_mult3): Likewise.
14395 (626te_mult4): Likewise.
14397 2013-06-18 Richard Biener <rguenther@suse.de>
14400 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
14402 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14405 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14406 with next_active_insn.
14408 2013-06-18 Alan Modra <amodra@gmail.com>
14410 * config/rs6000/rs6000.h (enum data_align): New.
14411 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
14412 (DATA_ABI_ALIGNMENT): Define.
14413 (CONSTANT_ALIGNMENT): Correct comment.
14414 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
14415 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
14417 2013-06-17 David Malcolm <dmalcolm@redhat.com>
14419 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
14420 ATTRIBUTE_UNUSED marking.
14422 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
14424 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
14425 alternative and update.
14426 (aarch64_dup_lanedi): Delete.
14427 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
14428 * config/aarch64/aarch64-simd-builtins.def: Update.
14430 2013-06-17 Richard Biener <rguenther@suse.de>
14432 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
14433 (lto_input_scc): Declare.
14434 (lto_input_tree_1): Likewise.
14435 (struct lto_stats_d): Add num_tree_bodies_output and
14436 num_pickle_refs_output.
14437 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
14438 (lto_read_tree_1): Split out from ...
14439 (lto_read_tree): ... this.
14440 (lto_input_scc): New function.
14441 (lto_input_tree_1): Split out from ...
14442 (lto_input_tree): ... this. Handle LTO_tree_scc.
14443 (lto_data_in_create): Create the streamer cache without hashes.
14444 * lto-streamer-out.c (create_output_block): Create the streamer
14445 cache with hashes when not doing WPA.
14446 (lto_write_tree_1): Split out from ...
14447 (lto_write_tree): ... this.
14448 (get_symbol_initial_value): New function.
14449 (lto_output_tree_1): Split out from ...
14450 (lto_output_tree): ... this. Write trees as series of SCCs
14451 using a DFS walk via DFS_write_tree.
14452 (struct sccs, struct scc_entry): New types.
14453 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
14454 (DFS_write_tree_body): New function.
14455 (DFS_write_tree): Likewise.
14456 (hash_tree): Likewise.
14457 (scc_entry_compare): Likewise.
14458 (hash_scc): Likewise.
14459 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
14460 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
14461 TREE_CHAIN as regular reference.
14462 (streamer_read_integer_cst): Remove.
14463 (streamer_get_pickled_tree): Adjust.
14464 * tree-streamer-out.c (streamer_write_chain): Disable streaming
14465 of DECL_EXTERNALs in BLOCK_VARS for now.
14466 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
14468 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
14469 Add hash value argument and record that if hashes are recorded
14471 (streamer_tree_cache_insert_1): Adjust.
14472 (streamer_tree_cache_insert): Likewise.
14473 (streamer_tree_cache_insert_at): Rename to ...
14474 (streamer_tree_cache_replace_tree): ... this and adjust.
14475 (streamer_tree_cache_append): Adjust.
14476 (record_common_node): Likewise.
14477 (streamer_tree_cache_create): Add argument whether to
14478 record hash values together with trees.
14479 (streamer_tree_cache_delete): Adjust.
14480 * tree-streamer.h (struct streamer_tree_cache_d): Add
14482 (streamer_read_integer_cst): Remove.
14483 (streamer_tree_cache_insert): Adjust.
14484 (streamer_tree_cache_append): Likewise.
14485 (streamer_tree_cache_insert_at): Rename to ...
14486 (streamer_tree_cache_replace_tree): ... this and adjust.
14487 (streamer_tree_cache_create): Add argument whether to record hashes.
14488 (streamer_tree_cache_get): Rename to ...
14489 (streamer_tree_cache_get_tree): ... this.
14490 (streamer_tree_cache_get_hash): New function.
14491 * tree.c (cache_integer_cst): New function.
14492 * tree.h (cache_integer_cst): Declare.
14493 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
14494 * lto-symtab.c (lto_varpool_replace_node): Only release
14495 DECL_INITIAL of non-prevailing decls.
14496 * varpool.c (varpool_remove_initializer): Do not release
14497 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
14499 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14501 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
14502 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
14503 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
14504 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
14505 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
14506 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
14507 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
14508 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
14509 instead of TARGET_64BIT.
14510 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
14511 Require ISA_HAS_<D>DIV.
14513 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
14515 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
14516 (mips*-*-linux*): Move default with_llsc setting to where other
14518 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
14520 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
14521 Likewise. Remove default with_tune setting. Move default float
14522 setting to its own block. Handle with_llsc in the same block as above.
14524 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
14526 PR rtl-optimization/57425
14527 PR rtl-optimization/57569
14528 * alias.c (write_dependence_p): Add new parameters mem_mode,
14529 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
14530 Changed all callers.
14531 (canon_anti_dependence): New function.
14532 * cse.c (check_dependence): Use canon_anti_dependence.
14533 * cselib.c (cselib_invalidate_mem): Likewise.
14534 * rtl.h (canon_anti_dependence): Declare.
14536 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14538 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
14539 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
14540 ".set mips3" for 64-bit targets.
14542 2013-06-15 Dehao Chen <dehao@google.com>
14544 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
14545 * gimple-low.c (gimple_check_call_matching_types): Likewise.
14546 (gimple_check_call_args): Likewise.
14547 * value-prof.c (check_ic_target): Likewise.
14548 * ipa-inline.c (early_inliner): Likewise.
14549 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
14550 * cgraph.c (cgraph_create_edge_1): Likewise.
14551 (cgraph_make_edge_direct): Likewise.
14553 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
14556 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
14557 rs6000_output_move_128bit to handle emitting quad memory
14558 operations. Set attribute length to 8 bytes.
14560 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
14562 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
14564 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
14565 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
14566 (aarch64_<su>mlsl<mode>): Likewise.
14568 2013-06-14 Mike Stump <mikestump@comcast.net>
14570 * Makefile.in (TARGET_H): Add insn-codes.h.
14572 2013-06-14 Alan Modra <amodra@gmail.com>
14574 PR middle-end/57134
14575 PR middle-end/57586
14576 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
14577 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
14578 bitfield expansion when EXPAND_MEMORY.
14579 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
14581 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
14583 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
14584 test for clearing quad memory on 32-bit later.
14586 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14588 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
14589 (fold_negate_expr): Likewise.
14590 (fold_real_zero_addition_p): Handle vectors.
14591 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
14593 2013-06-14 Alan Modra <amodra@gmail.com>
14595 * varasm.c (force_const_mem): Revert 2013-06-07 change.
14597 2013-06-13 Jan Hubicka <jh@suse.cz>
14599 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
14600 Local comdats are not externally visible.
14601 * symtab.c (dump_symtab_base): Dump externally visible.
14602 (verify_symtab_base): Verify back links in the symtab hash.
14604 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14606 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
14607 CONVERT_EXPR as equal nodes.
14609 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14611 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
14613 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14615 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
14616 Generalize to complex and vector.
14617 * tree.c (build_all_ones_cst): New function.
14618 * tree.h (build_all_ones_cst): Declare it.
14620 2013-06-13 Alan Modra <amodra@gmail.com>
14622 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
14623 * config/rs6000/rs6000.md (signbittf2): New insn.
14624 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
14625 (abstf2_internal, cmptf_internal2): Likewise.
14626 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
14628 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
14629 Pat Haugen <pthaugen@us.ibm.com>
14630 Peter Bergner <bergner@vnet.ibm.com>
14632 * config/rs6000/rs6000.c (emit_load_locked): Add support for
14633 power8 byte, half-word, and quad-word atomic instructions.
14634 (emit_store_conditional): Likewise.
14635 (rs6000_expand_atomic_compare_and_swap): Likewise.
14636 (rs6000_expand_atomic_op): Likewise.
14638 * config/rs6000/sync.md (larx): Add new modes for power8.
14640 (AINT): New mode iterator to include TImode as well as normal
14641 integer modes on power8.
14642 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
14643 that VSX registers are not considered. Use AINT mode iterator
14644 instead of INT1 to allow inclusion of quad word atomic operations
14646 (load_locked<mode>): Likewise.
14647 (store_conditional<mode>): Likewise.
14648 (atomic_compare_and_swap<mode>): Likewise.
14649 (atomic_exchange<mode>): Likewise.
14650 (atomic_nand<mode>): Likewise.
14651 (atomic_fetch_<fetchop_name><mode>): Likewise.
14652 (atomic_nand_fetch<mode>): Likewise.
14653 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
14655 (ATOMIC): On power8, add QImode, HImode modes.
14656 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
14657 modes that promote to SImode.
14658 (load_lockedti): Convert TImode arguments to PTImode, so that we
14659 get a guaranteed even/odd register pair.
14660 (load_lockedpti): Likewise.
14661 (store_conditionalti): Likewise.
14662 (store_conditionalpti): Likewise.
14664 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
14665 atomic load/store instructions.
14668 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
14670 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
14672 (insn_count): New attribute, with most cases extracted from...
14673 (length): ...here. Redefine most cases in terms of insn_count.
14674 (single_insn): Delete.
14675 (can_delay): Use insn_count to check for single instructions.
14676 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
14677 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
14678 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
14679 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
14680 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
14681 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
14682 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
14683 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
14684 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
14685 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
14686 rather than "length".
14687 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
14688 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
14689 Use "insn_count" rather than "length".
14690 * config/mips/mips-dsp.md
14691 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
14692 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
14695 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
14697 PR tree-optimization/57361
14698 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
14700 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
14702 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
14704 (aarch64_simd_combine<mode>): New instruction expansion.
14705 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
14706 function prototype.
14707 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
14708 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
14710 2013-06-12 Jan Hubicka <jh@suse.cz>
14712 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
14713 decl has when in streaming stage.
14714 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
14715 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
14717 2013-06-12 Roland Stigge <stigge@antcom.de>
14720 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
14722 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14724 PR tree-optimization/57537
14725 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
14726 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
14728 2013-06-12 Richard Biener <rguenther@suse.de>
14730 * data-streamer.h (streamer_write_char_stream): CSE
14731 obs->current_pointer.
14732 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
14733 streamer_write_char_stream manually and optimize the resulting loop.
14734 (streamer_write_hwi_stream): Likewise.
14736 2013-06-12 Jan Hubicka <jh@suse.cz>
14738 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
14739 * cgraph.h (varpool_create_empty_node): Declare.
14740 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
14742 * symtab.c (symtab_unregister_node): Be lax about missin entries
14744 (symtab_get_node): Update comment.
14745 * varpool.c (varpool_create_empty_node): Break out from ...
14746 (varpool_node_for_decl): ... here.
14747 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
14749 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
14751 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
14752 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
14753 part. Use straight-line flow at the end.
14754 <COMPONENT_REF>: Remove superfluous else.
14755 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
14757 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14760 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
14761 target hook even for !TREE_PUBLIC decls. If no resolution info
14762 is available, return false for common and external decls.
14764 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14766 * config/rl78/constraints.md (U): New constraint.
14767 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
14770 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
14773 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
14774 to allow returning address to AT_PLATFORM name.
14776 2013-06-11 Jan Hubicka <jh@suse.cz>
14778 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
14779 * cgraph.h (symtab_node_base): Add weakref flag.
14780 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
14781 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
14782 (output_weakrefs): Use weakref flag.
14783 * fold-const.c (simple_operand_p): Handle WEAK.
14784 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
14785 * ipa.c (varpool_externally_visible_p): Drop weakref.
14786 (function_and_variable_visibility): Update comment; fix weakref
14787 sanity checks; do not clear DECL_WEAK on them.
14788 * lto-cgraph.c (lto_output_node): update.
14789 (lto_output_varpool_node): Update.
14790 (input_overwrite_node): Update.
14791 (input_node): Update.
14792 (input_varpool_node): Update.
14793 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
14794 (lto_symtab_merge_symbols): Add sanity check.
14795 (lto_symtab_prevailing_decl): Do not special case weakrefs.
14796 * passes.c (rest_of_decl_compilation): Set static flag, too.
14797 * symtab.c (dump_symtab_base): Dump weakref.
14798 (verify_symtab_base): Sanity check weakrefs.
14799 (symtab_make_decl_local): Remove duplicated code.
14800 (symtab_alias_ultimate_target): Simplify.
14801 * varpool.c (varpool_create_variable_alias): Set weakref flag.
14803 2013-06-11 Tom de Vries <tom@codesourcery.com>
14805 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
14806 sequence_vect == NULL.
14808 2013-06-11 DJ Delorie <dj@redhat.com>
14810 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
14811 (rl78_unwind_word_mode): New.
14813 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14815 * final.c (debug_prefix_maps): Make static.
14817 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14819 * function.c (initial_trampoline): Remove stray copy.
14821 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
14823 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
14825 2013-06-11 Martin Jambor <mjambor@suse.cz>
14827 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
14828 within bounds at the beginning of the function.
14830 2013-06-11 Alan Modra <amodra@gmail.com>
14832 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
14834 (get_named_section): Don't NULL !DECL_P decl.
14836 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
14838 * doc/invoke.texi (core-avx2): Document.
14840 (atom): Updated with MOVBE.
14842 2013-06-11 Richard Biener <rguenther@suse.de>
14844 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
14846 2013-06-11 Anton Blanchard <anton@samba.org>
14848 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
14849 correct shift value in little-endian mode.
14851 2013-06-11 Jakub Jelinek <jakub@redhat.com>
14854 * varasm.c (get_variable_align): Move #endif to the right place.
14856 2013-06-10 Cary Coutant <ccoutant@google.com>
14858 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
14859 for hash so that hash table traversal order is deterministic.
14861 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
14862 Pat Haugen <pthaugen@us.ibm.com>
14863 Peter Bergner <bergner@vnet.ibm.com>
14865 * config/rs6000/vector.md (GPR move splitter): Do not split moves
14866 of vectors in GPRS if they are direct moves or quad word load or
14869 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
14871 (direct_move_p): Likewise.
14872 (quad_load_store_p): Likewise.
14874 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
14875 classes into bins based on the physical register type.
14876 (reg_class_to_reg_type): Likewise.
14877 (IS_STD_REG_TYPE): Likewise.
14878 (IS_FP_VECT_REG_TYPE): Likewise.
14879 (reload_fpr_gpr): Arrays to determine what insn to use if we can
14880 use direct move instructions.
14881 (reload_gpr_vsx): Likewise.
14882 (reload_vsx_gpr): Likewise.
14883 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
14884 information that is a simplification of register classes. Also
14885 precalculate direct move reload helpers.
14886 (direct_move_p): New function to return true if the operation can
14887 be done as a direct move instruciton.
14888 (quad_load_store_p): New function to return true if the operation
14889 is a quad memory operation.
14890 (rs6000_legitimize_address): If quad memory, only allow register
14891 indirect for TImode addresses.
14892 (rs6000_legitimate_address_p): Likewise.
14893 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
14894 (rs6000_reload_register_type): Likewise.
14895 (register_to_reg_type): Return register type.
14896 (rs6000_secondary_reload_simple_move): New helper function for
14897 secondary reload and secondary memory needed to identify anything
14898 that is a simple move, and does not need reloading.
14899 (rs6000_secondary_reload_direct_move): New helper function for
14900 secondary reload to identify cases that can be done with several
14901 instructions via the direct move instructions.
14902 (rs6000_secondary_reload_move): New helper function for secondary
14903 reload to identify moves between register types that can be done.
14904 (rs6000_secondary_reload): Add support for quad memory operations
14905 and for direct move.
14906 (rs6000_secondary_memory_needed): Likewise.
14907 (rs6000_debug_secondary_memory_needed): Change argument names.
14908 (rs6000_output_move_128bit): New function to return the move to
14909 use for 128-bit moves, including knowing about the various
14910 limitations of quad memory operations.
14912 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
14913 memory operations. call rs6000_output_move_128bit for the actual
14914 instruciton(s) to generate.
14915 (vsx_movti_64bit): Likewise.
14917 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
14918 (UNSPEC_P8V_MTVSRWZ): Likewise.
14919 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
14920 (UNSPEC_P8V_MTVSRD): Likewise.
14921 (UNSPEC_P8V_XXPERMDI): Likewise.
14922 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
14923 (UNSPEC_FUSION_GPR): Likewise.
14924 (FMOVE128_GPR): New iterator for direct move.
14925 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
14926 (f32_sv): Likewise.
14927 (f32_dm): Likewise.
14928 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
14929 loads and direct move instructions.
14930 (zero_extendsidi2_lfiwzx): Likewise.
14931 (extendsidi2_lfiwax): Likewise.
14932 (extendsidi2_nocell): Likewise.
14933 (floatsi<mode>2_lfiwax): Likewise.
14934 (lfiwax): Likewise.
14935 (floatunssi<mode>2_lfiwzx): Likewise.
14936 (lfiwzx): Likewise.
14937 (fix_trunc<mode>_stfiwx): Likewise.
14938 (fixuns_trunc<mode>_stfiwx): Likewise.
14939 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
14940 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
14941 (parity<mode>2_cmpb): Set length/type attr.
14942 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
14943 for 'mr.' to fast_compare.
14944 (bpermd_<mode>): Change type attr to popcnt.
14945 (p8_fmrgow_<mode>): New insns for power8 direct move support.
14946 (p8_mtvsrwz_1): Likewise.
14947 (p8_mtvsrwz_2): Likewise.
14948 (reload_fpr_from_gpr<mode>): Likewise.
14949 (p8_mtvsrd_1): Likewise.
14950 (p8_mtvsrd_2): Likewise.
14951 (p8_xxpermdi_<mode>): Likewise.
14952 (reload_vsx_from_gpr<mode>): Likewise.
14953 (reload_vsx_from_gprsf): Likewise.
14954 (p8_mfvsrd_3_<mode>): LIkewise.
14955 (reload_gpr_from_vsx<mode>): Likewise.
14956 (reload_gpr_from_vsxsf): Likewise.
14957 (p8_mfvsrd_4_disf): Likewise.
14958 (multi-word GPR splits): Do not split direct moves or quad memory
14961 2013-06-10 David Malcolm <dmalcolm@redhat.com>
14963 * tree-into-ssa.c (interesting_blocks): Make static.
14965 2013-06-10 Jakub Jelinek <jakub@redhat.com>
14968 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
14969 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
14970 Use DATA_ABI_ALIGNMENT for that case instead if defined.
14971 (get_variable_align): New function.
14972 (get_variable_section, emit_bss, emit_common,
14973 assemble_variable_contents, place_block_symbol): Use
14974 get_variable_align instead of DECL_ALIGN.
14975 (assemble_noswitch_variable): Add align argument, use it
14976 instead of DECL_ALIGN.
14977 (assemble_variable): Adjust caller. Use get_variable_align
14978 instead of DECL_ALIGN.
14979 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
14981 (DATA_ABI_ALIGNMENT): Define.
14982 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
14983 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
14984 opt is false, only return the psABI mandated alignment increase.
14985 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
14986 (DATA_ABI_ALIGNMENT): ... this.
14987 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
14988 (DATA_ABI_ALIGNMENT): ... this.
14989 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
14990 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
14991 (DATA_ABI_ALIGNMENT): ... this.
14992 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
14993 * doc/tm.texi: Regenerated.
14995 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
14997 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
14998 cmp_code to construct REG_EQUAL note.
15000 2013-06-09 Jakub Jelinek <jakub@redhat.com>
15003 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
15004 that operands[2] doesn't overlap with operands[0].
15006 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
15007 Jan Hubicka <jh@suse.cz>
15009 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
15010 hack to mark symbols as used.
15012 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
15014 PR rtl-optimization/57559
15015 * lra-constraints.c (process_alt_operands): Don't discourage
15016 memory with known offset for offsetable memory constraint.
15017 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
15019 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
15021 * varasm.c (struct oc_local_state): Reorder fields.
15022 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
15023 and adjust accordingly.
15024 (output_constructor): Reorder initialization code and adjust call to
15025 output_constructor_bitfield.
15027 2013-06-07 Jan Hubicka <jh@suse.cz>
15029 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
15031 2013-06-07 David Malcolm <dmalcolm@redhat.com>
15033 * tree-object-size.c (unknown): Make const.
15035 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15037 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
15038 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
15039 for last alternative in the cpu_facility attribute.
15041 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15044 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
15045 (xordi3): Change operand 2 constraint to arm_xordi_operand.
15046 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
15047 * config/arm/constraints.md (Dg): New constraint.
15048 * config/arm/neon.md (xordi3_neon): Remove.
15049 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
15050 * config/arm/predicates.md (arm_xordi_operand): New predicate.
15052 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15054 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
15055 Clean up alternatives.
15057 2013-06-07 Alan Modra <amodra@gmail.com>
15059 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
15062 2013-06-07 Alan Modra <amodra@gmail.com>
15064 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
15066 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15068 * config/arm/constraints.md (Df): New constraint.
15069 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
15070 Correct length attribute for last two alternatives.
15072 2013-06-07 Alan Modra <amodra@gmail.com>
15074 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
15075 override user -mfp-in-toc.
15076 (offsettable_ok_by_alignment): Consider just the current access
15077 rather than the whole object, unless BLKmode. Handle
15078 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
15079 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
15080 for -mcmodel=medium.
15081 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
15082 override user -mfp-in-toc or -msum-in-toc. Default to
15083 -mno-fp-in-toc for -mcmodel=medium.
15085 2013-06-06 DJ Delorie <dj@redhat.com>
15087 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
15088 TARGET_VALID_POINTER_MODE.
15090 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
15091 Pat Haugen <pthaugen@us.ibm.com>
15092 Peter Bergner <bergner@vnet.ibm.com>
15094 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
15095 Document new power8 builtins.
15097 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
15098 condition code register, to allow 128-bit logical operations to be
15099 done in the VSX or GPR registers.
15100 (nor<mode>3): Use the canonical form for nor.
15101 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
15102 vclz*, and vpopcnt* vector instructions.
15103 (nand<mode>3): Likewise.
15104 (orc<mode>3): Likewise.
15105 (clz<mode>2): LIkewise.
15106 (popcount<mode>2): Likewise.
15108 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
15109 that only the GPRs are recognized.
15111 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15112 support for new power8 builtins.
15114 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
15116 (xscvdpspn): Likewise.
15122 (vpopcnt): Likewise.
15123 (vpopcntb): Likewise.
15124 (vpopcnth): Likewise.
15125 (vpopcntw): Likewise.
15126 (vpopcntd): Likewise.
15128 (vmrgew): Likewise.
15129 (vmrgow): Likewise.
15131 (eqv_v16qi3): Likewise.
15132 (eqv_v8hi3): Likewise.
15133 (eqv_v4si3): Likewise.
15134 (eqv_v2di3): Likewise.
15135 (eqv_v4sf3): Likewise.
15136 (eqv_v2df3): Likewise.
15138 (nand_v16qi3): Likewise.
15139 (nand_v8hi3): Likewise.
15140 (nand_v4si3): Likewise.
15141 (nand_v2di3): Likewise.
15142 (nand_v4sf3): Likewise.
15143 (nand_v2df3): Likewise.
15145 (orc_v16qi3): Likewise.
15146 (orc_v8hi3): Likewise.
15147 (orc_v4si3): Likewise.
15148 (orc_v2di3): Likewise.
15149 (orc_v4sf3): Likewise.
15150 (orc_v2df3): Likewise.
15152 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
15153 allow power8 quad mode in 64-bit.
15154 (rs6000_builtin_vectorized_function): Add support to vectorize
15155 ISA 2.07 count leading zeros, population count builtins.
15156 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
15157 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
15158 (builtin_function_type): Add vgbbd builtin function which takes an
15160 (altivec_expand_vec_perm_const): Add support for new power8 merge
15163 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
15164 that does not include TImdoe for use with 32-bit.
15165 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
15167 (UNSPEC_VSX_CVDPSPN): Likewise.
15168 (vsx_xscvdpspn): Likewise.
15169 (vsx_xscvspdpn): Likewise.
15170 (vsx_xscvdpspn_scalar): Likewise.
15171 (vsx_xscvspdpn_directmove): Likewise.
15172 (vsx_and<mode>3): Split logical operations into 32-bit and
15173 64-bit. Add support to do logical operations on TImode as well as
15174 VSX vector types. Allow logical operations to be done in either
15175 VSX registers or in general purpose registers in 64-bit mode. Add
15176 splitters if GPRs were used. For AND, add clobber of CCmode to
15177 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
15179 (vsx_and<mode>3_32bit): Likewise.
15180 (vsx_and<mode>3_64bit): Likewise.
15181 (vsx_ior<mode>3): Likewise.
15182 (vsx_ior<mode>3_32bit): Likewise.
15183 (vsx_ior<mode>3_64bit): Likewise.
15184 (vsx_xor<mode>3): Likewise.
15185 (vsx_xor<mode>3_32bit): Likewise.
15186 (vsx_xor<mode>3_64bit): Likewise.
15187 (vsx_one_cmpl<mode>2): Likewise.
15188 (vsx_one_cmpl<mode>2_32bit): Likewise.
15189 (vsx_one_cmpl<mode>2_64bit): Likewise.
15190 (vsx_nor<mode>3): Likewise.
15191 (vsx_nor<mode>3_32bit): Likewise.
15192 (vsx_nor<mode>3_64bit): Likewise.
15193 (vsx_andc<mode>3): Likewise.
15194 (vsx_andc<mode>3_32bit): Likewise.
15195 (vsx_andc<mode>3_64bit): Likewise.
15196 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
15197 and xxlorc instructions.
15198 (vsx_eqv<mode>3_64bit): Likewise.
15199 (vsx_nand<mode>3_32bit): Likewise.
15200 (vsx_nand<mode>3_64bit): Likewise.
15201 (vsx_orc<mode>3_32bit): Likewise.
15202 (vsx_orc<mode>3_64bit): Likewise.
15204 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
15206 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
15208 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
15209 (p8_vmrgow): Likewise.
15210 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
15211 GPRs to be split under VSX.
15212 (p8v_clz<mode>2): Add power8 count leading zero support.
15213 (p8v_popcount<mode>2): Add power8 population count support.
15214 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
15217 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
15220 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
15222 (vec_nand): Likewise.
15223 (vec_vclz): Likewise.
15224 (vec_vclzb): Likewise.
15225 (vec_vclzd): Likewise.
15226 (vec_vclzh): Likewise.
15227 (vec_vclzw): Likewise.
15228 (vec_vgbbd): Likewise.
15229 (vec_vmrgew): Likewise.
15230 (vec_vmrgow): Likewise.
15231 (vec_vpopcnt): Likewise.
15232 (vec_vpopcntb): Likewise.
15233 (vec_vpopcntd): Likewise.
15234 (vec_vpopcnth): Likewise.
15235 (vec_vpopcntw): Likewise.
15237 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15239 PR rtl-optimization/57468
15240 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
15243 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15245 PR rtl-optimization/57459
15246 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
15247 type when setting live regs.
15249 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15251 * config/s390/s390.opt (mlra): New option.
15252 * config/s390/s390.c (s390_decompose_address): Check displacement
15253 for all registers for LRA.
15254 (s390_secondary_reload): Don't used secondary reloads for LRA.
15255 (s390_lra_p): New function.
15256 (TARGET_LRA_P): Define.
15257 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
15258 of attribute cpu_facility to zarch for the last alternative.
15259 (*cmpmem_short): Ditto.
15261 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
15263 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
15264 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
15265 (arm_expand_prologue): Likewise.
15267 2013-06-06 Teresa Johnson <tejohnson@google.com>
15270 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
15271 as this is now done by redirect_edge_and_branch_force.
15272 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
15273 barriers, and fix interaction with splitting.
15274 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
15275 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
15276 reflect changes made in the routine.
15277 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
15278 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
15279 since this is called in cfglayout mode, and replace partition fixup
15280 with assert as that is now done by force_nonfallthru_and_redirect.
15281 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
15282 already be marked with region crossing note.
15283 (insert_section_boundary_note): Make non-static, gate on flag
15284 has_bb_partition, rewrite to also check for multiple partitions.
15285 (rest_of_handle_reorder_blocks): Remove call to
15286 insert_section_boundary_note, now done later during free_cfg.
15287 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
15288 * bb-reorder.h (insert_section_boundary_note): Declare.
15289 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
15290 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
15291 invoke insert_section_boundary_note.
15292 (try_redirect_by_replacing_jump): Remove unnecessary
15293 check for region crossing note.
15294 (fixup_partition_crossing): New function.
15295 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
15296 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
15297 in non-cfglayout mode.
15298 (force_nonfallthru_and_redirect): Fixup partition boundaries,
15299 remove old code that tried to do this. Emit barrier correctly
15300 when we are in cfglayout mode.
15301 (last_bb_in_partition): New function.
15302 (rtl_split_edge): Correctly fixup partition boundaries.
15303 (commit_one_edge_insertion): Remove old code that tried to
15304 fixup region crossing edge since this is now handled in
15305 split_block, and set up insertion point correctly since
15306 block may now end in a jump.
15307 (verify_hot_cold_block_grouping): Guard against checking when not in
15308 linearized RTL mode.
15309 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
15311 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
15312 rtl_verify_flow_info, so not called in cfglayout mode.
15313 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
15314 (fixup_reorder_chain): Remove old code that attempted to fixup region
15315 crossing note as this is now handled in force_nonfallthru_and_redirect.
15316 (duplicate_insn_chain): Don't duplicate switch section notes.
15317 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
15319 * basic-block.h (emit_barrier_after_bb): Declare.
15321 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15323 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
15324 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
15325 arm_usatsihi): Adjust alternatives for arm_restrict_it.
15327 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15329 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
15331 * config/arm/ldmstm.md: Regenerate.
15333 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15335 * config/arm/sync.md (atomic_loaddi_1):
15336 Disable predication for arm_restrict_it.
15337 (arm_load_exclusive<mode>): Likewise.
15338 (arm_load_exclusivesi): Likewise.
15339 (arm_load_exclusivedi): Likewise.
15340 (arm_load_acquire_exclusive<mode>): Likewise.
15341 (arm_load_acquire_exclusivesi): Likewise.
15342 (arm_load_acquire_exclusivedi): Likewise.
15343 (arm_store_exclusive<mode>): Likewise.
15344 (arm_store_exclusive<mode>): Likewise.
15345 (arm_store_release_exclusivedi): Likewise.
15346 (arm_store_release_exclusive<mode>): Likewise.
15348 2013-06-06 Richard Biener <rguenther@suse.de>
15350 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
15352 (lto_tag_is_tree_code_p): Adjust.
15353 (lto_tag_is_gimple_code_p): Likewise.
15354 (lto_gimple_code_to_tag): Likewise.
15355 (lto_tag_to_gimple_code): Likewise.
15356 (lto_tree_code_to_tag): Likewise.
15357 (lto_tag_to_tree_code): Likewise.
15358 * data-streamer.h (streamer_write_hwi_in_range): Use
15359 uhwi streaming to stream the normalized range.
15360 (streamer_read_hwi_in_range): Likewise.
15362 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15364 * config/arm/arm.md (enabled_for_depr_it): New attribute.
15365 (predicable_short_it): Likewise.
15366 (predicated): Likewise.
15367 (enabled): Handle above.
15368 (define_cond_exec): Set predicated attribute to yes.
15370 2013-06-05 Mike Stump <mikestump@comcast.net>
15372 * gdbinit.in (__FUNCTION__): Add.
15374 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
15376 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
15377 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
15379 2013-06-05 Jan Hubicka <jh@suse.cz>
15381 * varasm.c (mark_decl_referenced): Revert the removal until targets
15384 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
15386 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
15387 instead of mark_decl_referenced.
15389 2013-06-05 Jan Hubicka <jh@suse.cz>
15391 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
15392 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
15393 and symtab_used_from_object_file_p.
15394 (cgraph_make_node_local_1): Clear forced_by_abi.
15395 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
15396 * cgraph.h (symtab_node_base): Add forced_by_abi.
15397 (decide_is_variable_needed): Remove.
15398 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
15399 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
15400 (decide_is_symbol_needed): ... this one; handle symbols in general;
15401 always analyze virtuals; honnor forced_by_abi.
15402 (cgraph_finalize_function): Update.
15403 (varpool_finalize_decl): Update.
15404 (symbol_defined_and_needed): Remove.
15405 (analyze_functions): Update.
15406 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
15407 output_refs, input_overwrite_node): Handle forced_by_abi.
15408 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
15409 (address_taken_from_non_vtable_p): ... this one.
15410 (comdat_can_be_unshared_p_1): New function.
15411 (cgraph_comdat_can_be_unshared_p): Rename to ...
15412 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
15413 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
15414 (function_and_variable_visibility): Clear forced_by_abi as needed.
15415 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
15416 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
15417 * symtab.c (dump_symtab_base): Dump forced_by_abi.
15418 * varpool.c (decide_is_variable_needed): Remove.
15420 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15422 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
15423 (arm_option_override): Override arm_restrict_it where appropriate.
15424 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
15425 * config/arm/arm.opt (mrestrict-it): New command-line option.
15426 * doc/invoke.texi: Document -mrestrict-it.
15428 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15430 * tsan.c (tsan_atomic_table): Make const.
15432 2013-06-05 Richard Biener <rguenther@suse.de>
15434 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
15435 index associated with the tree we are supposed to replace.
15436 * tree-streamer-out.c (pack_ts_base_value_fields): Output
15437 TREE_ASM_WRITTEN as zero for everything but SSA names.
15439 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15441 * tree-ssa-structalias.c (call_stmt_vars): Make static.
15443 2013-06-04 Jan Hubicka <jh@suse.cz>
15445 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
15446 (input_node, input_varpool_node): Handle correctly external same
15448 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
15449 nodes at ltrans stage.
15451 2013-06-04 Jan Hubicka <jh@suse.cz>
15453 * ipa-inline.c (update_caller_keys): Fix availability test.
15454 (update_callee_keys): Likewise.
15455 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
15456 to follow ELF standard.
15458 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
15460 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
15461 (mips64r5900el-*-elf*): New configurations.
15462 * config/mips/mips-cpus.def (r5900): New processor.
15463 * config/mips/mips-tables.opt: Regenerate.
15464 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
15465 (mips_issue_rate): Handle PROCESSOR_R5900.
15466 (mips_reorg_process_insns): Force reorder mode for the R5900.
15467 * config/mips/mips.h (TARGET_MIPS5900): Define.
15468 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
15470 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
15472 * config/mips/mips.md (processor): Add r5900.
15473 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
15475 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15477 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
15478 into function to generate MOVI instruction.
15479 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
15480 (aarch64_preferred_simd_mode): Turn into wrapper.
15481 (aarch64_output_scalar_simd_mov_immediate): New function.
15482 * config/aarch64/aarch64-protos.h: Add prototype for above.
15484 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15486 * config/aarch64/aarch64.c (simd_immediate_info): Remove
15487 element_char member.
15488 (sizetochar): Return signed char.
15489 (aarch64_simd_valid_immediate): Remove elchar and other
15490 unnecessary variables.
15491 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
15492 Calculate element_char as required.
15493 * config/aarch64/aarch64-protos.h: Update and move prototype
15494 for aarch64_output_simd_mov_immediate.
15495 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
15498 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15500 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
15501 information completed by aarch64_simd_valid_immediate.
15502 (aarch64_legitimate_constant_p): Update arguments.
15503 (aarch64_simd_valid_immediate): Work with struct rather than many
15505 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
15506 (aarch64_simd_make_constant): Update arguments.
15507 (aarch64_output_simd_mov_immediate): Work with struct rather than
15508 many pointers. Output immediate directly rather than as operand.
15509 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15511 * config/aarch64/constraints.md (Dn): Update arguments.
15513 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15515 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
15517 (aarch64_simd_immediate_valid_for_move): Remove.
15518 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
15519 (aarch64_simd_make_constant): Update call.
15520 (aarch64_output_simd_mov_immediate): Update call.
15521 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15523 * config/aarch64/constraints.md (Dn): Update call.
15525 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15527 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
15528 return type to bool for prototype.
15529 (aarch64_legitimate_constant_p): Check for true instead of not -1.
15530 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
15531 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
15533 2013-06-04 Catherine Moore <clm@codesourcery.com>
15535 * config/mips/mips.opt (meva): New.
15536 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
15537 (ASM_SPEC): Handle -meva.
15538 * doc/invoke.texi (meva): Document.
15540 2013-06-04 Alan Modra <amodra@gmail.com>
15542 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
15545 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15547 * rtl.def: Add extra fourth optional field to define_cond_exec.
15548 * gensupport.c (process_one_cond_exec): Process attributes from
15550 * doc/md.texi: Document fourth field in define_cond_exec.
15552 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
15554 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
15555 out the processing order as in store_bit_field_1.
15557 2013-06-04 Jan Hubicka <jh@suse.cz>
15559 PR middle-end/57500
15560 * cgraphunit.c (cgraph_process_same_body_aliases): Create
15561 non-VAR_DECL node if it does not exist yet.
15563 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
15565 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
15566 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
15567 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
15568 target_cpu_default setting.
15570 2013-06-03 Teresa Johnson <tejohnson@google.com>
15572 * dumpfile.c (opt_info_switch_p): Change -fopt-info
15573 default to -fopt-info=optimized instead of all.
15574 * doc/invoke.texi: Ditto.
15575 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
15576 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
15577 (execute_vect_slp): Emit BB vectorization success under
15578 MSG_OPTIMIZED_LOCATIONS.
15579 * tree-vect-slp.c (vect_slp_transform_bb): Change
15580 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
15581 * tree-vect-loop.c (vect_transform_loop): Ditto.
15583 2013-06-03 Jason Merrill <jason@redhat.com>
15586 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
15587 Use TARGET_EXPR for C++.
15589 2013-06-03 Jakub Jelinek <jakub@redhat.com>
15591 PR rtl-optimization/57268
15592 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
15593 if DEBUG_INSN_P (insn).
15596 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15598 PR rtl-optimization/57268
15599 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15600 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15602 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15604 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
15605 (ix86_avoid_lea_for_addr): Likewise.
15606 (exact_dependency_1): Likewise.
15607 (ix86_adjust_cost): Likewise.
15608 (swap_top_of_ready_list): Fix formatting and !reload_completed check
15610 (do_reorder_for_imul): Fix typo, formatting and
15611 !reload_completed check removed.
15612 (ix86_sched_reorder): Fix typo and formatting.
15613 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
15616 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
15618 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
15620 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
15622 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
15623 <STRING_CST>: Likewise.
15624 <VECTOR_CST>: Likewise.
15626 2013-06-01 Janus Weil <janus@gcc.gnu.org>
15627 Mikael Morin <mikael@gcc.gnu.org>
15629 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
15630 * config.in: Regenerated.
15631 * configure: Regenerated.
15633 2013-06-01 Jan Hubicka <jh@suse.cz>
15635 PR middle-end/57366
15636 * cgraphunit.c (compile): When weakref is not supported,
15637 set up transparent aliases before final output pass.
15638 * varasm.c (assemble_alias): Do not try to do it here.
15640 2013-06-01 Jan Hubicka <jh@suse.cz>
15642 PR middle-end/57467
15643 * passes.c (for_per_function): Skip unanalyzed functions.
15645 2013-06-01 Jan Hubicka <jh@suse.cz>
15647 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
15648 (lto_symtab_merge_symbols_1): ... this one.
15649 (lto_symtab_merge_cgraph_nodes): Rename to ...
15650 (lto_symtab_merge_symbols): ... this one; simplify.
15651 * cgraph.c (same_body_aliases_done): Rename to ...
15652 (cpp_implicit_aliases_done): ... this one.
15653 (cgraph_create_function_alias): Update.
15654 (cgraph_same_body_alias): Update.
15655 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
15656 (verify_edge_corresponds_to_fndecl): Simplify.
15657 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
15658 (cgraph_node): Remove same_body_alias.
15659 (varpool_node): Remove alias_of and extra_name_alias.
15660 (same_body_aliases_done): Rename to ..
15661 (cpp_implicit_aliases_done): ... this one.
15662 (symtab_alias_ultimate_target): Add default parameter.
15663 (symtab_resolve_alias): New function.
15664 (fixup_same_cpp_alias_visibility): Declare.
15665 (cgraph_function_node): Add default parameter.
15666 (cgraph_node_asm_name): Likewise.
15667 (cgraph_function_or_thunk_node): Add default parameter; do
15668 not ICE when it is NULL.
15669 (varpool_variable_node): Likewise.
15670 * tree-emutls.c (create_emultls_var): Update.
15671 (ipa_lower_emutls): Update.
15672 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
15673 (cgraph_reset_node): Reset alias info.
15674 (cgraph_finalize_function): Update.
15675 (fixup_same_cpp_alias_visibility): Move to symtab.c.
15676 (analyze_function): Simplify.
15677 (cgraph_process_same_body_aliases): Simplify.
15678 (analyze_functions): Fixup same body aliases.
15679 (handle_alias_pairs): Simplify.
15680 (assemble_thunk): Update.
15681 (assemble_thunks_and_aliases): Update.
15682 (output_weakrefs): Rewrite.
15683 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
15684 (lto_output_varpool_node): Likewise.
15685 (compute_ltrans_boundary): Remve assert.
15686 (get_alias_symbol): New functoin.
15687 (input_node): Rewrite alias handling.
15688 (input_varpool_node): Likewise.
15689 * ipa-pure-const.c (propagate_pure_const): Fix formating.
15690 * ipa.c (process_references): Handle weakrefs correctly.
15691 (symtab_remove_unreachable_nodes): Likewise.
15692 * trans-mem.c (get_cg_data): Update.
15693 (ipa_tm_create_version_alias): Update.
15694 (ipa_tm_execute): Update.
15695 * symtab.c (dump_symtab_base): Dump aliases.
15696 (verify_symtab_base): Verify aliases.
15697 (symtab_node_availability): New function.
15698 (symtab_alias_ultimate_target): Simplify.
15699 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
15700 handle all the fixup cases.
15701 (symtab_resolve_alias): New function.
15702 * passes.c (ipa_write_summaries): Handle weakrefs.
15703 * varpool.c (varpool_analyze_node): Simplify.
15704 (assemble_aliases): Update.
15705 (varpool_create_variable_alias): Simplify.
15706 (varpool_extra_name_alias): Simplify.
15707 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
15708 (lto_symtab_merge_symbols): ... this one.
15710 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
15713 PR rtl-optimization/57268
15714 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15715 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15717 2013-06-01 Tobias Burnus <burnus@net-b.de>
15719 Partially reverted:
15720 2013-05-31 Tobias Burnus <burnus@net-b.de>
15722 PR middle-end/57073
15723 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
15726 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15728 PR rtl-optimization/57268
15729 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15730 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15732 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15734 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
15735 unordered comparison operators when -fno-trapping-math is in effect
15737 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
15738 and implement unordered comparison operators properly on the e500.
15740 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15742 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
15743 for constant scalar integers.
15744 (simplify_relational_operation_1): Likewise.
15746 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
15748 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
15749 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
15752 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15753 Igor Zamyatin <igor.zamyatin@intel.com>
15755 Silvermont (SLM) architecture performance tuning.
15756 * config/i386/i386.h (enum ix86_tune_indices): Add
15757 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
15758 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
15760 * config/i386/i386.c (initial_ix86_tune_features)
15761 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
15762 (ix86_lea_outperforms): Handle Silvermont tuning.
15763 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
15765 (ix86_use_lea_for_mov): Likewise.
15766 (ix86_avoid_lea_for_addr): Likewise.
15767 (ix86_lea_for_add_ok): Likewise.
15768 (exact_dependency_1): New function.
15769 (exact_store_load_dependency): Likewise.
15770 (ix86_adjust_cost): Handle Silvermont tuning.
15771 (do_reoder_for_imul): Likewise.
15772 (swap_top_of_ready_list): New function.
15773 (ix86_sched_reorder): Changed to handle Silvermont tuning.
15775 * config/i386/i386.md (peepholes that split memory operand in fp
15778 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15780 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
15781 Remove un-necessary braces.
15783 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15785 * config/aarch64/aarch64.c (aarch64_classify_symbol):
15786 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
15788 2013-05-31 Tobias Burnus <burnus@net-b.de>
15790 PR middle-end/57073
15791 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
15793 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15796 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
15797 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
15798 * config/arm/neon.md (iordi3_neon): Remove.
15799 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
15800 * config/arm/predicates.md (imm_for_neon_logic_operand):
15801 Move to earlier in the file.
15802 (neon_logic_op2): Likewise.
15803 (arm_iordi_operand_neon): New predicate.
15805 2013-05-31 Richard Biener <rguenther@suse.de>
15807 PR tree-optimization/57478
15808 PR tree-optimization/57453
15809 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
15812 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15814 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
15815 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
15817 2013-05-30 Tobias Burnus <burnus@net-b.de>
15818 Thomas Koenig <tkoenig@gcc.gnu.org>
15820 PR middle-end/57073
15821 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
15822 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
15824 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
15826 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
15828 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
15830 * target.def (register_usage_leveling_p): New hook.
15831 * targhooks.c (default_register_usage_leveling_p): New.
15832 * targhooks.h (default_register_usage_leveling_p): New prototype.
15833 * lra-assigns.c (register_usage_leveling_p): Use the hook.
15834 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
15835 * doc/tm.texi: Update.
15836 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
15838 2013-05-30 Ian Bolton <ian.bolton@arm.com>
15840 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
15841 (*insv_reg<mode>): New define_insn.
15843 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
15845 PR rtl-optimization/57439
15846 * postreload.c (move2add_valid_value_p): Check that we have
15847 a zero subreg_regno_offset when accessing the register in
15848 the requested mode.
15850 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15851 Igor Zamyatin <igor.zamyatin@intel.com>
15853 Silvermont (SLM) architecture pipeline model, tuning and
15855 * config.gcc: Add slm config options and target.
15857 * config/i386/slm.md: New.
15859 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
15861 * config/i386/i386-c.c (ix86_target_macros_internal): New case
15863 (ix86_target_macros_internal): Likewise.
15865 * config/i386/i386.c (slm_cost): New cost.
15866 (m_SLM): New macro flag.
15867 (initial_ix86_tune_features): Set m_SLM.
15868 (x86_accumulate_outgoing_args): Likewise.
15869 (x86_arch_always_fancy_math_387): Likewise.
15870 (processor_target_table): Add slm cost.
15871 (cpu_names): Add slm cpu name.
15872 (x86_option_override_internal): Set SLM ISA.
15873 (ix86_issue_rate): New case PROCESSOR_SLM.
15874 (ia32_multipass_dfa_lookahead): Likewise.
15875 (fold_builtin_cpu): Add slm.
15877 * config/i386/i386.h (TARGET_SLM): New target macro.
15878 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
15879 (processor_type): Add PROCESSOR_SLM.
15881 * config/i386/i386.md (cpu): Add new value "slm".
15882 (slm.md): Include slm.md.
15884 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
15885 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15887 * config/arm/arm-protos.h: Add and update function protos.
15888 * config/arm/arm.c (use_simple_return_p): New added.
15889 (thumb2_expand_return): Check simple_return flag.
15890 * config/arm/arm.md: Add simple_return and conditional simple_return.
15891 * config/arm/iterators.md: Add iterator for return and simple_return.
15893 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15895 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
15896 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
15897 (arm_emit_vfp_multi_reg_pop): Likewise.
15898 (thumb2_emit_ldrd_pop): Likewise.
15899 (arm_expand_epilogue): Add misc REG_CFA notes.
15900 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
15902 2013-05-29 Lawrence Crowl <crowl@google.com>
15904 * config/arm/t-arm: Update for below.
15906 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
15907 Change type to hash_table. Update dependent calls and types.
15909 * config/i386/t-cygming: Update for below.
15911 * config/i386/t-interix: Update for below.
15913 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
15914 Change type to hash_table. Update dependent calls and types.
15915 (i386_find_on_wrapper_list::wrappers): Likewise.
15917 * config/ia64/t-ia64: Update for below.
15919 * config/ia64/ia64.c (bundle_state_table):
15920 Change type to hash_table. Update dependent calls and types.
15922 * config/mips/mips.c (mips_reorg_process_insns::htab):
15923 Change type to hash_table. Update dependent calls and types.
15925 * config/sol2.c (solaris_comdat_htab):
15926 Change type to hash_table. Update dependent calls and types.
15928 * config/t-sol2: Update for above.
15930 2013-05-29 Teresa Johnson <tejohnson@google.com>
15932 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
15933 functions are not yet marked as defined.
15935 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
15936 Pat Haugen <pthaugen@us.ibm.com>
15937 Peter Bergner <bergner@vnet.ibm.com>
15939 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
15943 (vrotl<mode>3): Likewise.
15944 (vashl<mode>3): Likewise.
15945 (vlshr<mode>3): Likewise.
15946 (vashr<mode>3): Likewise.
15948 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15949 support for power8 V2DI builtins.
15951 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
15952 power8 V2DI builtins.
15953 (vupkhsw): Likewise.
15954 (vupklsw): Likewise.
15955 (vaddudm): Likewise.
15956 (vminsd): Likewise.
15957 (vmaxsd): Likewise.
15958 (vminud): Likewise.
15959 (vmaxud): Likewise.
15960 (vpkudum): Likewise.
15961 (vpksdss): Likewise.
15962 (vpkudus): Likewise.
15963 (vpksdus): Likewise.
15968 (vsubudm): Likewise.
15969 (vcmpequd): Likewise.
15970 (vcmpgtsd): Likewise.
15971 (vcmpgtud): Likewise.
15972 (vcmpequd_p): Likewise.
15973 (vcmpgtsd_p): Likewise.
15974 (vcmpgtud_p): Likewise.
15975 (vupkhsw): Likewise.
15976 (vupklsw): Likewise.
15977 (vaddudm): Likewise.
15978 (vmaxsd): Likewise.
15979 (vmaxud): Likewise.
15980 (vminsd): Likewise.
15981 (vminud): Likewise.
15982 (vpksdss): Likewise.
15983 (vpksdus): Likewise.
15984 (vpkudum): Likewise.
15985 (vpkudus): Likewise.
15990 (vsubudm): Likewise.
15992 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
15993 support for power8 V2DI instructions.
15995 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
15996 power8 V2DI instructions. Combine pack and unpack insns to use an
15997 iterator for each mode. Check whether a particular mode supports
15998 Altivec instructions instead of just checking TARGET_ALTIVEC.
15999 (UNSPEC_VPKUWUM): Likewise.
16000 (UNSPEC_VPKSHSS): Likewise.
16001 (UNSPEC_VPKSWSS): Likewise.
16002 (UNSPEC_VPKUHUS): Likewise.
16003 (UNSPEC_VPKSHUS): Likewise.
16004 (UNSPEC_VPKUWUS): Likewise.
16005 (UNSPEC_VPKSWUS): Likewise.
16006 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
16007 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
16008 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
16009 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
16010 (UNSPEC_VUPKHSB): Likewise.
16011 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
16012 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
16013 (UNSPEC_VUPKHSH): Likewise.
16014 (UNSPEC_VUPKLSB): Likewise.
16015 (UNSPEC_VUPKLSH): Likewise.
16017 (VI_char): Likewise.
16018 (VI_scalar): Likewise.
16019 (VI_unit): Likewise.
16021 (VP_small): Likewise.
16022 (VP_small_lc): Likewise.
16023 (VU_char): Likewise.
16024 (add<mode>3): Likewise.
16025 (altivec_vaddcuw): Likewise.
16026 (altivec_vaddu<VI_char>s): Likewise.
16027 (altivec_vadds<VI_char>s): Likewise.
16028 (sub<mode>3): Likewise.
16029 (altivec_vsubcuw): Likewise.
16030 (altivec_vsubu<VI_char>s): Likewise.
16031 (altivec_vsubs<VI_char>s): Likewise.
16032 (altivec_vavgs<VI_char>): Likewise.
16033 (altivec_vcmpbfp): Likewise.
16034 (altivec_eq<mode>): Likewise.
16035 (altivec_gt<mode>): Likewise.
16036 (altivec_gtu<mode>): Likewise.
16037 (umax<mode>3): Likewise.
16038 (smax<mode>3): Likewise.
16039 (umin<mode>3): Likewise.
16040 (smin<mode>3): Likewise.
16041 (altivec_vpkuhum): Likewise.
16042 (altivec_vpkuwum): Likewise.
16043 (altivec_vpkshss): Likewise.
16044 (altivec_vpkswss): Likewise.
16045 (altivec_vpkuhus): Likewise.
16046 (altivec_vpkshus): Likewise.
16047 (altivec_vpkuwus): Likewise.
16048 (altivec_vpkswus): Likewise.
16049 (altivec_vpks<VI_char>ss): Likewise.
16050 (altivec_vpks<VI_char>us): Likewise.
16051 (altivec_vpku<VI_char>us): Likewise.
16052 (altivec_vpku<VI_char>um): Likewise.
16053 (altivec_vrl<VI_char>): Likewise.
16054 (altivec_vsl<VI_char>): Likewise.
16055 (altivec_vsr<VI_char>): Likewise.
16056 (altivec_vsra<VI_char>): Likewise.
16057 (altivec_vsldoi_<mode>): Likewise.
16058 (altivec_vupkhsb): Likewise.
16059 (altivec_vupkhs<VU_char>): Likewise.
16060 (altivec_vupkls<VU_char>): Likewise.
16061 (altivec_vupkhsh): Likewise.
16062 (altivec_vupklsb): Likewise.
16063 (altivec_vupklsh): Likewise.
16064 (altivec_vcmpequ<VI_char>_p): Likewise.
16065 (altivec_vcmpgts<VI_char>_p): Likewise.
16066 (altivec_vcmpgtu<VI_char>_p): Likewise.
16067 (abs<mode>2): Likewise.
16068 (vec_unpacks_hi_v16qi): Likewise.
16069 (vec_unpacks_hi_v8hi): Likewise.
16070 (vec_unpacks_lo_v16qi): Likewise.
16071 (vec_unpacks_hi_<VP_small_lc>): Likewise.
16072 (vec_unpacks_lo_v8hi): Likewise.
16073 (vec_unpacks_lo_<VP_small_lc>): Likewise.
16074 (vec_pack_trunc_v8h): Likewise.
16075 (vec_pack_trunc_v4si): Likewise.
16076 (vec_pack_trunc_<mode>): Likewise.
16078 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
16080 (vec_vmaxsd): Likewise.
16081 (vec_vmaxud): Likewise.
16082 (vec_vminsd): Likewise.
16083 (vec_vminud): Likewise.
16084 (vec_vpksdss): Likewise.
16085 (vec_vpksdus): Likewise.
16086 (vec_vpkudum): Likewise.
16087 (vec_vpkudus): Likewise.
16088 (vec_vrld): Likewise.
16089 (vec_vsld): Likewise.
16090 (vec_vsrad): Likewise.
16091 (vec_vsrd): Likewise.
16092 (vec_vsubudm): Likewise.
16093 (vec_vupkhsw): Likewise.
16094 (vec_vupklsw): Likewise.
16096 2013-05-29 Jan Hubicka <jh@suse.cz>
16098 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
16099 flags; reorder rest of fields in more consistent way.
16100 (varpool_node): Remove analyzed, finalized and alias.
16101 (cgraph_ndoe): Likewise.
16102 (symtab_alias_ultimate_target): New function.
16103 (cgraph_function_node): Move offline.
16104 (cgraph_reset_node): Declare.
16105 (cgraph_comdat_can_be_unshared_p): Remove.
16106 (varpool_remove_initializer): Declare.
16107 (varpool_first_defined_variable, varpool_next_defined_variable
16108 cgraph_first_defined_function, cgraph_next_defined_function): Update.
16109 (cgraph_function_with_gimple_body_p): Update.
16110 (varpool_all_refs_explicit_p): Update.
16111 (symtab_alias_target): New function.
16112 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
16113 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
16114 (cgraph_function_or_thunk_node): Simplify using
16115 symtab_alias_ultimate_target.
16116 (varpool_variable_node): Likewise.
16117 * cgraph.c (cgraph_create_function_alias): Update.
16118 (cgraph_add_thunk): Update.
16119 (cgraph_remove_node): Update.
16120 (dump_cgraph_node): Do not dump removed flags.
16121 (cgraph_function_body_availability): Update.
16122 (cgraph_propagate_frequency): Update.
16123 (verify_cgraph_node): Check sanity of local flag.
16124 (cgraph_function_node): Move here from cgraph.h; revamp for
16125 cgraph_function_or_thunk_node.
16126 * lto-symtab.c (lto_varpool_replace_node): Update.
16127 (lto_symtab_resolve_can_prevail_p): Update.
16128 (lto_symtab_merge_cgraph_nodes): Update.
16129 * ipa-cp.c (determine_versionability, initialize_node_lattices,
16130 propagate_constants_accross_call, devirtualization_time_bonus,
16131 ipcp_propagate_stage): Update.
16132 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
16133 * ipa-inline-transform.c (clone_inlined_nodes,
16134 preserve_function_body_p): Update.
16135 * ipa-reference.c (propagate): Update.
16136 (write_node_summary_p): Update.
16137 * toplev.c (wrapup_global_declaration_2): Update.
16138 * cgraphunit.c (cgraph_analyze_function): Rename to ...
16139 (analyze_function) ... this one.
16140 (cgraph_process_new_functions): Update.
16141 (cgraph_reset_node): Export.
16142 (cgraph_finalize_function): Update.
16143 (cgraph_add_new_function): Update.
16144 (process_function_and_variable_attributes): Update.
16145 (varpool_finalize_decl): Update.
16146 (symbol_finalized): Remove.
16147 (symbol_finalized_and_needed): Rename to ...
16148 (symbol_defined_and_needed): ... update.
16149 (cgraph_analyze_functions): Update.
16150 (handle_alias_pairs): Update.
16151 (mark_functions_to_output): Update.
16152 (assemble_thunk): Update.
16153 (output_in_order): Update.
16154 (output_weakrefs): Update.
16155 (finalize_compilation_unit): Update.
16156 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
16157 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
16158 input_node, input_varpool_node): Update.
16159 * dbxout.c (dbxout_expand_expr): Update.
16160 * cgraphclones.c (cgraph_clone_node): Update.
16161 (cgraph_copy_node_for_versioning): Update.
16162 (cgraph_materialize_clone): Update.
16163 (cgraph_materialize_all_clones): Update.
16164 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
16165 propagate_pure_const, propagate_nothrow): Update.
16166 * lto-streamer-out.c (lto_output, write_symbol): Update.
16167 * ipa-utils.c (ipa_reverse_postorder): Update.
16168 * ipa-inline.c (can_inline_edge_p): Update.
16169 (update_caller_keys, ipa_inline): Update.
16170 * dwarf2out.c (reference_to_unused,
16171 premark_types_used_by_global_vars_helper): Update.
16172 * tree-eh.c (tree_could_trap_p): Update.
16173 * ipa-split.c (consider_split, execute_split_functions): Update.
16174 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
16175 has_addr_references_p): Update; move ahead in file for better
16177 (process_references): Simplify.
16178 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
16179 bodies are removed.
16180 (cgraph_comdat_can_be_unshared_p): Make static.
16181 (cgraph_externally_visible_p): Update.
16182 (varpool_externally_visible_p): Update.
16183 (function_and_variable_visibility): Update.
16184 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
16185 ipa_tm_mark_force_output_node): Update.
16186 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
16187 estimate_edge_devirt_benefit, inline_generate_summary,
16188 inline_write_summary): Update.
16189 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
16190 * ipa-prop.c (ipa_compute_jump_functions): Update.
16191 (ipa_print_node_params, ipa_prop_read_section,
16192 ipa_update_after_lto_read, read_replacements_section): Update.
16193 * varasm.c (mark_decl_referenced): Update.
16194 (assemble_alias, dump_tm_clone_pairs): Update.
16195 * tree-inline.c (copy_bb): Update.
16196 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
16198 * symtab.c (dump_symtab_base): Print new flags.
16199 (verify_symtab_base): Verify new flags.
16200 (symtab_alias_ultimate_target): New function.
16201 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
16202 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
16204 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
16206 * i386.c (ix86_get_function_versions_dispatcher,
16207 ix86_generate_version_dispatcher_body): Update.
16208 (fold_builtin_cpu): Use varpool_add_new_variable.
16209 * varpool.c (varpool_remove_initializer): Break out from ...
16210 (varpool_remove_node): ... this one.
16211 (dump_varpool_node, varpool_node_for_asm,
16212 cgraph_variable_initializer_availability, varpool_analyze_node,
16213 varpool_assemble_decl, varpool_remove_unreferenced_decls,
16214 varpool_finalize_named_section_flags, varpool_create_variable_alias):
16217 2013-05-29 Jan Hubicka <jh@suse.cz>
16219 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
16221 2013-05-29 Easwaran Raman <eraman@google.com>
16223 PR tree-optimization/57442
16224 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
16225 when control exits the main loop.
16227 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
16229 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
16231 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
16232 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
16233 * rx/t-rx: Add rx100 under multi library matches option for nofpu
16236 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16238 PR tree-optimization/57441
16239 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
16240 Don't limit size of incr_vec to number of candidates.
16242 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
16244 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
16245 and mips16 directories.
16246 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
16247 (MULTILIB_DIRNAMES): Ditto.
16248 (MULTILIB_EXCEPTIONS): Add new exceptions.
16249 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
16250 (MULTILIB_DIRNAMES): Ditto.
16251 (MULTILIB_EXCEPTIONS): Add new exceptions.
16253 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16254 Marcus Shawcroft <marcus.shawcroft@arm.com>
16256 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
16257 SYMBOL_TINY_ABSOLUTE.
16258 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
16259 SYMBOL_TINY_ABSOLUTE.
16260 (aarch64_expand_mov_immediate): Likewise.
16261 (aarch64_classify_symbol): Likewise.
16262 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
16263 Permit SYMBOL_TINY_ABSOLUTE.
16264 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
16266 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16267 Marcus Shawcroft <marcus.shawcroft@arm.com>
16269 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
16270 Refactor if/switch. Replace gcc_assert with if.
16272 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16274 * config/i386/i386.c (initial_ix86_tune_features): Enable
16275 FP Reassociation for AMD bdver1 and bdver2.
16277 2013-05-29 Martin Jambor <mjambor@suse.cz>
16279 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
16280 and IMAGPART_EXPR do not occur within other handled_components.
16282 2013-05-29 Richard Biener <rguenther@suse.de>
16284 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
16285 access on whether the use is in the BB we currently try to
16287 (vect_bb_vectorization_profitable_p): Pass the BB we currently
16288 vectorize to vect_bb_slp_scalar_cost.
16290 2013-05-29 Richard Biener <rguenther@suse.de>
16292 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
16293 computing scalar cost offsetted by stmts that are kept live
16295 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
16296 for computation of scalar cost.
16298 2013-05-28 Steve Ellcey <sellcey@mips.com>
16300 * config/mips/mips-cpus.def (mips32r2): Change processor type.
16302 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
16304 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
16305 array notation built-in reduction functions.
16306 * doc/passes.texi (Passes): Added documentation about changes done
16308 * doc/invoke.texi (C Dialect Options): Added documentation about
16309 the -fcilkplus flag.
16310 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
16311 (BUILTINS_DEF): Depend on cilkplus.def.
16312 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
16313 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
16314 * cilkplus.def: New file.
16316 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
16318 PR rtl-optimization/57439
16319 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
16321 2013-05-28 Easwaran Raman <eraman@google.com>
16323 PR tree-optimization/57337
16324 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
16325 (find_insert_point): Correctly identify the insertion point
16326 when two statements with the same UID is compared.
16328 2013-05-28 Richard Biener <rguenther@suse.de>
16330 PR tree-optimization/56787
16331 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
16332 from the list of data references.
16333 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
16335 (vect_analyze_loop_operations): Likewise.
16336 (vect_transform_loop): Remove clobbers.
16338 2013-05-28 Martin Jambor <mjambor@suse.cz>
16340 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
16341 and REALPART_EXPRs have scalar type.
16343 2013-05-28 Richard Biener <rguenther@suse.de>
16345 PR tree-optimization/57411
16346 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
16348 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
16349 virtual operand propagation.
16351 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16353 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
16354 destination register for bmasksi_vis.
16355 (vector_init_bshuffle): Likewise.
16356 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
16358 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16360 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
16361 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
16362 mode if the instruction isn't available in the original mode.
16363 * config/sparc/sparc.opt (mfix-ut699): New option.
16364 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
16365 (divdf3): Turn into expander.
16366 (divdf3_nofix): New insn.
16367 (divdf3_fix): Likewise.
16368 (divsf3): Disable if -mfix-ut699.
16369 (sqrtdf2): Turn into expander.
16370 (sqrtdf2_nofix): New insn.
16371 (sqrtdf2_fix): Likewise.
16372 (sqrtsf2): Disable if -mfix-ut699.
16374 2013-05-27 Richard Biener <rguenther@suse.de>
16376 PR middle-end/57412
16377 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
16378 block for the new loop.
16380 2013-05-27 Richard Biener <rguenther@suse.de>
16382 PR tree-optimization/57343
16383 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
16384 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
16385 (number_of_iterations_cond): Do not build the folded tree.
16387 2013-05-27 Richard Biener <rguenther@suse.de>
16390 PR middle-end/57381
16391 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16392 OEP_CONSTANT_ADDRESS_OF retained.
16394 PR tree-optimization/57417
16395 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
16396 for unchanged base.
16397 (set_ssa_val_to): Compare addresses using
16398 get_addr_base_and_unit_offset.
16400 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
16402 PR rtl-optimization/56833
16403 * postreload.c (move2add_record_mode): New function.
16404 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
16405 (move2add_use_add2_insn): Use move2add_record_sym_value.
16406 (move2add_use_add3_insn): Likewise.
16407 (reload_cse_move2add): Use move2add_valid_value_p and
16408 move2add_record_mode. Invalidate call-clobbered and REG_INC
16409 affected regs by setting reg_mode to VOIDmode.
16410 (move2add_note_store): Don't pretend the inside of a SUBREG is
16411 the actual destination. Invalidate single/leading registers by
16412 setting reg_mode to VOIDmode.
16413 Use move2add_record_sym_value, move2add_valid_value_p and
16414 move2add_record_mode.
16416 2013-05-27 Richard Biener <rguenther@suse.de>
16418 PR tree-optimization/57396
16419 * tree-affine.c (double_int_constant_multiple_p): Properly
16420 return false for val == 0 and div != 0.
16422 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16424 * config/mips/mips.h: Use #elif in preprocessor conditions.
16426 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16429 * config/mips/constraints.md (kl): New constraint.
16430 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
16431 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
16432 constraint for operand 0. Split after CSE for MIPS16. Emit a move
16433 from LO for MIPS16.
16434 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
16436 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16439 * config/mips/mips.c (mips_can_inline_p): New function.
16440 (TARGET_CAN_INLINE_P): Define.
16442 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
16444 * sched-int.h (ds_t, dw_t): Make unsigned int.
16445 Fix documentation that describes how all the ds_t bits are used.
16446 Reserve the last bit for delayed-branch scheduling.
16447 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
16448 (BITS_PER_DEP_WEAK): Fix definition and documentation.
16449 (gen_dep_weak_1): Remove prototype.
16450 * sched-deps.c (get_dep_weak_1): Make static.
16451 * target.def (speculate_insn, needs_block_p, gen_spec_check,
16452 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
16453 * doc/tm.texi: Regenerate.
16454 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
16456 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
16459 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
16461 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
16462 Sandra Loosemore <sandra@codesourcery.com>
16464 * config.gcc (powerpc-*): Allow native for with-cpu.
16466 2013-05-24 Jeff Law <law@redhat.com>
16468 PR tree-optimization/57124
16469 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
16470 conversion feeding a condition if the range has an overflow
16471 if -fstrict-overflow. Add warnings for when we do make the
16474 2013-05-24 Dehao Chen <dehao@google.com>
16476 * tree-cfg.c (locus_discrim_map): Fix the typo.
16477 (locus_discrim_hasher): Likewise.
16478 (locus_discrim_hasher::hash): Likewise.
16479 (locus_discrim_hasher::equal): Likewise.
16481 2013-05-24 Martin Jambor <mjambor@suse.cz>
16483 PR tree-optimization/57294
16484 * cgraph.h (ipa_record_stmt_references): Declare.
16485 * cgraphbuild.c (ipa_record_stmt_references): New function.
16486 (build_cgraph_edges): Use ipa_record_stmt_references.
16487 (rebuild_cgraph_edges): Likewise.
16488 (cgraph_rebuild_references): Likewise.
16489 * ipa-prop.c (ipa_modify_call_arguments): Discard references
16490 associated with the old statement and build references from the
16491 newly built statements.
16492 * ipa-ref.c (ipa_remove_stmt_references): New function.
16493 * ipa-ref.h (ipa_remove_stmt_references): Declare.
16495 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
16497 * lra-constraints.c (emit_spill_move): Use smaller mode for
16499 (check_and_process_move): Consider mem-reg moves for secondary
16501 (curr_insn_transform): Don't lose insns emitted before for
16502 secondary memory moves.
16503 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
16504 reg set up in the current insn.
16506 2013-05-24 Dehao Chen <dehao@google.com>
16508 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
16510 (locus_descrim_hasher::equal): Likewise.
16511 (build_gimple_cfg): New discriminator assignment algorithm.
16512 (make_edges): Likewise.
16513 (next_discriminator_for_locus): Likewise.
16514 (same_line_p): Likewise.
16515 (assign_discriminators): Likewise.
16516 (make_cond_expr_edges): Likewise.
16517 (make_gimple_switch_edges): Likewise.
16518 (make_goto_expr_edges): Likewise.
16519 (make_gimple_asm_edges): Likewise.
16521 2013-05-24 Ian Bolton <ian.bolton@arm.com>
16523 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
16524 X format specifier to only display bottom 16 bits.
16525 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
16526 immediate to match for operand 2, since it will be masked.
16528 2013-05-24 Richard Biener <rguenther@suse.de>
16530 PR tree-optimization/57287
16531 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
16532 all SSA names that occur in abnormal PHIs.
16534 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
16537 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
16538 that index is not negative.
16540 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
16542 PR rtl-optimization/55177
16543 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
16544 (simplify_byte_swapping_operation): New.
16545 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
16546 (simplify_relational_operation_1): Deal with BSWAP.
16548 2013-05-23 Richard Henderson <rth@redhat.com>
16551 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
16552 (ix86_reorg): Call it.
16554 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
16557 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
16558 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
16559 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
16561 2013-05-23 Christian Bruel <christian.bruel@st.com>
16564 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
16566 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16567 Marcus Shawcroft <marcus.shawcroft@arm.com>
16569 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
16570 * config/aarch64/constraints.md (Usa): Remove.
16571 * doc/md.texi (AArch64 Usa): Remove.
16573 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16574 Marcus Shawcroft <marcus.shawcroft@arm.com>
16576 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
16577 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
16578 * config/aarch64/predicates.md (aarch64_const_address): Remove.
16579 (aarch64_mov_operand): Use aarch64_mov_operand_p.
16581 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
16583 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
16584 instruction (AdvSIMD).
16585 * config/aarch64/aarch64-builtins.c
16586 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
16587 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
16589 2013-05-23 Martin Jambor <mjambor@suse.cz>
16591 PR middle-end/57347
16592 * tree.h (contains_bitfld_component_ref_p): Declare.
16593 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
16594 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
16596 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
16597 not access a bit-field. Assert all final offsets are byte-aligned.
16599 2013-05-23 Richard Biener <rguenther@suse.de>
16601 PR tree-optimization/57380
16602 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
16603 least one invariant or re-used load.
16604 * passes.c (init_optimization_passes): Move pass_phiprop before
16607 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
16609 * config/aarch64/aarch64-simd.md
16610 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
16612 2013-05-23 Richard Biener <rguenther@suse.de>
16614 PR middle-end/57381
16615 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16616 OEP_CONSTANT_ADDRESS_OF retained.
16618 2013-05-23 Jakub Jelinek <jakub@redhat.com>
16620 PR middle-end/57344
16621 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
16622 don't lower unit. Handle unit not being always BITS_PER_WORD.
16624 2013-05-23 Richard Biener <rguenther@suse.de>
16626 PR rtl-optimization/57341
16627 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
16628 instead of true_dependence.
16630 2013-05-22 David Malcolm <dmalcolm@redhat.com>
16632 * bb-reorder.c (branch_threshold): Make const.
16633 (exec_threshold): Ditto.
16635 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16636 Pat Haugen <pthaugen@us.ibm.com>
16637 Peter Bergner <bergner@vnet.ibm.com>
16639 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
16640 documentation for the power8 crypto builtins.
16642 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
16644 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
16645 macros for defining power8 builtin functions.
16646 (BU_P8V_AV_2): Likewise.
16647 (BU_P8V_AV_P): Likewise.
16648 (BU_P8V_VSX_1): Likewise.
16649 (BU_P8V_OVERLOAD_1): Likewise.
16650 (BU_P8V_OVERLOAD_2): Likewise.
16651 (BU_CRYPTO_1): Likewise.
16652 (BU_CRYPTO_2): Likewise.
16653 (BU_CRYPTO_3): Likewise.
16654 (BU_CRYPTO_OVERLOAD_1): Likewise.
16655 (BU_CRYPTO_OVERLOAD_2): Likewise.
16656 (XSCVSPDP): Fix typo, point to the correct instruction.
16657 (VCIPHER): Add power8 crypto builtins.
16658 (VCIPHERLAST): Likewise.
16659 (VNCIPHER): Likewise.
16660 (VNCIPHERLAST): Likewise.
16661 (VPMSUMB): Likewise.
16662 (VPMSUMH): Likewise.
16663 (VPMSUMW): Likewise.
16664 (VPERMXOR_V2DI): Likewise.
16665 (VPERMXOR_V4SI: Likewise.
16666 (VPERMXOR_V8HI: Likewise.
16667 (VPERMXOR_V16QI: Likewise.
16668 (VSHASIGMAW): Likewise.
16669 (VSHASIGMAD): Likewise.
16670 (VPMSUM): Likewise.
16671 (VPERMXOR): Likewise.
16672 (VSHASIGMA): Likewise.
16674 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
16675 __CRYPTO__ if the crypto instructions are available.
16676 (altivec_overloaded_builtins): Add support for overloaded power8
16679 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
16680 support for power8 crypto builtins.
16681 (builtin_function_type): Likewise.
16682 (altivec_init_builtins): Add support for builtins that take vector
16683 long long (V2DI) arguments.
16685 * config/rs6000/crypto.md: New file, define power8 crypto
16688 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16689 Pat Haugen <pthaugen@us.ibm.com>
16690 Peter Bergner <bergner@vnet.ibm.com>
16692 * doc/invoke.texi (Option Summary): Add power8 options.
16693 (RS/6000 and PowerPC Options): Likewise.
16695 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
16696 constraints.md instead of rs6000.h. Reorder w* constraints. Add
16697 wm, wn, wr documentation.
16699 * config/rs6000/constraints.md (wm): New constraint for VSX
16700 registers if direct move instructions are enabled.
16701 (wn): New constraint for no registers.
16702 (wq): New constraint for quad word even GPR registers.
16703 (wr): New constraint if 64-bit instructions are enabled.
16704 (wv): New constraint if power8 vector instructions are enabled.
16705 (wQ): New constraint for quad word memory locations.
16707 * config/rs6000/predicates.md (const_0_to_15_operand): New
16708 constraint for 0..15 for crypto instructions.
16709 (gpc_reg_operand): If VSX allow registers in VSX registers as well
16710 as GPR and floating point registers.
16711 (int_reg_operand): New predicate to match only GPR registers.
16712 (base_reg_operand): New predicate to match base registers.
16713 (quad_int_reg_operand): New predicate to match even GPR registers
16714 for quad memory operations.
16715 (vsx_reg_or_cint_operand): New predicate to allow vector logical
16716 operations in both GPR and VSX registers.
16717 (quad_memory_operand): New predicate for quad memory operations.
16718 (reg_or_indexed_operand): New predicate for direct move support.
16720 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
16721 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
16722 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
16723 (POWERPC_MASKS): Add power8 options.
16724 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
16727 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
16728 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
16730 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
16731 (-mpower8-fusion): New power8 options.
16732 (-mpower8-fusion-sign): Likewise.
16733 (-mpower8-vector): Likewise.
16734 (-mcrypto): Likewise.
16735 (-mdirect-move): Likewise.
16736 (-mquad-memory): Likewise.
16738 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
16739 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
16741 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
16742 (rs6000_debug_vector_unit): Add p8_vector.
16743 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
16744 definitions. Also print fusion state.
16745 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
16746 (rs6000_builtin_mask_calculate): Add power8 builtin support.
16747 (rs6000_option_override_internal): Add support for power8.
16748 (rs6000_common_init_builtins): Add debugging for skipped builtins
16749 if -mdebug=builtin.
16750 (rs6000_adjust_cost): Add power8 support.
16751 (rs6000_issue_rate): Likewise.
16752 (insn_must_be_first_in_group): Likewise.
16753 (insn_must_be_last_in_group): Likewise.
16754 (force_new_group): Likewise.
16755 (rs6000_register_move_cost): Likewise.
16756 (rs6000_opt_masks): Likewise.
16758 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
16759 power8 capable assembler, default to power7 options.
16760 (TARGET_DIRECT_MOVE): Likewise.
16761 (TARGET_CRYPTO): Likewise.
16762 (TARGET_P8_VECTOR): Likewise.
16763 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
16764 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
16765 (VECTOR_MEM_P8_VECTOR_P): Likewise.
16766 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
16767 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
16768 (TARGET_XSCVDPSPN): Likewise.
16769 (TARGET_XSCVSPDPN): Likewsie.
16770 (TARGET_SYNC_HI_QI): Likewise.
16771 (TARGET_SYNC_TI): Likewise.
16772 (MASK_CRYPTO): Likewise.
16773 (MASK_DIRECT_MOVE): Likewise.
16774 (MASK_P8_FUSION): Likewise.
16775 (MASK_P8_VECTOR): Likewise.
16776 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
16777 temporary used by some of the direct move instructions to get two FP
16778 temporary registers does not force creation of a stack frame.
16779 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
16780 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
16781 that any VSX registers are tieable, even if they are also an
16782 Altivec vector mode.
16783 (r6000_reg_class_enum): Add wm, wr, wv constraints.
16784 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
16785 (RS6000_BTM_CRYPTO): Likewise.
16786 (RS6000_BTM_COMMON): Likewise.
16788 * config/rs6000/rs6000.md (cpu attribute): Add power8.
16789 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
16790 (enum rs6000_vector): Add power8 vector support.
16792 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16796 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
16797 (any_sibcall_could_use_r3): this and handle indirect calls.
16798 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
16800 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16802 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
16804 2013-05-22 Richard Biener <rguenther@suse.de>
16806 PR middle-end/57349
16807 * profile.c (branch_prob): Do not split blocks that are
16808 abnormally receiving from ECF_RETURNS_TWICE functions.
16810 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
16812 * recog.c (offsettable_address_addr_space_p): Fix calculation of
16813 address mode. Move pointer mode initialization to the same place.
16815 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
16817 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
16818 while it has any effect.
16820 2013-05-21 Easwaran Raman <eraman@google.com>
16822 PR tree-optimization/57322
16823 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
16824 UID of the statement added to the BB to be 1.
16826 2013-05-21 Jakub Jelinek <jakub@redhat.com>
16828 PR tree-optimization/57331
16829 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
16830 of conversion from pointer type to integral type with integer.
16832 2013-05-21 Martin Jambor <mjambor@suse.cz>
16835 * ipa-prop.c (ipa_read_node_info): Process param_used and
16836 controlled_uses in the same order as when writing.
16838 2013-05-21 Magnus Granberg <baldrick@free.fr>
16841 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
16843 2013-05-21 Richard Biener <rguenther@suse.de>
16845 PR tree-optimization/57318
16846 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
16847 estimate stmts with side-effects as likely eliminated.
16849 2013-05-21 Richard Biener <rguenther@suse.de>
16851 PR tree-optimization/57330
16852 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
16853 preserve the call stmts fntype.
16855 2013-05-21 Richard Biener <rguenther@suse.de>
16857 PR tree-optimization/57303
16858 * tree-ssa-sink.c (statement_sink_location): Improve killing
16859 stmt detection and properly handle self-assignments.
16861 2013-05-21 Christian Bruel <christian.bruel@st.com>
16863 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
16864 spanning registers. LEAF_REG_REMAP is supported only for contiguous
16865 registers. Set register size out of the PARALLEL loop.
16867 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
16870 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
16871 (*fmasf4, *fmasf4_media): New insns.
16873 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16875 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
16876 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
16877 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
16878 (mips_idiv_insns): Update the comments to say that the returned
16879 instruction counts are in units of BASE_INSN_LENGTH.
16880 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
16881 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
16882 using 2 rather than 4 as the length of indirect MIPS16 and
16883 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
16884 length of a NOP. Don't divide MIPS16 lengths by 2.
16885 (mips16_split_long_branches): Assume a branch is long if the
16886 length is greater than 4 rather than 8.
16887 * config/mips/mips.md (length): Give MIPS16 lengths directly,
16888 rather than multiplying them by 2. Multiply instruction counts
16889 by BASE_INSN_LENGTH rather than 4.
16890 (*jump_mips16, tls_get_tp_mips16_<mode>)
16891 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
16893 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16895 * config/mips/mips.md (extended_mips16): Remove branch case.
16896 (length): Remove duplicated extended_mips16 test.
16898 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16900 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
16902 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
16904 * recog.h (Recog_data): Rename to...
16905 (recog_data_d): ...this.
16906 (recog_data): Update accordingly.
16907 * recog.c (recog_data): Likewise.
16908 * reload.c (save_recog_data): Likewise.
16909 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
16910 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
16912 2013-05-17 Julian Brown <julian@codesourcery.com>
16914 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
16915 found in a REG_EQUAL note, invalidate it.
16917 2013-05-17 Easwaran Raman <eraman@google.com>
16919 * tree-ssa-reassoc.c (find_insert_point): New function.
16920 (insert_stmt_after): Likewise.
16921 (get_def_stmt): Likewise.
16922 (ensure_ops_are_available): Likewise.
16923 (not_dominated_by): Likewise.
16924 (rewrite_expr_tree): Do not move statements beyond what is
16925 necessary. Remove call to swap_ops_for_binary_stmt...
16926 (reassociate_bb): ... and move it here.
16927 (build_and_add_sum): Assign UIDs for new statements.
16928 (linearize_expr): Likewise.
16929 (do_reassoc): Renumber gimple statement UIDs.
16931 2013-05-17 Jan Hubicka <jh@suse.cz>
16933 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
16935 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
16936 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
16938 (output_weakrefs): Update.
16940 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
16941 Martin Jambor <mjambor@suse.cz>
16943 PR middle-end/57276
16944 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
16945 value that corresponds to the given aggval is found in values vector.
16947 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
16949 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
16950 sse, sse2, sse3, ssse3 and sse4a flags to options.
16952 2013-05-17 David Malcolm <dmalcolm@redhat.com>
16954 * gengtype-state.c: (s_expr_writer): New class, to handle
16955 prettifying of output layout of s-expressions.
16956 (state_writer): New class, to write out gtype.state.
16957 (state_written_type_count): Move this variable into member data of
16959 (s_expr_writer::s_expr_writer): New code: constructor for new class
16960 (state_writer::state_writer(): ditto
16961 (s_expr_writer::write_new_line): New function
16962 (s_expr_writer::write_any_indent): ditto
16963 (s_expr_writer::begin_s_expr): ditto
16964 (s_expr_writer::end_s_expr): ditto
16965 (write_state_fileloc): convert to method of state_writer...
16966 (state_writer:: write_state_fileloc): ...and use methods of
16967 s_expr_writer to write indentation into the gtype.state output file
16968 to visually represent the hierarchical structure of the list
16970 (write_state_fields): ditto, renaming to...
16971 (state_writer::write_state_fields)
16972 (write_state_a_string): ditto, renaming to...
16973 (state_writer::write_state_a_string)
16974 (write_state_string_option): ditto, renaming to...
16975 (state_writer::write_state_string_option)
16976 (write_state_type_option): ditto, renaming to...
16977 (state_writer::write_state_type_option)
16978 (write_state_nested_option): ditto, renaming to...
16979 (state_writer::write_state_nested_option)
16980 (write_state_option): ditto, renaming to...
16981 (state_writer::write_state_option)
16982 (write_state_options): ditto, renaming to...
16983 (state_writer::write_state_options)
16984 (write_state_lang_bitmap): ditto, renaming to...
16985 (state_writer::write_state_lang_bitmap)
16986 (write_state_version): ditto, renaming to...
16987 (state_writer::write_state_version)
16988 (write_state_scalar_type): ditto, renaming to...
16989 (state_writer::write_state_scalar_type)
16990 (write_state_string_type): ditto, renaming to...
16991 (state_writer::write_state_string_type)
16992 (write_state_undefined_type): ditto, renaming to...
16993 (state_writer::write_state_undefined_type)
16994 (write_state_struct_union_type): ditto, renaming to...
16995 (state_writer::write_state_struct_union_type)
16996 (write_state_struct_type): ditto, renaming to...
16997 (state_writer::write_state_struct_type)
16998 (write_state_user_struct_type): ditto, renaming to...
16999 (state_writer::write_state_user_struct_type)
17000 (write_state_lang_struct_type): ditto, renaming to...
17001 (state_writer::write_state_lang_struct_type)
17002 (write_state_param_struct_type): ditto, renaming to...
17003 (state_writer::write_state_param_struct_type)
17004 (write_state_pointer_type): ditto, renaming to...
17005 (state_writer::write_state_pointer_type)
17006 (write_state_array_type): ditto, renaming to...
17007 (state_writer::write_state_array_type)
17008 (write_state_gc_used): ditto, renaming to...
17009 (state_writer::write_state_gc_used)
17010 (write_state_common_type_content): ditto, renaming to...
17011 (state_writer::write_state_common_type_content)
17012 (write_state_type): ditto, renaming to...
17013 (state_writer::write_state_type)
17014 (write_state_pair_list): ditto, renaming to...
17015 (state_writer::write_state_pair_list)
17016 (write_state_pair): ditto, renaming to...
17017 (state_writer::write_state_pair)
17018 (write_state_typedefs): ditto, renaming to...
17019 (state_writer::write_state_typedefs)
17020 (write_state_structures): ditto, renaming to...
17021 (state_writer::write_state_structures)
17022 (write_state_param_structs): ditto, renaming to...
17023 (state_writer::write_state_param_structs)
17024 (write_state_variables): ditto, renaming to...
17025 (state_writer::write_state_variables)
17026 (write_state_srcdir): ditto, renaming to...
17027 (state_writer::write_state_srcdir)
17028 (write_state_files_list): ditto, renaming to...
17029 (state_writer::write_state_files_list)
17030 (write_state_languages): ditto, renaming to...
17031 (state_writer::write_state_languages)
17032 (write_state): create a state_writer instance and use it when
17033 writing out the state file
17035 2013-05-17 Mike Stump <mikestump@comcast.net>
17037 PR rtl-optimization/57304
17038 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
17039 accessing DF_REF_REAL_LOC.
17041 2013-05-17 Jakub Jelinek <jakub@redhat.com>
17043 PR rtl-optimization/57281
17044 PR rtl-optimization/57300
17045 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
17046 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
17047 what the other splitter did if the registers are dead.
17049 2013-05-17 Richard Biener <rguenther@suse.de>
17051 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
17054 2013-05-17 Jakub Jelinek <jakub@redhat.com>
17056 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
17059 2013-05-17 Marek Polacek <polacek@redhat.com>
17061 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
17062 length when doing non-zero store of storing '\0' to '\0'.
17064 2013-05-17 Jakub Jelinek <jakub@redhat.com>
17066 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
17067 vect_external_def oprnd1 with loop_vinfo, try to emit
17068 optional cast, negation and and stmts on the loop preheader
17069 edge instead of into the pattern def seq.
17071 PR tree-optimization/57051
17072 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
17073 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
17075 2013-05-16 Nick Clifton <nickc@redhat.com>
17077 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
17078 (rl78_is_naked_func): New function.
17079 (rl78_expand_prologue): Skip prologue generation for naked functions.
17080 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
17081 * doc/extend.texi (naked): Add RL78 to the list of processors
17082 that supports this attribute.
17084 2013-05-16 Jeff Law <law@redhat.com>
17086 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
17088 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
17090 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
17091 cache parameters using detect_caches_amd also for CYRIX,
17092 NSC and TM2 signatures.
17094 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
17095 Dzianis Kahanovich <mahatma@eu.by>
17099 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
17100 VIA/Centaur processors and determine their cache parameters
17101 using detect_caches_amd.
17103 2013-05-16 Teresa Johnson <tejohnson@google.com>
17105 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
17106 (rtl_verify_edges): New function.
17107 (rtl_verify_bb_insns): Ditto.
17108 (rtl_verify_bb_pointers): Ditto.
17109 (rtl_verify_bb_insn_chain): Ditto.
17110 (rtl_verify_fallthru): Ditto.
17111 (rtl_verify_bb_layout): Ditto.
17112 (rtl_verify_flow_info_1): Outline checks into new functions.
17113 (rtl_verify_flow_info): Ditto.
17115 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
17117 * cfghooks.c (copy_bbs): Add update_dominance argument.
17118 * cfghooks.h (copy_bbs): Update prototype.
17119 * tree-cfg.c (gimple_duplicate_sese_region):
17120 Add update_dominance argument.
17121 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
17122 * tree-ssa-loop-ch.c (copy_loop_headers): Update
17123 gimple_duplicate_sese_region call.
17124 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
17125 Update copy_bbs call.
17126 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
17127 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
17129 2013-05-16 Jakub Jelinek <jakub@redhat.com>
17131 * tree-vectorizer.h (NUM_PATTERNS): Increment.
17132 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
17133 vect_recog_rotate_pattern.
17134 (vect_recog_rotate_pattern): New function.
17136 2013-05-16 Jason Merrill <jason@redhat.com>
17138 * Makefile.in (LLINKER): New variable.
17139 (mostlyclean): Remove link mutex.
17140 * configure.ac: Handle --enable-link-mutex.
17141 * lock-and-run.sh: New script.
17143 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17146 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
17149 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17151 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
17153 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
17155 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
17156 * config/arm/arm.c (next_consecutive_mem): New function.
17157 (gen_movmem_ldrd_strd): Likewise.
17158 * config/arm/arm.md (movmemqi): Update condition and code.
17159 (unaligned_loaddi, unaligned_storedi): New patterns.
17161 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17163 * config.gcc: Obsolete *-*-solaris2.9*.
17164 * doc/install.texi (Specific, *-*-solaris2*): Document it.
17166 2013-05-16 Richard Biener <rguenther@suse.de>
17168 * passes.c (init_optimization_passes): Move pass_parallelize_loops
17169 earlier, after GRAPHITE transforms and IV canonicalization.
17171 2013-05-16 Jakub Jelinek <jakub@redhat.com>
17173 * omp-low.c (extract_omp_for_data): For collapsed loops,
17174 if at least one of the loops is known at compile time to
17175 iterate zero times, set count to 0.
17176 (expand_omp_regimplify_p): New function.
17177 (expand_omp_for_generic): For collapsed loops, if at least
17178 one of the loops isn't known to iterate at least once,
17179 add runtime check with setting count to 0.
17180 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
17181 For unsigned types if it isn't known at compile time that
17182 the loop will iterate at least once, add runtime check to bypass
17183 the whole loop if initial condition isn't true.
17185 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
17187 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
17189 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
17191 PR middle-end/57286
17192 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
17193 transformations to avoid an infinite loop.
17195 2013-05-16 Marek Polacek <polacek@redhat.com>
17197 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
17199 2013-05-15 Leif Ekblad <leif@rdos.net>
17201 * config/i386/i386.c (ix86_decompose_address): Use
17202 DEFAULT_TLS_SEG_REG to access TLS segment register.
17203 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
17204 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
17205 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
17207 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
17210 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
17211 sibling calls to functions that would normally be lazily bound,
17212 unless $gp is call-clobbered.
17214 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
17216 * config/i386/i386.c (ix86_option_override_internal): Update
17217 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
17218 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
17219 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
17220 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
17222 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
17224 2013-05-15 Andreas Schwab <schwab@suse.de>
17226 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
17227 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
17230 2013-05-15 Teresa Johnson <tejohnson@google.com>
17232 * loop-unroll.c (report_unroll_peel): Check decision before
17233 emitting unroll/peel message.
17235 2013-05-15 Teresa Johnson <tejohnson@google.com>
17237 * function.h (has_bb_partition): New rtl_data flag.
17238 (bb_reorder_complete): Ditto.
17239 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
17240 instead of flag_reorder_blocks_and_partition.
17241 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
17242 with some enhancements.
17243 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
17244 * bb-reorder.c (connect_traces): Check for has_bb_partition
17245 instead of flag_reorder_blocks_and_partition.
17246 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
17247 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
17248 verify_hot_cold_block_grouping.
17249 (partition_hot_cold_basic_blocks): Set has_bb_partition.
17251 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17254 * config/arm/predicates.md (call_insn_operand): New predicate.
17255 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
17256 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
17257 if insn is not a tail call.
17258 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
17260 * config/arm/arm.h (enum reg_class): New caller save register class.
17261 (REG_CLASS_NAMES): Likewise.
17262 (REG_CLASS_CONTENTS): Likewise.
17263 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
17266 2013-05-15 Richard Biener <rguenther@suse.de>
17268 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
17269 of MSG_OPTIMIZED_LOCATIONS.
17270 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
17271 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
17273 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
17274 of MSG_OPTIMIZED_LOCATIONS.
17275 (execute_vect_slp): Likewise.
17276 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
17277 (vect_create_cond_for_alias_checks): Likewise.
17278 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
17279 (vect_recog_widen_mult_pattern): Likewise.
17280 (vect_recog_widen_sum_pattern): Likewise.
17281 (vect_recog_over_widening_pattern): Likewise.
17282 (vect_recog_widen_shift_pattern): Likewise.
17283 (vect_recog_vector_vector_shift_pattern): Likewise.
17284 (vect_recog_divmod_pattern): Likewise.
17285 (vect_recog_mixed_size_cond_pattern): Likewise.
17286 (vect_recog_bool_pattern): Likewise.
17287 (vect_pattern_recog_1): Likewise.
17289 2013-05-15 Martin Jambor <mjambor@suse.cz>
17291 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
17292 non-functions to builtin_unreachable.
17293 * ipa-inline-transform.c (inline_call): Do not assert estimates were
17294 correct when new direct edges were discovered.
17296 2013-05-15 Martin Jambor <mjambor@suse.cz>
17298 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
17299 header, print symbol order instead of node uid, print more information
17300 about indirect edge targets.
17301 (ipa_make_edge_direct_to_target): Print symbol order instead of node
17303 (ipa_make_edge_direct_to_target): Likewise.
17304 (remove_described_reference): Likewise.
17305 (propagate_controlled_uses): Likewise.
17306 (ipa_print_node_params): Also print symbol order.
17307 (ipcp_transform_function): Print symbol order instead of node uids.
17308 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
17309 (cgraph_get_create_real_symbol_node): Likewise.
17310 * ipa-cp.c (print_lattice): Likewise.
17311 (print_all_lattices): Likewise.
17312 (determine_versionability): Likewise.
17313 (initialize_node_lattices): Likewise.
17314 (estimate_local_effects): Likewise.
17315 (update_profiling_info): Likewise.
17316 (create_specialized_node): Likewise.
17317 (perhaps_add_new_callers): Likewise.
17318 (decide_about_value): Likewise.
17319 (decide_whether_version_node): Likewise.
17320 (identify_dead_nodes): Likewise.
17321 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
17322 (dump_inline_summary): Likewise.
17323 (estimate_node_size_and_time): Likewise.
17324 (inline_analyze_function): Likewise.
17325 * ipa-inline.c (report_inline_failed_reason): Likewise.
17326 (want_early_inline_function_p): Likewise.
17327 (edge_badness): Likewise.
17328 (update_edge_key): Likewise.
17329 (inline_small_functions): Likewise. Add dumping of order to two other
17331 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
17332 instead of node uids.
17333 (propagate_pure_const): Likewise.
17334 (propagate_pure_const): Likewise.
17335 * ipa-utils.c (dump_cgraph_node_set): Likewise.
17336 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
17337 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
17339 * tree-pretty-print.c (dump_function_header): Likewise.
17340 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
17341 Print symbol order instead of node uids.
17343 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17345 * config/s390/s390.c (s390_register_move_cost): Don't impose the
17346 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
17348 2013-05-15 Richard Biener <rguenther@suse.de>
17350 PR tree-optimization/57275
17351 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
17352 return value for fail to do runtime alias checks for gather loads.
17354 2013-05-15 Jan Hubicka <jh@suse.cz>
17358 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
17359 weakrefs are not external.
17360 (lto_symtab_merge_decls): Fix thinko when dealing with
17361 non-lto_symtab decls.
17362 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
17363 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
17364 * varpool.c (dump_varpool_node): Dump more flags.
17366 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17368 * config/i386/i386.c (processor_alias_table): Add instruction
17369 FSGSBASE for AMD bdver3 architecture.
17371 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17373 * tree.c (warn_deprecated_use): Print file:line using locus color.
17374 * diagnostic.c (diagnostic_report_current_module): Print file:line
17375 and file:line:column using locus color.
17377 2013-05-14 Mike Stump <mikestump@comcast.net>
17379 * gdbinit.in: Add __null.
17381 2013-05-14 Mike Stump <mikestump@comcast.net>
17383 * recog.h: Rename struct recog_data to Recog_data.
17384 * recog.c: Likewise.
17385 * reload.c (can_reload_into): Likewise.
17386 * config/picochip/picochip.c: Likewise.
17388 2013-05-14 Mike Stump <mikestump@comcast.net>
17390 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
17392 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
17394 * resource.h (struct resources): Remove unch_memory member.
17395 (CLEAR_RESOURCE): Don't clear unch_memory.
17396 * resource.c (mark_referenced_resources): Don't set it.
17397 (mark_set_resources): Likewise.
17398 (mark_target_live_regs): Don't clear it.
17399 (init_resource_info): Likewise.
17400 * reorg.c (resource_conflicts_p): Don't compare it.
17401 (redundant_insn): Don't set it.
17403 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
17405 * emit-rtl.c (next_label): Remove unused function.
17406 (skip_consecutive_labels, link_cc0_insns): Move to ...
17407 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
17408 only place where these functions are used, and make them static.
17410 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17412 * fold-const.c (fold_negate_expr): Handle vectors.
17413 (fold_truth_not_expr): Make it static.
17414 (fold_invert_truthvalue): New static function.
17415 (invert_truthvalue_loc): Handle vectors. Do not call
17416 fold_truth_not_expr directly.
17417 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
17418 <TRUTH_NOT_EXPR>: Do not cast to boolean.
17419 (fold_comparison): Handle vector constants.
17420 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
17421 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
17422 * tree.h (fold_truth_not_expr): Remove declaration.
17424 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
17426 * config/aarch64/aarch64-simd.md
17427 (aarch64_vcond_internal<mode>): Rename to...
17428 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
17429 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
17430 float modes. Clarify all iterator modes.
17431 (vcond<mode><mode>): Use new name for vcond expanders.
17432 (vcond<v_cmp_result><mode>): Likewise.
17433 (vcondu<mode><mode>: Likewise.
17434 * config/aarch64/iterators.md (VDQF_COND): New.
17436 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17439 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
17440 variable for the shift amount. Check that we shift by non-negative
17443 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
17446 * config/arm/arm.h (EPILOGUE_USES): Only return true
17447 for LR_REGNUM after epilogue_completed.
17449 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
17451 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
17452 is error_mark_node.
17454 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17457 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
17458 and Solaris 11+/x86 with gld.
17459 * configure: Regenerate.
17461 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17463 * expmed.c (expand_shift_1): Canonicalize rotates by
17464 constant bitsize / 2 to bitsize - 1.
17465 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
17466 case ROTATERT>: Likewise.
17469 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17471 * config/i386/i386.md (rotateinv): New code attr.
17472 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17473 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17474 roll $31, %eax, etc.
17476 2013-05-14 Richard Biener <rguenther@suse.de>
17478 PR middle-end/57235
17479 * tree-eh.c (sink_clobbers): Give up for successors with
17480 multiple predecessors and no virtual uses.
17482 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
17484 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
17485 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
17487 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17489 PR middle-end/57251
17490 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
17491 the case when both op0 and op1 have VOIDmode.
17493 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17495 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
17496 in multiply-accumulate mode.
17498 2013-05-13 Guozhi Wei <carrot@google.com>
17500 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
17502 2013-05-13 Kai Tietz <ktietz@redhat.com>
17505 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
17506 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
17507 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
17508 * config/i386/i386.c (ix86_option_override_internal): Likewise.
17509 (ix86_expand_prologue): Likewise.
17510 (ix86_expand_split_stack_prologue): Likewise.
17511 (legitimate_pic_address_disp_p): Likewise.
17512 (legitimize_pic_address): Likewise.
17513 (legitimize_tls_address): Likewise.
17514 (legitimize_pe_coff_symbol): Likewise.
17515 (output_pic_addr_const): Likewise.
17516 (construct_plt_address): Likewise.
17517 (ix86_expand_call): Likewise.
17518 (x86_output_mi_thunk): Likewise.
17519 (x86_function_profiler): Likewise.
17521 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
17523 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
17524 similar switch cases.
17525 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
17526 (aarch64_simd_mov_to_<mode>low): Delete.
17527 (aarch64_simd_mov_to_<mode>high): Delete.
17528 (move_lo_quad_<mode>): Add w<-r alternative.
17529 (aarch64_simd_move_hi_quad_<mode>): Likewise.
17530 (aarch64_simd_mov_from_*): Update type attribute.
17531 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
17534 2013-05-13 Jan Hubicka <jh@suse.cz>
17536 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
17537 * config/i386/i386.c (ix86_compute_frame_layout,
17538 ix86_expand_epilogue, emit_i387_cw_initialization,
17539 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
17540 ix86_local_alignment): Fix use of size/speed predicates.
17542 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17544 PR tree-optimization/45216
17545 PR tree-optimization/57157
17546 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
17547 the (-Y) & (B - 1) variant if OP is |.
17548 * expmed.c (expand_shift_1): For rotations by const0_rtx just
17549 return shifted. Use (-op1) & (prec - 1) as other_amount
17550 instead of prec - op1.
17552 2013-05-13 Martin Jambor <mjambor@suse.cz>
17554 PR middle-end/42371
17555 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
17556 (ipa_constant_data): New type.
17557 (ipa_jump_func): Use ipa_constant_data to hold information about
17558 constant jump functions.
17559 (ipa_get_jf_constant): Adjust to jump function type changes.
17560 (ipa_get_jf_constant_rdesc): New function.
17561 (ipa_param_descriptor): New field controlled_uses.
17562 (ipa_get_controlled_uses): New function.
17563 (ipa_set_controlled_uses): Likewise.
17564 * ipa-ref.h (ipa_find_reference): Declare.
17565 * ipa-prop.c (ipa_cst_ref_desc): New type.
17566 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
17568 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
17569 New parameter cs. Adjust all callers.
17570 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
17571 (remove_described_reference): New function.
17572 (jfunc_rdesc_usable): Likewise.
17573 (try_make_edge_direct_simple_call): Decrement controlled use count,
17574 attempt to remove reference if it hits zero.
17575 (combine_controlled_uses_counters): New function.
17576 (propagate_controlled_uses): Likewise.
17577 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
17578 (ipa_edge_duplication_hook): Duplicate reference descriptions.
17579 (ipa_print_node_params): Print described use counter.
17580 (ipa_write_jump_function): Adjust to jump function type changes.
17581 (ipa_read_jump_function): New parameter CS, pass it to
17582 ipa_set_jf_constant. Adjust caller.
17583 (ipa_write_node_info): Stream controlled use count
17584 (ipa_read_node_info): Likewise.
17585 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
17587 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
17588 count. Remove cloning-added reference if it reaches zero.
17589 * ipa-ref.c (ipa_find_reference): New function.
17591 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17593 * config/i386/i386.c (processor_target_table): Modified default
17594 alignment values for AMD BD and BT architectures.
17596 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
17598 * tree-vect-generic.c (uniform_vector_p): Move ...
17599 * tree.c (uniform_vector_p): ... here.
17600 * tree.h (uniform_vector_p): Declare it.
17601 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
17604 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17606 PR tree-optimization/57230
17607 * tree-ssa-strlen.c (handle_char_store): Record length for
17608 array store from STRING_CST.
17610 PR tree-optimization/57230
17611 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
17614 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
17616 * config/epiphany/epiphany.c (epiphany_init): Check size of
17617 NUM_MODES_FOR_MODE_SWITCHING.
17618 (epiphany_expand_prologue):
17619 Remove CONFIG_REGNUM initial value handling code.
17620 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
17621 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
17622 (emit_set_fp_mode, epiphany_mode_after): Likewise.
17623 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
17624 Don't return 1 for FP_MODE_NONE.
17625 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
17626 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
17627 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
17628 * config/epiphany/epiphany.md (save_config): New pattern.
17630 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
17632 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
17634 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
17636 * config/i386/i386.md (memory): Handle sseishft1.
17637 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
17638 (*vec_extractv2di_1): Ditto.
17640 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
17642 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
17645 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
17647 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
17648 Add mthumb/march=armv7-a multilib.
17649 Add mthumb/march=armv7-r multilib.
17650 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
17652 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
17654 * config/v850/t-rtems: Add more multilibs.
17656 2013-05-10 Richard Biener <rguenther@suse.de>
17658 PR tree-optimization/57214
17659 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
17660 not propagate from SSA names that occur in abnormal PHI nodes.
17662 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
17664 * stor-layout.c (element_precision): New function.
17665 * machmode.h (element_precision): Declare it.
17666 * tree.c (build_minus_one_cst): New function.
17667 (element_precision): Likewise.
17668 * tree.h (build_minus_one_cst): Declare new function.
17669 (element_precision): Likewise.
17670 * fold-const.c (operand_equal_p): Use element_precision.
17671 (fold_binary_loc): Handle vector types.
17672 * convert.c (convert_to_integer): Use element_precision.
17673 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
17676 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
17678 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
17679 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
17680 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
17681 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17682 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
17683 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
17684 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
17685 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17686 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
17687 (Uuw8): New constraints.
17688 (Usb4): Move into alphabetical order.
17689 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
17690 (sd8_operand, ub8_operand, uw8_operand): New predicates.
17691 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
17692 previously unnamed patterns.
17693 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
17694 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
17695 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
17696 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
17697 of set_attr_alternative/if_then_else. Use extended_mips16 instead
17698 of specific lengths.
17700 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17702 * config/i386/i386.md (rotateinv): New code attr.
17703 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17704 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17705 roll $31, %eax, etc.
17707 PR tree-optimization/45216
17708 PR tree-optimization/57157
17709 * tree-ssa-forwprop.c (simplify_rotate): New function.
17710 (ssa_forward_propagate_and_combine): Call it.
17712 2013-05-10 Richard Biener <rguenther@suse.de>
17714 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
17715 disable peeling when we version for aliasing.
17716 (vector_alignment_reachable_p): Honor explicit user alignment.
17717 (vect_supportable_dr_alignment): Likewise.
17718 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
17719 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
17720 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
17721 then peeling to arrange for the cost-model check to come first.
17723 2013-05-10 Alan Modra <amodra@gmail.com>
17725 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
17726 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
17727 * configure: Regenerate.
17729 2013-05-10 Alan Modra <amodra@gmail.com>
17732 * varasm.c (default_elf_select_section): Move !DECL_P check..
17733 (get_named_section): ..to here before calling get_section_name.
17735 (default_section_type_flags): Add DECL_P check.
17736 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
17737 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
17739 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
17741 * config/epiphany/epiphany.c (epiphany_expand_prologue):
17742 When using gen_stack_adjust_str with a register offset, add a
17743 REG_FRAME_RELATED_EXPR note.
17745 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
17747 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
17748 (*vec_extractv4si_zext_mem): Ditto.
17749 (*vec_extractv2di): Add 0->x and x->x alternatives.
17750 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
17751 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
17753 2013-05-09 Jason Merrill <jason@redhat.com>
17755 N3639 C++1y VLA support
17756 * gimplify.c (gimplify_vla_decl): Don't touch an existing
17759 * tree.c (build_constructor_va): New.
17760 * tree.h: Declare it.
17762 2013-05-09 Martin Jambor <mjambor@suse.cz>
17765 * gimple-fold.c (canonicalize_constructor_val): Call
17766 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
17768 2013-05-09 Jan Hubicka <jh@suse.cz>
17769 Richard Biener <rguenther@suse.de>
17772 * symtab.c (symtab_make_decl_local): Do not add private names.
17774 2013-05-09 Jan Hubicka <jh@suse.cz>
17777 * symtab.c (insert_to_assembler_name_hash): Handle clones.
17778 (unlink_from_assembler_name_hash): Likewise.
17779 (symtab_prevail_in_asm_name_hash, symtab_register_node,
17780 symtab_unregister_node, symtab_initialize_asm_name_hash,
17781 change_decl_assembler_name): Update.
17783 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
17785 * config/aarch64/aarch64.md: New movtf split.
17786 (*movtf_aarch64): Update.
17787 (aarch64_movdi_tilow): Handle TF modes and rename to
17788 aarch64_movdi_<mode>low.
17789 (aarch64_movdi_tihigh): Handle TF modes and rename to
17790 aarch64_movdi_<mode>high
17791 (aarch64_movtihigh_di): Handle TF modes and rename to
17792 aarch64_mov<mode>high_di
17793 (aarch64_movtilow_di): Handle TF modes and rename to
17794 aarch64_mov<mode>low_di
17795 (aarch64_movtilow_tilow): Remove spurious whitespace.
17796 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
17798 (aarch64_print_operand): Update.
17800 2013-05-09 Alan Modra <amodra@gmail.com>
17802 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
17804 * configure: Regenerate.
17806 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17808 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
17809 splitter preparation statements.
17810 * config/i386/sse.md (*vec_extract* splitters): Ditto.
17811 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
17814 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17816 * gimple-ssa-strength-reduction.c (count_candidates): Change
17817 return value to int.
17818 (analyze_candidates_and_replace): Change type of length to int.
17820 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17822 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
17823 (*vec_extract<mode>): Use VI12_128 mode iterator.
17824 (*vec_extract<mode>_mem): Ditto.
17825 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
17828 2013-05-08 Diego Novillo <dnovillo@google.com>
17833 2012-08-17 Diego Novillo <dnovillo@google.com>
17836 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
17837 * config.in: Regenerate.
17838 * configure: Regenerate.
17839 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
17841 2013-05-08 Jan Hubicka <jh@suse.cz>
17844 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
17845 * cgraph.h (symtab_node_base): Add unique_name.
17846 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17847 input_overwrite_node, input_varpool_node): Stream unique_name.
17848 * cgraphclones.c (cgraph_create_virtual_clone,
17849 cgraph_function_versioning): Set unique_name.
17850 * ipa.c (function_and_variable_visibility): Set unique_name.
17852 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17854 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
17855 (alloc_cand_and_find_basis): Restrict conditional candidate
17856 processing to CAND_MULTs.
17858 2013-05-08 Jan Hubicka <jh@suse.cz>
17861 lto-symtab.c (lto_symtab_symbol_p): New function.
17862 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
17863 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
17864 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
17865 Skip static symbols.
17867 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
17869 PR tree-optimization/57200
17870 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
17871 Only call inform if the preceding warning_at returns true.
17873 2013-05-07 Han Shen <shenhan@google.com>
17875 * cfgexpand.c (record_or_union_type_has_array_p): New function.
17876 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
17877 * common.opt (fstack-protector-strong): New option.
17878 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
17879 * doc/invoke.texi (Optimization Options): Document
17880 "-fstack-protector-strong".
17881 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
17883 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17885 * config/mips/mips.c (mips_machine_reorg2): Return 0.
17887 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
17889 * ira.c (update_equiv_regs): Add insn having equiv memory even if
17890 it is not lhs of the insn.
17891 (setup_reg_equiv): Remove insn having equiv memory which it is not
17893 * lra-constraints.c (process_address): Try to improve generation
17894 code for address base + disp.
17895 (lra_constraints): Make correct the code for checking insn setting
17896 up backward equivalence. Remove insn only if it is in the init
17898 * lra-eliminations.c (update_reg_eliminate): Change return value.
17899 (lra_eliminate): Use the result.
17901 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
17903 * config/i386/sse.md (ssescalarnummask): New mode attribute.
17904 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
17905 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
17906 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
17907 register target operands.
17908 (*vec_extractv8hi_sse2): New pattern.
17909 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
17910 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
17911 (*vec_extract<mode>_mem): New insn and split pattern.
17913 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
17915 * config/arm/arm.c (arm_asan_shadow_offset): New function.
17916 (TARGET_ASAN_SHADOW_OFFSET): Define.
17917 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
17918 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
17920 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17922 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
17923 (incr_vec_index): Return -1 if increment not found.
17924 (create_add_on_incoming_edge): Assert if increment not found.
17925 (record_increment): Limit number of increments recorded.
17926 (all_phi_incrs_profitable): Return false if an increment not found.
17927 (replace_profitable_candidates): Don't process increments that were
17929 (analyze_candidates_and_replace): Limit size of incr_vec.
17931 2013-05-07 Richard Biener <rguenther@suse.de>
17933 * calls.c (special_function_p): setjmp-like functions are leaf.
17934 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
17935 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
17937 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
17939 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
17940 (aarch64_simd_mov<mode>): New expander.
17941 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
17942 (aarch64_simd_mov_to_<mode>high): Likewise.
17943 (aarch64_simd_mov_from_<mode>low): Likewise.
17944 (aarch64_simd_mov_from_<mode>high): Likewise.
17945 (aarch64_dup_lane<mode>): Update.
17946 (aarch64_dup_lanedi): New instruction pattern.
17947 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
17948 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
17950 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17952 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
17953 (replace_mult_candidate): Remove unnecessary argument; remove
17954 unnecessary parameter from call to introduce_cast_before_cand.
17955 (replace_unconditional_candidate): Remove unnecessary parameter
17956 from call to replace_mult_candidate.
17957 (replace_conditional_candidate): Likewise.
17958 (insert_initializers): Use make_temp_ssa_name.
17959 (introduce_cast_before_cand): Remove unnecessary argument; use
17960 make_temp_ssa_name.
17961 (replace_one_candidate): Remove unnecessary argument; remove
17962 unnecessary parameter from calls to introduce_cast_before_cand.
17963 (replace_profitable_candidates): Remove unnecessary parameters
17964 from calls to replace_one_candidate.
17966 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17968 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
17969 phi def as possibly hiding a basis for a CAND_ADD whose operands
17970 have been commuted in the analysis.
17971 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
17973 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17975 * config/aarch64/aarch64.md
17976 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
17977 shift value between 0-4.
17979 2013-05-07 Richard Biener <rguenther@suse.de>
17981 * double-int.h (rshift): New overload.
17982 * double-int.c (rshift): New function.
17983 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
17984 (create_reference_ops_from_ref): Remove.
17985 (vn_reference_insert): Use shared ops for constructing the
17986 reference and copy it.
17988 2013-05-07 Richard Biener <rguenther@suse.de>
17990 PR middle-end/57190
17991 * tree-eh.c (sink_clobbers): Properly propagate
17992 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
17994 2013-05-07 Jakub Jelinek <jakub@redhat.com>
17996 PR tree-optimization/57149
17997 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
17998 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
17999 collect_phi_def_edges, execute_late_warn_uninitialized): Use
18000 uninit_undefined_value_p instead of ssa_undefined_value_p.
18003 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
18004 for modifier == EXPAND_INITIALIZER.
18006 2013-05-07 Anton Blanchard <anton@samba.org>
18008 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
18009 for powerpc64 little endian.
18010 * configure: Regenerate.
18012 2013-05-06 Graham Stott <grahams@btinternet.com>
18014 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
18015 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
18016 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
18017 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
18019 2013-05-06 Graham Stott <grahams@btinternet.com>
18021 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
18022 codes which allow non-lvalues.
18024 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
18026 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
18027 components are all 1s.
18028 (integer_minus_onep): New function.
18029 * tree.h (integer_minus_onep): Declare it.
18030 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
18031 integer_minus_onep instead of integer_all_onesp.
18033 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18036 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
18037 variations of these patterns.
18039 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
18041 * config/i386/i386.md (isa): Add x64_sse4 member.
18042 (enabled): Handle x64_sse4.
18043 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
18044 instruction for 64bit SSE4_1 targets. Update insn attributes.
18045 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
18046 instruction for SSE4_1 targets. Update insn attributes.
18047 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
18048 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
18049 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
18051 (*vec_extractv4si): Rename from *sse4_1_pextrd.
18052 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
18053 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
18055 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18058 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
18060 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
18062 * final.c (do_assembler_dialects): Don't handle curly braces and
18063 vertical bar escaped by % as dialect delimiters.
18064 (output_asm_insn): Print curly braces and vertical bar if escaped
18065 by % and ASSEMBLER_DIALECT defined.
18066 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
18067 * doc/tm.texi: Regenerated.
18069 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
18071 * config/mips/mips.c: Include tree-pass.h.
18072 (mips_reorg): Split in pre- and post-dbr_schedule parts.
18073 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
18074 (pass_mips_machine_reorg2): New machine specific pass.
18075 (insert_pass_mips_machine_reorg2): New pass plugin definition.
18076 (mips_option_override): Register the new pass.
18077 * rtl.h (cleanup_barriers): Remove prototype.
18078 (dbr_schedule): Likewise.
18079 * jump.c (cleanup_barriers): Make static.
18080 * reorg.c (dbr_schedule): Likewise.
18082 2013-05-06 Richard Biener <rguenther@suse.de>
18084 PR tree-optimization/57185
18085 * tree-parloops.c (add_field_for_reduction): Handle anonymous
18086 SSA names properly.
18088 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
18091 * config/i386/i386.c (add_parameter_dependencies): Add dependence
18092 between "first_arg" and "insn", not "last" and "insn".
18094 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18096 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
18097 (find_candidates_in_block): Re-enable slsr_process_phi.
18098 (create_phi_basis): Fix double counting of candidate adjustment.
18100 2013-05-06 Richard Biener <rguenther@suse.de>
18102 PR middle-end/57147
18103 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
18104 the edge is also fallthru, preserve it and just clear the
18106 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
18107 also complex, preserve that and just clear the fallthru flag.
18108 * tree-inline.c (update_ssa_across_abnormal_edges): Also
18109 update virtual operands.
18111 2013-05-06 Alan Modra <amodra@gmail.com>
18113 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
18114 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
18115 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
18116 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
18117 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
18118 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
18120 2013-05-06 Alan Modra <amodra@gmail.com>
18122 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
18123 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
18124 (DEFAULT_ASM_ENDIAN): Define.
18125 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
18126 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
18127 Update -K PIC clause from sysv4.h.
18128 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
18129 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
18131 2013-05-06 Alan Modra <amodra@gmail.com>
18133 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
18134 twice for little-endian.
18135 (ashrdi3_no_power, ashrdi3): Support little-endian.
18137 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18140 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
18141 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
18143 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
18144 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
18145 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
18148 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
18149 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18151 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
18152 * config/pa/pa.opt: Make mbig-switch a no-op.
18153 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
18154 (CASE_VECTOR_MODE): Always return SImode.
18155 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
18156 for the !TARGET_BIG_SWITCH case.
18157 * config/pa/pa-linux.h: Likewise.
18158 * config/pa/pa-openbsd.h: Likewise.
18159 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
18160 * config/pa/pa.md (short_jump): Remove define_insn.
18161 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
18162 (casesi0): Remove define_insn.
18163 (type): Remove btable_branch.
18164 (pa_combine_type): Likewise.
18165 (in_nullified_branch_delay): Likewise.
18166 (in_call_delay): Likewise.
18167 (define_delay): Likewise.
18168 (define_insn_reservation "Z3"): Likewise.
18169 (define_insn_reservation "Z4"): Likewise.
18170 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
18171 (pa_adjust_insn_length): Remove adjustment for btable branches.
18172 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
18175 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
18177 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
18178 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
18179 Add m->r,x alternatives.
18180 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
18181 splitters using SWI48x mode iterator.
18182 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
18183 TARGET_64BIT. Add m->x alternative.
18184 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
18185 Add o->x alternative. Enable for TARGET_SSE.
18186 (sse_storeq): Remove expander.
18187 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
18188 with memory input operand.
18189 (*vec_extractv2di_1 splitter): New.
18190 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
18191 * config/i386/i386.md (ssevecmodelower): New mode attribute.
18193 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
18195 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
18196 (INT_LOWPART): Delete.
18197 (extract_MB): Adjust.
18198 (extract_ME): Adjust.
18199 (print_operand): Adjust.
18201 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
18203 * config/rs6000/predicates.md (reg_or_add_cint_operand,
18204 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
18205 (reg_or_logical_cint_operand, easy_fp_constant,
18206 logical_const_operand): Delete "CONST_DOUBLE" case.
18207 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
18208 "HOST_BITS_PER_WIDE_INT == 64" test.
18209 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
18210 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
18211 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
18212 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
18214 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
18215 CONST_DOUBLE DImode/VOIDmode case.
18216 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
18217 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
18218 CONST_DOUBLE VOIDmode case.
18219 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
18220 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
18221 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
18222 Delete CONST_DOUBLE case.
18223 (splitters for mov FMOVE64 const_double): Delete
18224 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
18225 "HOST_BITS_PER_WIDE_INT >= 64" test.
18226 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
18228 (mov DI const_double): Delete.
18230 2013-05-04 Jakub Jelinek <jakub@redhat.com>
18232 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
18233 on op shows all bits zero in mode of a lowpart subreg, return zero.
18235 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
18238 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
18239 to save TFmode registers and DImode to save TImode registers for
18240 caller save operations.
18241 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
18242 mark being partially clobbered since they only use the first
18245 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
18246 and TDmode only use the upper 64-bits of each VSX register.
18248 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18250 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
18251 (find_candidates_in_block): Disable slsr_process_phi.
18253 2013-05-03 Guozhi Wei <carrot@google.com>
18255 * coverage.c (coverage_obj_init): Move the construction of gcov
18257 (build_init_ctor): ... here.
18259 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18261 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
18262 (slsr_cand_d): Redefine def_phi.
18263 (stride_status, phi_adjust_status, count_phis_status): New enums.
18264 (find_phi_def): New.
18265 (find_basis_for_base_expr): New.
18266 (find_basis_for_candidate): Handle hidden bases.
18267 (alloc_cand_and_find_basis): Handle phi candidates.
18268 (slsr_process_phi): New.
18269 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
18270 (create_mul_imm_cand): Likewise.
18271 (create_add_ssa_cand): Exclude phi base candidates.
18272 (create_add_imm_cand): Likewise.
18273 (slsr_process_cast): Likewise.
18274 (slsr_process_copy): Likewise.
18275 (find_candidates_in_block): Handle phi candidates.
18276 (dump_candidate): Likewise.
18277 (unconditional_cands): Delete.
18278 (unconditional_cands_with_known_stride_p): Delete.
18279 (phi_dependent_cand_p): New.
18280 (cand_increment): Handle phi-dependent candidates.
18281 (replace_dependent): Delete.
18282 (replace_mult_candidate): New.
18283 (replace_unconditional_candidate): New.
18284 (incr_vec_index): Move to avoid forward reference.
18285 (create_add_on_incoming_edge): New.
18286 (create_phi_basis): New.
18287 (replace_dependents): Delete.
18288 (replace_conditional_candidate): New.
18289 (phi_add_costs): New.
18290 (replace_uncond_cands_and_profitable_phis): New.
18291 (record_increment): Handle phi adjustments.
18292 (record_phi_increments): New.
18293 (record_increments): Handle phi adjustments.
18294 (phi_incr_cost): New.
18295 (lowest_cost_path): Handle phis.
18296 (total_savings): Likewise.
18297 (analyze_increments): Likewise.
18298 (ncd_with_phi): New.
18299 (ncd_of_cand_and_phis): New.
18300 (nearest_common_dominator_for_cands): Handle phi increments.
18301 (all_phi_incrs_profitable): New.
18302 (replace_profitable_candidates): Handle phi-dependent candidates.
18303 (analyze_candidates_and_replace): Likewise.
18305 2013-05-03 Teresa Johnson <tejohnson@google.com>
18308 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
18309 do not exceed REG_BR_PROB_BASE.
18311 2013-05-03 Jeff Law <law@redhat.com>
18313 PR tree-optimization/57144
18314 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
18315 operand of the condition will bit into the new type when eliminating
18316 a cast feeding a condition.
18318 2013-05-03 Jakub Jelinek <jakub@redhat.com>
18320 PR rtl-optimization/57130
18321 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
18322 of COMPARE as in_code to the recursive call if needed.
18324 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
18326 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
18327 (enabled): Handle new members.
18328 * config/i386/sse.md (*vec_concatv2si): Merge from
18329 *vec_concatv2si_sse2 and vec_concatv2si_sse.
18330 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
18332 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
18334 PR tree-optimization/57027
18335 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
18336 for fnms opportunity, check we got the prerequisite kind
18337 of tree / gimple before using accessor functions.
18339 2013-05-03 Richard Biener <rguenther@suse.de>
18341 * double-int.h (lshift): New overload without precision
18342 and arith argument.
18343 (operator *=, operator +=, operator -=): Move ...
18344 * double-int.c (operator *=, operator +=, operator -=): ... here
18345 and implement more efficiently.
18346 (mul_double_with_sign): Remove.
18347 (lshift_double): Adjust to take unsinged shift argument, push
18348 dispatching code to callers.
18349 (mul_double_wide_with_sign): Add early out for callers that
18350 are not interested in high parts or overflow.
18351 (lshift): New function.
18352 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
18353 dispatch code here.
18354 (lrotate, rrotate): Use logical shifts.
18355 * expr.c (get_inner_reference): Use lshift.
18356 * fixed-value.c (do_fixed_divide): Likewise.
18357 * tree-dfa.c (get_ref_base_and_extent): Likewise.
18358 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
18359 (indirect_refs_may_alias_p): Likewise.
18360 (stmt_kills_ref_p_1): Likewise.
18362 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18364 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
18366 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18368 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
18369 scalar form of FABD instruction.
18371 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
18373 * lra-constraints.c (process_alt_operands): Add checking alt
18374 number to choose the best alternative.
18376 2013-05-02 Richard Biener <rguenther@suse.de>
18378 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
18379 bitmap and its handling.
18380 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
18382 2013-05-02 Richard Biener <rguenther@suse.de>
18384 PR middle-end/57140
18385 * tree-inline.c (copy_loops): Properly handle removed loops.
18386 (copy_cfg_body): Mark destination loops for fixup if source
18387 loops needed fixup.
18389 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
18392 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
18393 generating simple_return for naked functions.
18395 2013-05-02 Martin Jambor <mjambor@suse.cz>
18397 PR middle-end/56988
18398 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
18399 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
18401 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
18402 ipa_agg_replacement_value structures.
18403 (known_aggs_to_agg_replacement_list): Likewise.
18404 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
18405 (read_agg_replacement_chain): Likewise.
18406 (ipcp_transform_function): Also check that by_ref flags match.
18408 2013-05-02 Richard Biener <rguenther@suse.de>
18410 * graphds.h (struct graph): Add obstack member.
18411 * graphds.c (new_graph): Initialize obstack and allocate
18413 (add_edge): Allocate edge from the obstack.
18414 (free_graph): Free the obstack instead of all edges and vertices.
18416 2013-05-02 Teresa Johnson <tejohnson@google.com>
18418 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
18420 * cfg.c (update_bb_profile_for_threading): Ditto.
18421 * tree-inline.c (copy_bb): Ditto.
18422 (copy_edges_for_bb): Ditto.
18423 (initialize_cfun): Ditto.
18424 (copy_cfg_body): Ditto.
18425 (expand_call_inline): Ditto.
18426 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
18427 (estimate_node_size_and_time): Ditto.
18428 (inline_merge_summary): Ditto.
18429 * cgraphclones.c (cgraph_clone_edge): Ditto.
18430 (cgraph_clone_node): Ditto.
18431 * sched-rgn.c (compute_dom_prob_ps): Ditto.
18432 (compute_trg_info): Ditto.
18434 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18436 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
18437 S reg when fp attribute set.
18438 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
18440 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18442 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
18444 (*and_one_cmplsi3_compare0_uxtw): Likewise.
18445 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
18446 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
18448 2013-05-02 Richard Biener <rguenther@suse.de>
18450 * tree-scalar-evolution.c (scev_info_hasher): Remove.
18451 (struct instantiate_cache_entry): New type.
18452 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
18453 (struct instantiate_cache_type): New type.
18454 (set_instantiated_value, get_instantiated_value): Remove.
18455 (get_instantiated_value_entry): New function.
18456 (instantiate_scev_name): Use the new cache and adjust.
18457 (instantiate_scev_poly): Adjust.
18458 (instantiate_scev_binary): Likewise.
18459 (instantiate_array_ref): Likewise.
18460 (instantiate_scev_convert): Likewise.
18461 (instantiate_scev_not): Likewise.
18462 (instantiate_scev_3): Likewise.
18463 (instantiate_scev_2): Likewise.
18464 (instantiate_scev_r): Likewise.
18465 (instantiate_scev): Likewise.
18466 (resolve_mixers): Likewise.
18468 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
18471 * lra-constraints.c (best_small_class_operands_num): Remove.
18472 (process_alt_operands): Remove small_class_operands_num. Take
18473 small classes operands into losers and only if the operand is not
18474 matched. Modify debugging output.
18475 (curr_insn_transform): Remove best_small_class_operands_num.
18478 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18480 * config/aarch64/aarch64-builtins.c
18481 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
18482 * config/aarch64/aarch64-simd-builtins.def
18483 (reduc_splus_): Add new modes.
18484 (reduc_uplus_): New.
18485 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
18486 (reduc_uplus_v4sf): Likewise.
18487 (reduc_splus_v4sf): Likewise.
18488 (aarch64_addv<mode>): Likewise.
18489 (reduc_uplus_<mode>): Likewise.
18490 (reduc_splus_<mode>): Likewise.
18491 (aarch64_addvv2di): Likewise.
18492 (reduc_uplus_v2di): Likewise.
18493 (reduc_splus_v2di): Likewise.
18494 (aarch64_addvv2si): Likewise.
18495 (reduc_uplus_v2si): Likewise.
18496 (reduc_splus_v2si): Likewise.
18497 (reduc_<sur>plus_<mode>): New.
18498 (reduc_<sur>plus_v2di): Likewise.
18499 (reduc_<sur>plus_v2si): Likewise.
18500 (reduc_<sur>plus_v4sf): Likewise.
18501 (aarch64_addpv4sf): Likewise.
18502 * config/aarch64/arm_neon.h
18503 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
18504 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
18505 add UNSPEC_SADDV, UNSPEC_UADDV.
18507 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
18509 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18511 * config/aarch64/arm_neon.h
18512 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
18514 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18516 * config/aarch64/aarch64-builtins
18517 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
18519 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18521 * config/aarch64/aarch64-simd-builtins.def
18522 (reduc_smax_): New.
18523 (reduc_smin_): Likewise.
18524 (reduc_umax_): Likewise.
18525 (reduc_umin_): Likewise.
18526 (reduc_smax_nan_): Likewise.
18527 (reduc_smin_nan_): Likewise.
18530 (smax): Update for V2SF, V4SF and V2DF modes.
18533 (smin_nan): Likewise.
18534 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
18535 (<su><maxmin><mode>3): ...This, refactor.
18536 (s<maxmin><mode>3): New.
18537 (<maxmin_uns><mode>3): Likewise.
18538 (reduc_<maxmin_uns>_<mode>): Refactor.
18539 (reduc_<maxmin_uns>_v4sf): Likewise.
18540 (reduc_<maxmin_uns>_v2si): Likewise.
18541 (aarch64_<fmaxmin><mode>: Remove.
18542 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
18544 (vmin<q>_f<32,64>): Likewise.
18545 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18547 (su): Add mappings for smax, smin, umax, umin.
18549 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18550 (FMAXMIN): Rename as...
18551 (FMAXMIN_UNS): ...This.
18553 (fmaxminv): Likewise.
18554 (fmaxmin): Likewise.
18556 (maxmin_uns_op): Likewise.
18558 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18560 * config/aarch64/arm_neon.h
18561 (vac<ge, gt><sd>_f<32, 64>): Rename to...
18562 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
18563 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
18565 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18567 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
18568 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
18570 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18572 * config/aarch64/aarch64-simd.md
18573 (vcond<mode>_internal): Handle special cases for constant masks.
18574 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
18575 (vcondu<mode><mode>): Likewise.
18576 (vcond<v_cmp_result><mode>): New.
18578 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18580 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
18581 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
18582 * config/aarch64/aarch64-simd-builtins.def
18583 (cmeq): Update to BUILTIN_VALLDI.
18588 * config/aarch64/arm_neon.h
18589 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
18590 to builtins or C as appropriate.
18592 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18594 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
18596 (cmhi): Rename to...
18598 * config/aarch64/aarch64-simd.md
18599 (simd_mode): Add SF.
18600 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
18601 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
18602 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
18603 (cstore<mode>_neg): ...This.
18604 * config/aarch64/iterators.md
18606 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
18607 (COMPARISONS): New.
18608 (UCOMPARISONS): Likewise.
18609 (optab): Add missing comparisons.
18615 (VCMP_S): Likewise.
18616 (VCMP_U): Likewise.
18617 (V_cmp_result): Add DF, SF modes.
18618 (v_cmp_result): Likewise.
18620 (vmtype): Likewise.
18621 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
18623 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
18625 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
18626 define_insn to define_insn_and_split.
18627 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
18628 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
18629 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
18630 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
18631 (thumb2_negscc): Likewise.
18633 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18635 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
18637 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18639 * config/arm/thumb2.md: Remove trailing whitespaces.
18641 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
18643 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
18644 Use gen_int_mode rather than GEN_INT.
18646 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
18648 * value-prof.c (stream_in_histogram_value): Remove the strayed
18651 2013-04-30 Richard Biener <rguenther@suse.de>
18653 PR middle-end/57122
18654 * cfghooks.c (split_edge): Properly check for the loop latch edge.
18656 2013-04-30 Richard Biener <rguenther@suse.de>
18658 PR middle-end/57107
18659 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
18661 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
18663 PR rtl-optimization/56957
18664 PR rtl-optimization/57105
18665 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
18666 variable. Use just INSN_UID for determining whether an insn
18667 should be only disconnected from the insn stream.
18668 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
18670 2013-04-30 Jakub Jelinek <jakub@redhat.com>
18672 PR tree-optimization/57104
18673 * tsan.c (instrument_expr): Don't instrument accesses to
18674 DECL_HARD_REGISTER VAR_DECLs.
18676 2013-04-30 Richard Biener <rguenther@suse.de>
18678 * function.h (loops_for_fn): New inline function.
18679 (set_loops_for_fn): Likewise.
18680 * cfgloop.h (place_new_loop): Add struct function parameter.
18681 (get_loop): Likewise.
18682 (get_loops): Likewise.
18683 (number_of_loops): Likewise.
18684 (fel_next): Adjust.
18685 (fel_init): Likewise.
18686 * cfg.c (get_loop_copy): Adjust.
18687 * cfgloop.c (flow_loops_dump): Likewise.
18688 (record_loop_exits): Likewise.
18689 (verify_loop_structure): Likewise.
18690 * cfgloopanal.c (mark_irreducible_loops): Likewise.
18691 (estimate_reg_pressure_cost): Likewise.
18692 (mark_loop_exit_edges): Likewise.
18693 * cfgloopmanip.c (place_new_loop): Likewise.
18694 (add_loop): Likewise.
18695 (duplicate_loop): Likewise.
18696 * graph.c (draw_cfg_nodes): Likewise.
18697 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
18698 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
18699 (extract_affine_chrec): Likewise.
18700 (build_scop_iteration_domain): Likewise.
18701 * graphite.c (graphite_initialize): Likewise.
18702 * ira-build.c (create_loop_tree_nodes): Likewise.
18703 (more_one_region_p): Likewise.
18704 (rebuild_regno_allocno_maps): Likewise.
18705 (mark_loops_for_removal): Likewise.
18706 (mark_all_loops_for_removal): Likewise.
18707 (remove_unnecessary_regions): Likewise.
18708 (ira_build): Likewise.
18709 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
18710 * loop-init.c (fix_loop_structure): Likewise.
18711 (gate_rtl_move_loop_invariants): Likewise.
18712 (gate_rtl_unswitch): Likewise.
18713 (gate_rtl_unroll_and_peel_loops): Likewise.
18714 (rtl_doloop): Likewise.
18715 * lto-streamer-in.c (input_cfg): Likewise.
18716 * lto-streamer-out.c (output_cfg): Likewise.
18717 * modulo-sched.c (sms_schedule): Likewise.
18718 * predict.c (tree_estimate_probability): Likewise.
18719 (tree_estimate_probability_driver): Likewise.
18720 (estimate_loops): Likewise.
18721 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
18722 (move_sese_region_to_fn): Likewise.
18723 (debug_loop_num): Likewise.
18724 * tree-chrec.c (chrec_evaluate): Likewise.
18725 (hide_evolution_in_other_loops_than_loop): Likewise.
18726 (chrec_component_in_loop_num): Likewise.
18727 (reset_evolution_in_loop): Likewise.
18728 (evolution_function_is_invariant_rec_p): Likewise.
18729 * tree-if-conv.c (main_tree_if_conversion): Likewise.
18730 * tree-inline.c (copy_loops): Likewise.
18731 (copy_cfg_body): Likewise.
18732 (tree_function_versioning): Likewise.
18733 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
18734 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
18736 (add_to_evolution_1): Likewise.
18737 (scev_const_prop): Likewise.
18738 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
18739 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
18740 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
18741 (tree_ssa_lim_initialize): Likewise.
18742 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
18743 (verify_loop_closed_ssa): Likewise.
18744 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
18745 (tree_ssa_loop_im): Likewise.
18746 (tree_ssa_loop_unswitch): Likewise.
18747 (tree_vectorize): Likewise.
18748 (check_data_deps): Likewise.
18749 (tree_ssa_loop_ivcanon): Likewise.
18750 (tree_ssa_loop_bounds): Likewise.
18751 (tree_complete_unroll): Likewise.
18752 (tree_complete_unroll_inner): Likewise.
18753 (tree_parallelize_loops): Likewise.
18754 (tree_ssa_loop_prefetch): Likewise.
18755 (tree_ssa_loop_ivopts): Likewise.
18756 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
18757 * tree-vectorizer.c (vectorize_loops): Likewise.
18759 2013-04-29 Mike Frysinger <vapier@gentoo.org>
18761 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
18762 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
18763 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
18764 with EABI_LINK_SPEC.
18766 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18769 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
18772 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
18775 * lra-constraints.c (process_alt_operands): Discourage a bit more
18776 using memory for pseudos. Print cost dump for alternatives.
18777 Modify cost values for conflicts with early clobbers.
18778 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
18780 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18783 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
18785 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18787 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
18788 from/to S register.
18789 (movdi_aarch64): Support LDR/STR from/to D register.
18791 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18793 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
18794 or higher by default.
18796 2013-04-29 Richard Biener <rguenther@suse.de>
18798 PR middle-end/57075
18799 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
18800 even if not adding abnormal edges for calls that can make
18803 2013-04-29 Richard Biener <rguenther@suse.de>
18805 PR middle-end/57103
18806 * tree-cfg.c (move_stmt_op): Fix condition under which to update
18808 (move_stmt_r): Remove redundant checking.
18810 2013-04-29 Teresa Johnson <tejohnson@google.com>
18813 * basic-block.h (apply_scale): New function.
18814 (apply_probability): Use apply_scale.
18815 * gimple-streamer-in.c (input_bb): Ditto.
18816 * lto-streamer-in.c (input_cfg): Ditto.
18817 * lto-cgraph.c (merge_profile_summaries): Ditto.
18818 * tree-optimize.c (execute_fixup_cfg): Ditto.
18819 * tree-inline.c (copy_bb): Update comment to use apply_scale.
18820 (copy_edges_for_bb): Ditto.
18821 (copy_cfg_body): Ditto.
18823 2013-04-29 Tom de Vries <tom@codesourcery.com>
18825 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
18826 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
18827 (tail_merge_optimize): Handle current_loops == NULL.
18829 2013-04-26 Jeff Law <law@redhat.com>
18831 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
18832 (simplify_cond_using_ranges): Generalize code to simplify
18833 COND_EXPRs where one argument is a constant and the other
18834 is an SSA_NAME created by an integral type conversion.
18836 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18838 * config/arm/arm.md (store_minmaxsi): Use only when
18839 optimize_insn_for_size_p.
18841 2013-04-29 Christian Bruel <christian.bruel@st.com>
18844 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
18846 2013-04-29 Richard Biener <rguenther@suse.de>
18848 PR middle-end/57089
18849 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
18850 loop tree make sure to schedule a fixup for the child as well.
18851 (expand_omp_for_generic): Properly add loops.
18852 (expand_omp_for_static_nochunk): Likewise.
18853 (expand_omp_for_static_chunk): Likewise.
18854 (expand_omp_for): For the degenerate case fixup loops.
18855 (expand_omp_sections): Fix default bb placement in loops.
18856 (expand_omp_atomic_pipeline): Properly add loops.
18858 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18860 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
18862 2013-04-29 Tom de Vries <tom@codesourcery.com>
18864 * tree-ssa-tail-merge.c: Update header comment.
18866 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18868 * config/aarch64/arm_neon.h
18869 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
18870 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
18871 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
18872 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
18873 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
18874 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
18875 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
18876 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
18878 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18880 * config/aarch64/aarch64-simd.md
18881 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
18882 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
18883 fix_trunc, fixuns_trunc.
18884 (ftrunc<VDQF:mode>2): New.
18885 * config/aarch64/iterators.md (optab): Add fix, fixuns.
18886 (fix_trunc_optab): New.
18888 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18890 * config/aarch64/aarch64-builtins.c
18891 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
18892 iceilf, lround, iroundf.
18894 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18897 * config/i386/i386.h (enum ix86_tune_indices)
18898 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
18899 New, split from X86_TUNE_INTER_UNIT_MOVES.
18900 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
18901 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
18902 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
18903 (TARGET_INTER_UNIT_MOVES): Remove.
18904 * config/i386/i386.c (initial_ix86_tune_features): Update.
18905 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
18906 (ix86_expand_convert_uns_didf_sse): Use
18907 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18908 (ix86_expand_vector_init_one_nonzero): Ditto.
18909 (ix86_expand_vector_init_interleave): Ditto.
18910 (inline_secondary_memory_needed): Return true for moves from SSE class
18911 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
18912 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
18913 * config/i386/constraints.md (Yi, Ym): Depend on
18914 TARGET_INTER_UNIT_MOVES_TO_VEC.
18915 (Yj, Yn): New constraints.
18916 * config/i386/i386.md (*movdi_internal): Change constraints of
18917 operand 1 from Yi to Yj and from Ym to Yn.
18918 (*movsi_internal): Ditto.
18919 (*movdf_internal): Ditto.
18920 (*movsf_internal): Ditto.
18921 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
18922 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18923 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
18924 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
18925 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
18926 * config/i386/sse.md (movdi_to_sse): Ditto.
18927 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
18928 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
18929 TARGET_INTER_UNIT_MOVES.
18930 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
18931 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
18932 instead of TARGET_INTER_UNIT_MOVES.
18933 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
18934 operand 1 from Yi to Yj and from Ym to Yn.
18936 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18938 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
18939 (float_truncate_hi_): Likewise.
18940 (float_extend_lo_): Likewise.
18941 (float_truncate_lo_): Likewise.
18942 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
18943 (aarch64_float_extend_lo_v2df): Likewise.
18944 (vec_unpacks_hi_v4sf): Likewise.
18945 (aarch64_float_truncate_lo_v2sf): Likewise.
18946 (aarch64_float_truncate_hi_v4sf): Likewise.
18947 (vec_pack_trunc_v2df): Likewise.
18948 (vec_pack_trunc_df): Likewise.
18950 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18952 * config/aarch64/aarch64-builtins.c
18953 (aarch64_fold_builtin): Fold float conversions.
18954 * config/aarch64/aarch64-simd-builtins.def
18955 (floatv2si, floatv4si, floatv2di): New.
18956 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
18957 * config/aarch64/aarch64-simd.md
18958 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
18959 * config/aarch64/iterators.md (FLOATUORS): New.
18960 (optab): Add float, floatuns.
18961 (su_optab): Likewise.
18963 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18965 * config/aarch64/aarch64-builtins.c
18966 (aarch64_builtin_vectorized_function): Use new names for
18968 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
18969 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
18970 (fcvtzu): Split as...
18971 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
18972 (fcvtas): Split as...
18973 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
18974 (fcvtau): Split as...
18975 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
18976 (fcvtps): Split as...
18977 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
18978 (fcvtpu): Split as...
18979 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
18980 (fcvtms): Split as...
18981 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
18982 (fcvtmu): Split as...
18983 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
18984 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
18985 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
18986 (lfrintnusf, lfrintnudf): Likewise.
18987 * config/aarch64/aarch64-simd.md
18988 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
18990 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
18991 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
18992 (fcvt_pattern): Likewise.
18994 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18996 * config/aarch64/aarch64-simd.md
18997 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
18998 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
19000 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19002 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
19003 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
19004 (vrnd<a,m,n,p>_f32): Implement using builtins.
19005 (vrnd<i,x><q>_f<32, 64>): New.
19007 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19009 * config/aarch64/aarch64-builtins.c
19010 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
19011 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
19012 (frintz): Rename to...
19014 (frintp): Rename to...
19016 (frintm): Rename to...
19018 (frinti): Rename to...
19019 (nearbyint): ...this.
19020 (frintx): Rename to...
19022 (frinta): Rename to...
19024 * config/aarch64/aarch64-simd.md
19025 (aarch64_frint<frint_suffix><mode>): Delete.
19026 (<frint_pattern><mode>2): Convert to insn.
19027 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
19028 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
19029 (frint_pattern): Likewise.
19030 (frint_suffix): Likewise.
19032 2013-04-29 Richard Biener <rguenther@suse.de>
19034 PR tree-optimization/57081
19035 * loop-init.c: Include tree-flow.h.
19036 (loop_optimizer_finalize): Free number of iteration estimates.
19037 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
19039 2013-04-29 Jakub Jelinek <jakub@redhat.com>
19041 PR tree-optimization/57083
19042 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
19043 non-singleton shift count range, zero extend low_bound for uns case.
19045 * config/i386/predicates.md (general_vector_operand): New predicate.
19046 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
19047 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
19048 if they aren't nonimmediate operands. If their original values
19049 satisfy const_vector_equal_evenodd_p, don't shift them.
19050 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
19051 predicates. For the SSE4.1 case force operands[{1,2}] into registers
19052 if not nonimmediate_operand.
19053 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
19054 instead of register_operand.
19055 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
19057 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
19059 * stor-layout.c (finalize_size_functions): Allocate a structure and
19060 reset cfun before dumping the functions.
19062 2013-04-27 Jakub Jelinek <jakub@redhat.com>
19064 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
19067 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
19068 use xop_pmacsdqh if uns_p.
19069 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
19070 the immediate rotate count.
19072 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
19074 * rtl.h (struct rtx_def): Add comment for field jump.
19075 (LRA_SUBREG_P): New macro.
19076 * recog.c (register_operand): Check LRA_SUBREG_P.
19077 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19079 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19080 (lra_final_code_change): Skip subreg change for operators.
19081 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19082 if there are no operand changes.
19083 * lra-constraints.c (curr_insn_set): New.
19084 (match_reload): Set LRA_SUBREG_P.
19085 (emit_spill_move): Ditto.
19086 (check_and_process_move): Use curr_insn_set. Process only single
19087 set insns. Don't initialize sec_mem_p and change_p.
19088 (simplify_operand_subreg): Use LRA_SUBREG_P.
19089 (reg_in_class_p): New function.
19090 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19091 of #ifdef. Add code to remove cycling.
19092 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19093 non-null disp. Reload inner instead of disp when base and index
19094 are null. Try to put lo_sum into register.
19095 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19096 (check_and_process_move): Move code for move cost check to
19097 simple_move_p. Remove equiv_substitution.
19098 (simple_move_p): New function.
19099 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19100 curr_insn_set. Call check_and_process_move only for single set
19101 insns. Use the new function. Move call of check_and_process_move
19102 after operand equiv substitution and address process.
19104 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19107 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
19108 with nonlocal goto receivers or returns twice calls, ignore
19109 unininitialized values from abnormal edges to nl goto receiver
19110 or returns twice call.
19112 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19114 PR tree-optimization/57051
19115 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
19116 and VEC_RSHIFT_EXPR if shift count is a multiple of element
19119 2013-04-26 Richard Biener <rguenther@suse.de>
19121 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
19122 (expand_omp_taskreg): Likewise. Mark loops for fixup.
19123 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
19124 (fixup_loop_arrays_after_move): New function.
19125 (move_sese_region_to_fn): Properly outline the loop tree parts
19126 of the SESE region.
19128 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
19130 * config/i386/i386.md (type, unit): Fix long lines.
19132 2013-04-26 Richard Biener <rguenther@suse.de>
19134 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
19135 (lto-streamer-out.o): Likewise.
19136 * cfgloop.c (init_loops_structure): Export, add struct function
19137 argument and adjust.
19138 (flow_loops_find): Adjust.
19139 * cfgloop.h (enum loop_estimation): Add EST_LAST.
19140 (init_loops_structure): Declare.
19141 * lto-streamer-in.c: Include cfgloop.h.
19142 (input_cfg): Input the loop tree.
19143 * lto-streamer-out.c: Include cfgloop.h.
19144 (output_cfg): Output the loop tree.
19145 (output_struct_function_base): Do not drop PROP_loops.
19147 2013-03-26 Richard Biener <rguenther@suse.de>
19149 * tree-cfg.c (execute_build_cfg): Build the loop tree.
19150 (pass_build_cfg): Provide PROP_loops.
19151 (move_sese_region_to_fn): Remove loops that are outlined into fn
19153 * tree-inline.c: Include cfgloop.h.
19154 (initialize_cfun): Do not drop PROP_loops.
19155 (copy_loops): New function.
19156 (copy_cfg_body): Copy loop structure.
19157 (tree_function_versioning): Initialize destination loop tree.
19158 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
19159 (pass_parallelize_loops): Do IL verification.
19160 * loop-init.c (loop_optimizer_init): Fixup loops if required.
19161 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
19162 the CFG make sure we fixup loops as well.
19163 * tree-ssa-tail-merge.c: Include cfgloop.h.
19164 (replace_block_by): When merging loop latches mark loops for fixup.
19165 * lto-streamer-out.c (output_struct_function_base): Drop
19166 PROP_loops for now.
19167 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
19168 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
19169 * ipa-split.c: Include cfgloop.h.
19170 (split_function): Add the new return block to the loop tree root.
19171 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
19172 whether we have removed the forwarder block.
19173 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
19174 * cfgloop.h (place_new_loop): Declare.
19175 * cfgloopmanip.c (place_new_loop): Export.
19176 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
19177 (tree-switch-conversion.o): Likewise.
19178 (tree-complex.o): Likewise.
19179 (tree-inline.o): Likewise.
19180 (tree-ssa-tailmerge.o): Likewise.
19181 (ipa-split.o): Likewise.
19182 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
19183 (tree-ssa-copy.o): Likewise.
19184 * tree-switch-conversion.c: Include cfgloop.h
19185 (process_switch): If we emit a bit-test cascade, schedule loops
19187 * tree-complex.c: Include cfgloop.h.
19188 (expand_complex_div_wide): Properly add new basic-blocks to loops.
19189 * asan.c: Include cfgloop.h.
19190 (create_cond_insert_point): Properly add new basic-blocks to
19191 loops, schedule loop fixup.
19192 * cfgloop.c (verify_loop_structure): Check that looks are not
19194 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
19196 (expand_omp_for_generic): Likewise.
19197 (expand_omp_sections): Likewise.
19198 (expand_omp_atomic_pipeline): Schedule loops for fixup.
19199 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
19200 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
19201 is initialized, not when loops are present.
19202 * tree-parloops.c (parallelize_loops): Remove checking here.
19203 * passes.c (init_optimization_passes): Schedule a copy-propagation
19204 pass before complete unrolling of inner loops.
19206 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19208 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
19209 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
19210 (colorize_init): Add argument to _WIN32 version.
19211 * toplev.c: Include diagnostic-color.h.
19212 (process_options): Default to -fdiagnostics-color=auto if
19213 GCC_COLORS env var is in the environment.
19214 * common.opt (fdiagnostics-color=): Add Var and Init.
19215 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
19216 env var is in the environment, the default is auto rather than never.
19218 * diagnostic.h (file_name_as_prefix): Add context argument.
19219 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
19220 the string as locus.
19221 * langhooks.c (lhd_print_error_function): Adjust caller.
19223 2013-04-25 Lawrence Crowl <crowl@google.com>
19225 * var-tracking.c (shared_hash_def::htab):
19226 Change type to hash_table. Update dependent calls and types.
19228 2013-04-25 Lawrence Crowl <crowl@google.com>
19230 * Makefile.in: Update as needed below.
19232 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
19233 Move declaration to after the type's method definitons.
19235 * attribs.c (htab_t scoped_attributes::attribute_hash):
19236 Change type to hash_table. Update dependent calls and types.
19238 * bitmap.c (htab_t bitmap_desc_hash):
19239 Change type to hash_table. Update dependent calls and types.
19241 * cselib.c (htab_t cselib_hash_table):
19242 Change type to hash_table. Update dependent calls and types.
19244 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
19245 (hash_string_slot_node): Move implementation into lto-streamer.h
19246 struct string_slot_hasher.
19247 (eq_string_slot_node): Likewise.
19249 * data-streamer-out.c: Update output_block::string_hash_table
19250 dependent calls and types.
19252 * dwarf2cfi.c (htab_t trace_index):
19253 Change type to hash_table. Update dependent calls and types.
19255 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
19256 Change type to hash_table. Update dependent calls and types.
19257 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
19258 (htab_t optimize_external_refs::map): Likewise.
19259 (htab_t output_comp_unit::extern_map): Likewise.
19260 (htab_t output_comdat_type_unit::extern_map): Likewise.
19261 (htab_t output_macinfo::macinfo_htab): Likewise.
19262 (htab_t optimize_location_lists::htab): Likewise.
19263 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
19265 * except.c (htab_t ehspec_hash_type):
19266 Change type to hash_table. Update dependent calls and types.
19267 (assign_filter_values::ttypes): Likewise.
19268 (assign_filter_values::ehspec): Likewise.
19269 (sjlj_assign_call_site_values::ar_hash): Likewise.
19270 (convert_to_eh_region_ranges::ar_hash): Likewise.
19272 * gcse.c (htab_t pre_ldst_table):
19273 Change type to hash_table. Update dependent calls and types.
19275 * ggc-common.c (htab_t saving_htab):
19276 Change type to hash_table. Update dependent calls and types.
19277 (htab_t loc_hash): Likewise.
19278 (htab_t ptr_hash): Likewise.
19279 (call_count): Rename ggc_call_count.
19280 (call_alloc): Rename ggc_call_alloc.
19281 (loc_descriptor): Rename make_loc_descriptor.
19282 (add_statistics): Rename ggc_add_statistics.
19284 * ggc-common.c (saving_htab):
19285 Change type to hash_table. Update dependent calls and types.
19287 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
19288 (push_gimplify_context): Likewise.
19289 (pop_gimplify_context): Likewise.
19290 (struct gimple_temp_hash_elt): Added.
19291 (struct gimplify_hasher): Likewise.
19292 (struct gimplify_ctx.temp_htab):
19293 Change type to hash_table. Update dependent calls and types.
19295 * gimple-fold.c: Include gimplify-ctx.h.
19297 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
19298 Change type to hash_table. Update dependent calls and types.
19299 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
19300 avoid potential global name collision.
19302 * gimplify.c: Include gimplify-ctx.h.
19303 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
19304 (htab_t gimplify_ctx::temp_htab):
19305 Update dependent calls and types for new type hash_table.
19306 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
19307 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
19309 * gimplify-ctx.h: New.
19310 (struct gimple_temp_hash_elt): Move from gimplify.c.
19311 (class gimplify_hasher): New.
19312 (struct gimplify_ctx): Move from gimple.h.
19313 (htab_t gimplify_ctx::temp_htab):
19314 Change type to hash_table. Update dependent calls and types.
19316 * graphite-clast-to-gimple.c: Include graphite-htab.h.
19317 (htab_t ivs_params::newivs_index):
19318 Change type to hash_table. Update dependent calls and types.
19319 (htab_t ivs_params::params_index): Likewise.
19320 (htab_t print_generated_program::params_index): Likewise.
19321 (htab_t gloog::newivs_index): Likewise.
19322 (htab_t gloog::params_index): Likewise.
19324 * graphite.c: Include graphite-htab.h.
19325 4htab_t graphite_transform_loops::bb_pbb_mapping):
19326 Change type to hash_table. Update dependent calls and types.
19328 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
19329 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
19330 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
19332 * graphite-dependences.c: Include graphite-htab.h.
19333 (loop_is_parallel_p): Change hash table type of parameter.
19335 * graphite-htab.h: New.
19336 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
19337 (extern find_pbb_via_hash): Move from graphite-poly.h.
19338 (extern loop_is_parallel_p): Move from graphite-poly.h.
19339 (extern get_loop_body_pbbs): Move from graphite-poly.h.
19341 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
19342 (extern loop_is_parallel_p): Move to graphite-htab.h.
19343 (extern get_loop_body_pbbs): Move to graphite-htab.h.
19345 * haifa-sched.c (htab_t delay_htab):
19346 Change type to hash_table. Update dependent calls and types.
19347 (htab_t delay_htab_i2): Likewise.
19349 * ira-color.c (htab_t allocno_hard_regs_htab):
19350 Change type to hash_table. Update dependent calls and types.
19352 * ira-costs.c (htab_t cost_classes_htab):
19353 Change type to hash_table. Update dependent calls and types.
19355 * loop-invariant.c (htab_t merge_identical_invariants::eq):
19356 Change type to hash_table. Update dependent calls and types.
19358 * loop-iv.c (htab_t bivs):
19359 Change type to hash_table. Update dependent calls and types.
19361 * loop-unroll.c (htab_t opt_info::insns_to_split):
19362 Change type to hash_table. Update dependent calls and types.
19363 (htab_t opt_info::insns_with_var_to_expand): Likewise.
19365 * lto-streamer.h (struct string_slot): Move from data-streamer.h
19366 (struct string_slot_hasher): New.
19367 (htab_t output_block::string_hash_table):
19368 Change type to hash_table. Update dependent calls and types.
19370 * lto-streamer-in.c (freeing_string_slot_hasher): New.
19371 (htab_t file_name_hash_table):
19372 Change type to hash_table. Update dependent calls and types.
19374 * lto-streamer-out.c: Update output_block::string_hash_table dependent
19377 * lto-streamer.c (htab_t tree_htab):
19378 Change type to hash_table. Update dependent calls and types.
19380 * omp-low.c: Include gimplify-ctx.h.
19382 * passes.c (htab_t name_to_pass_map):
19383 Change type to hash_table. Update dependent calls and types.
19384 (pass_traverse): Rename to passes_pass_traverse.
19386 * plugin.c (htab_t event_tab):
19387 Change type to hash_table. Update dependent calls and types.
19389 * postreload-gcse.c (htab_t expr_table):
19390 Change type to hash_table. Update dependent calls and types.
19391 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
19393 * sese.c (debug_rename_map_1): Make extern.
19394 (htab_t copy_bb_and_scalar_dependences::rename_map):
19395 Change type to hash_table. Update dependent calls and types.
19397 * sese.h (extern debug_rename_map): Move to .c file.
19399 * store-motion.c (htab_t store_motion_mems_table):
19400 Change type to hash_table. Update dependent calls and types.
19402 * trans-mem.c (htab_t tm_new_mem_hash):
19403 Change type to hash_table. Update dependent calls and types.
19405 * tree-browser.c (htab_t TB_up_ht):
19406 Change type to hash_table. Update dependent calls and types.
19408 * tree-cfg.c (htab_t discriminator_per_locus):
19409 Change type to hash_table. Update dependent calls and types.
19411 * tree-complex.c: Include tree-hasher.h
19412 (htab_t complex_variable_components):
19413 Change type to hash_table. Update dependent calls and types.
19415 * tree-eh.c (htab_t finally_tree):
19416 Change type to hash_table. Update dependent calls and types.
19418 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
19419 struct int_tree_hasher.
19420 (extern int_tree_map_eq): Likewise.
19421 (uid_decl_map_hash): Removed.
19422 (extern decl_tree_map_eq): Likewise.
19424 * tree-hasher.h: New.
19425 (struct int_tree_hasher): New.
19426 (typedef int_tree_htab_type): New.
19428 * tree-inline.c: Include gimplify-ctx.h.
19430 * tree-mudflap.c: Include gimplify-ctx.h.
19432 * tree-parloops.c: Include tree-hasher.h.
19433 (htab_t eliminate_local_variables_stmt::decl_address):
19434 Change type to hash_table. Update dependent calls and types.
19435 (htab_t separate_decls_in_region::decl_copies): Likewise.
19437 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
19438 Change type to hash_table. Update dependent calls and types.
19440 * tree-sra.c (candidates):
19441 Change type to hash_table. Update dependent calls and types.
19443 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
19445 (int_tree_map_hash): Likewise.
19447 * tree-ssa-dom.c (htab_t avail_exprs):
19448 Change type to hash_table. Update dependent calls and types.
19450 * tree-ssa-live.c (var_map_base_init::tree_to_index):
19451 Change type to hash_table. Update dependent calls and types.
19453 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
19454 Change type to hash_table. Update dependent calls and types.
19456 * tree-ssa-phiopt.c (seen_ssa_names):
19457 Change type to hash_table. Update dependent calls and types.
19459 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
19460 Change type to hash_table. Update dependent calls and types.
19462 * tree-ssa-uncprop.c (equiv):
19463 Change type to hash_table. Update dependent calls and types.
19465 2013-04-25 Jakub Jelinek <jakub@redhat.com>
19467 PR rtl-optimization/57003
19468 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
19469 call note_stores with kill_clobbered_value callback again after
19470 killing regs_invalidated_by_call.
19472 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19474 * config/aarch64/aarch64-simd.md
19475 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
19476 (aarch64_simd_bsl<mode>): Likewise.
19477 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
19479 2013-04-25 Marek Polacek <polacek@redhat.com>
19481 PR tree-optimization/57066
19482 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
19484 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
19486 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
19488 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19490 * config/aarch64/aarch64-builtins.c
19491 (aarch64_fold_builtin): New.
19492 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
19493 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
19494 * config/aarch64/aarch64-simd-builtins.def (abs): New.
19495 * config/aarch64/arm_neon.h
19496 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
19498 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19499 Tejas Belagod <tejas.belagod@arm.com>
19501 * config/aarch64/aarch64-builtins.c
19502 (aarch64_gimple_fold_builtin): New.
19503 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
19504 * config/aarch64/aarch64-simd-builtins.def (addv): New.
19505 * config/aarch64/aarch64-simd.md (addpv4sf): New.
19506 (addvv4sf): Update.
19507 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
19509 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19511 * config/aarch64/aarch64.md
19512 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
19514 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19516 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
19517 (*ngcsi_uxtw): New pattern.
19519 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19520 Julian Brown <julian@codesourcery.com>
19522 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
19523 (TB_DREG): Add T_V4HF.
19524 (v4hf_UP): New macro.
19525 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19526 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19527 Handle initialisation of V4HF. Adjust initialisation of reinterpret
19529 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19530 (arm_vector_mode_supported_p): Handle V4HF.
19531 (arm_mangle_map): Handle V4HFmode.
19532 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
19533 * config/arm/arm_neon_builtins.def: Add entries for
19534 vcvtv4hfv4sf, vcvtv4sfv4hf.
19535 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
19536 (neon_vcvtv4hfv4sf): Likewise.
19537 * config/arm/neon-gen.ml: Handle half-precision floating point
19539 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
19540 * config/arm/arm_neon.h: Regenerate.
19541 * config/arm/neon.ml (type elts): Add F16.
19542 (type vectype): Add T_float16x4, T_floatHF.
19543 (type vecmode): Add V4HF.
19544 (type features): Add Requires_FP_bit feature.
19545 (elt_width): Handle F16.
19546 (elt_class): Likewise.
19547 (elt_of_class_width): Likewise.
19548 (mode_of_elt): Refactor.
19549 (type_for_elt): Handle F16, fix error messages.
19550 (vectype_size): Handle T_float16x4.
19551 (vcvt_sh): New function.
19552 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
19553 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
19554 (string_of_mode): Handle V4HF.
19555 * doc/arm-neon-intrinsics.texi: Regenerate.
19557 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19559 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
19560 format specifier in 'X' case.
19562 2013-04-25 Alan Modra <amodra@gmail.com>
19565 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
19566 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
19567 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
19568 Repeat for many other rotate/shift and mask patterns using subregs.
19569 Name lshiftrt insns.
19570 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
19571 on WORDS_BIG_ENDIAN.
19573 2013-04-25 Alan Modra <amodra@gmail.com>
19575 * config.gcc: Support little-endian powerpc-linux targets.
19576 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
19577 (LINK_OS_LINUX_SPEC): Define.
19578 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
19579 Preserve MASK_LITTLE_ENDIAN.
19580 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
19581 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
19582 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
19583 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
19584 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
19585 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
19586 Correct fp word order for little-endian. Don't shift toc entries
19587 smaller than a word for little-endian.
19588 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
19589 (bswapdi2 splits): Correct low-part subreg for little-endian.
19590 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
19591 low/high where such is correct only for be.
19592 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
19593 little-endian for -mcall-aixdesc.
19595 2013-04-25 Alan Modra <amodra@gmail.com>
19597 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
19598 replace_equiv_address_nv.
19600 2013-04-25 Alan Modra <amodra@gmail.com>
19602 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
19604 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19607 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19608 * rtl.h (struct rtx_def): ...
19610 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19612 PR rtl-optimizations/57046
19613 * lra-constraints (split_reg): Set up lra_risky_transformations_p
19614 for multi-reg splits.
19616 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
19618 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
19620 2013-04-24 Sterling Augustine <saugustine@google.com>
19622 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
19623 (comp_dir_string, debug_str_dwo_section): New.
19624 (DEBUG_STR_DWO_SECTION): Rename to ...
19625 (DEBUG_DWO_STR_SECTION): ... this.
19626 (DEBUG_NORM_STR_SECTION): Delete.
19627 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
19628 (DEBUG_STR_DWO_SECTION_FLAGS): New.
19629 (find_AT_string): Move most logic to ...
19630 (find_AT_string_in_table): ... here. New.
19631 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
19632 add_skeleton_AT_string. Delete logic.
19633 (output_skeleton_debug_sections): Remove call to
19634 add_top_level_skeleton_die_attrs.
19635 (add_comp_dir_attribute): Move logic to comp_dir_string.
19636 (dwarf2out_init): Initialize debug_str_dwo_section.
19637 (output_indirect_string): Call find_string_form.
19638 (output_indirect_strings): Rewrite.
19639 (prune_unused_types): Empty skeleton_debug_str_hash.
19640 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
19641 (dwarf2out_finish): Call output_indirect_strings.
19643 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19645 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
19647 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19649 * rtl.h (struct rtx_def): Add comment for field jump.
19650 (LRA_SUBREG_P): New macro.
19651 * recog.c (register_operand): Check LRA_SUBREG_P.
19652 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19654 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19655 (lra_final_code_change): Skip subreg change for operators.
19656 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19657 if there are no operand changes.
19658 * lra-constraints.c (curr_insn_set): New.
19659 (match_reload): Set LRA_SUBREG_P.
19660 (emit_spill_move): Ditto.
19661 (check_and_process_move): Use curr_insn_set. Process only single
19662 set insns. Don't initialize sec_mem_p and change_p.
19663 (simplify_operand_subreg): Use LRA_SUBREG_P.
19664 (reg_in_class_p): New function.
19665 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19666 of #ifdef. Add code to remove cycling.
19667 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19668 non-null disp. Reload inner instead of disp when base and index
19669 are null. Try to put lo_sum into register.
19670 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19671 (check_and_process_move): Move code for move cost check to
19672 simple_move_p. Remove equiv_substitution.
19673 (simple_move_p): New function.
19674 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19675 curr_insn_set. Call check_and_process_move only for single set
19676 insns. Use the new function. Move call of check_and_process_move
19677 after operand equiv substitution and address process.
19679 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
19681 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
19682 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
19683 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
19685 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19687 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
19689 2013-04-24 Marek Polacek <polacek@redhat.com>
19691 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
19692 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
19693 (select_loops_exit_conditions): Likewise.
19694 (number_of_iterations_for_all_loops): Likewise.
19695 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
19696 (scev_analysis): Likewise.
19698 2013-04-02 Catherine Moore <clm@codesourcery.com>
19699 Chao-ying Fu <fu@mips.com>
19701 * config/mips/micromips.md (jraddiusp): New pattern.
19702 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
19703 instruction if possible.
19705 2013-04-24 Alan Modra <amodra@gmail.com>
19707 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
19709 2013-04-24 Julian Brown <julian@codesourcery.com>
19710 Chung-Lin Tang <cltang@codesourcery.com>
19712 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
19713 dependency behavior in enumeration type DIE generation. Add TODO note
19714 to comments about future DW_FORM_sdata/udata re-work of related code.
19716 2013-04-23 Lawrence Crowl <crowl@google.com>
19718 * Makefile.in: Update as needed below.
19720 * hash-table.h (class hash_table):
19721 Correct many methods with parameter types compare_type to the correct
19722 value_type. (Correct code was unlikely to notice the change.)
19723 (hash_table::elements_with_deleted) New.
19724 (class hashtable::iterator): New.
19725 (hashtable::begin()): New.
19726 (hashtable::end()): New.
19727 (FOR_EACH_HASH_TABLE_ELEMENT): New.
19729 * statistics.c (statistics_hashes):
19730 Change type to hash_table. Update dependent calls and types.
19732 * tree-into-ssa.c (var_infos):
19733 Change type to hash_table. Update dependent calls and types.
19735 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
19736 Change type to hash_table. Update dependent calls and types.
19738 * tree-ssa-loop-im.c (struct mem_ref.refs):
19739 Change type to hash_table. Update dependent calls and types.
19741 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
19742 Change type to hash_table. Update dependent calls and types.
19744 * tree-ssa-sccvn.c (vn_tables_s::nary):
19745 Change type to hash_table. Update dependent calls and types.
19746 (vn_tables_s::phis): Likewise.
19747 (vn_tables_s::references): Likewise.
19749 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
19750 (vn_reference_eq): Update parameter and return types.
19752 * tree-ssa-structalias.c (pointer_equiv_class_table):
19753 Change type to hash_table. Update dependent calls and types.
19754 (location_equiv_class_table): Likewise.
19756 * tree-vect-data-refs.c: Consequential changes for making
19757 peeling a hash_table.
19759 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
19760 (destroy_loop_vec_info): Dependent hash_table update.
19762 * tree-vectorizer.h (peeling_htab):
19763 Change type to hash_table. Update dependent calls and types.
19765 2013-04-23 Shiva Chen <shiva0217@gmail.com>
19767 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
19768 to check the register content is equal or not.
19769 * lra-constraints.c (match_reload): Use lra_assign_reg_val
19770 to assign register content record.
19771 * lra-eliminations.c (update_reg_eliminate): Use
19772 lra_update_reg_val_offset to update register content offset.
19773 * lra-int.h (struct lra_reg): Add offset member.
19774 (lra_reg_val_equal_p): New static inline function.
19775 (lra_update_reg_val_offset): New static inline function.
19776 (lra_assign_reg_val): New static inline function.
19777 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
19778 to assign register content record.
19779 (initialize_lra_reg_info_element): Initial offset to zero.
19781 2013-04-23 Catherine Moore <clm@codesourcery.com>
19783 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
19784 operands. Record compression.
19786 2013-04-23 Xinliang David Li <davidxl@google.com>
19788 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
19790 2013-04-23 Richard Biener <rguenther@suse.de>
19792 PR middle-end/57036
19793 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
19794 parameter, only add abnormal goto edges from the copied body
19795 if the call could perform abnormal gotos.
19796 (copy_cfg_body): Adjust.
19798 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
19800 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
19802 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
19804 * coretypes.h (gimple_stmt_iterator): Add struct to make
19807 2013-04-23 Richard Biener <rguenther@suse.de>
19809 PR tree-optimization/57026
19810 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
19811 from SSA names occuring in abnormal PHI nodes.
19813 2013-04-22 Andi Kleen <ak@linux.intel.com>
19815 * lto/lto.c (print_lto_report_1): Fix LTO report names.
19817 2013-04-22 Andi Kleen <ak@linux.intel.com>
19819 * lto/lto.c (print_lto_report_1): Declare early.
19820 (read_cgraph_and_symbols): Call print_lto_report_1 early.
19822 2013-04-22 Andi Kleen <ak@linux.intel.com>
19824 * common.opt (-flto-report-wpa): Add.
19825 * doc/invoke.texi (-flto-report-wpa): Add.
19826 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
19829 2013-04-22 Xinliang David Li <davidxl@google.com>
19831 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
19832 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
19833 * Makefile.in: New dependency
19835 David Daney <ddaney.cavm@gmail.com>
19837 * configure.ac (gcc_cv_as_micromips_support): Use the
19838 --fatal-warnings option.
19839 * configure: Regenerate.
19841 2013-04-22 Marek Polacek <polacek@redhat.com>
19844 * tsan.c (instrument_expr): Don't instrument expression
19845 in case its size is zero.
19847 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
19851 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19853 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19855 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19857 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
19858 (gimple_stmt_iterator): New typedef.
19859 * gimple.h (gimple_stmt_iterator): Rename to...
19860 (gimple_stmt_iterator_d): ... This.
19861 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
19862 trees be valid for GIMPLE and GENERIC.
19863 (TARGET_GIMPLE_FOLD_BUILTIN): New.
19864 * gimple-fold.c (gimple_fold_call): Call target hook
19865 gimple_fold_builtin.
19866 * hooks.c (hook_bool_gsiptr_false): New.
19867 * hooks.h (hook_bool_gsiptr_false): New.
19868 * target.def (fold_stmt): New.
19869 * doc/tm.texi: Regenerate.
19871 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
19874 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
19875 a set sp if no stack realignment.
19877 2013-04-22 Nick Clifton <nickc@redhat.com>
19879 * config.gcc (tilegx-linux): Extend extra_objs rather than
19881 (tilepro-linux): Likewise.
19883 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19885 * config/aarch64/aarch64-builtins.c
19887 (CF0, CF1, CF2, CF3, CF4, CF10): New.
19888 (VAR<1-12>): Add MAP parameter.
19889 (BUILTIN_*): Likewise.
19890 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
19891 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
19892 (aarch64_ushl_n<mode>): Likewise.
19893 (aarch64_sshr_n<mode>): Likewise.
19894 (aarch64_ushr_n<mode>): Likewise.
19895 (aarch64_<maxmin><mode>): Likewise.
19896 (aarch64_sqrt<mode>): Likewise.
19897 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
19898 (vshr<q>_n_*): Likewise.
19900 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19902 * config/aarch64/aarch64-builtins.c
19903 (aarch64_simd_builtin_type_mode): Handle SF types.
19905 (BUILTIN_GPF): Define.
19906 (aarch64_init_simd_builtins): Handle SF types.
19907 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
19908 (frecps): Likewise.
19909 (frecpx): Likewise.
19910 * config/aarch64/aarch64-simd.md
19911 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
19912 (aarch64_frecpe<mode>): New.
19913 (aarch64_frecps<mode>): Likewise.
19914 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
19915 (v8type): Add frecp<esx>.
19916 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
19917 (aarch64_frecps<mode>): Likewise.
19918 * config/aarch64/iterators.md (FRECP): New.
19919 (frecp_suffix): Likewise.
19920 * config/aarch64/arm_neon.h
19921 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
19923 2013-04-22 Christian Bruel <christian.bruel@st.com>
19926 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
19927 (REG_CLASS_NAMES): Idem.
19928 (REG_CLASS_CONTENTS): Idem.
19929 (REGCLASS_HAS_FP_REG): Idem.
19930 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
19931 (sh_conditional_register_usage): Idem.
19933 2013-04-21 Jeff Law <law@redhat.com>
19935 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
19936 (ssa_forward_propagate_and_combine): Use it.
19938 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19940 * lra.c: Update the flow chart diagram.
19942 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19944 PR rtl-optimization/56847
19945 * lra-constraints.c (process_alt_operands): Discourage alternative
19946 with non-matche doffsettable memory constraint fro memory with
19949 2013-04-19 Richard Biener <rguenther@suse.de>
19951 PR tree-optimization/56982
19952 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
19954 * gimplify.c (gimplify_call_expr): Notice special calls.
19955 (gimplify_modify_expr): Likewise.
19956 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
19957 abnormal control flow receivers.
19958 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
19959 in the same way as cfun->has_nonlocal_labels.
19960 (gimple_purge_dead_abnormal_call_edges): Likewise.
19961 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
19962 receivers start a basic-block.
19964 2013-04-19 Richard Biener <rguenther@suse.de>
19966 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
19968 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
19969 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
19970 (SLP_TREE_LOAD_PERMUTATION): Add.
19971 (vect_transform_slp_perm_load): Adjust prototype.
19972 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
19973 (vect_free_slp_instance): Likewise.
19974 (vect_create_new_slp_node): Likewise.
19975 (vect_supported_slp_permutation_p): Remove.
19976 (vect_slp_rearrange_stmts): Adjust.
19977 (vect_supported_load_permutation_p): Likewise. Inline
19978 vect_supported_slp_permutation_p here.
19979 (vect_analyze_slp_instance): Compute load permutations per
19980 slp node instead of per instance.
19981 (vect_get_slp_defs): Adjust.
19982 (vect_transform_slp_perm_load): Likewise.
19983 (vect_schedule_slp_instance): Remove redundant code.
19984 (vect_schedule_slp): Remove hack for PR56270, add it ...
19985 * tree-vect-stmts.c (vectorizable_load): ... here, do not
19986 CSE loads for SLP. Adjust.
19988 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19990 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
19991 spelling in two comments.
19993 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19996 * config/arm/arm.c (load_multiple_sequence): Require SP
19997 as base register for loads if SP is in the register list.
19999 2013-04-19 Martin Jambor <mjambor@suse.cz>
20001 PR tree-optimization/56718
20002 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
20003 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
20004 and made public. Adjusted all callers.
20005 (ipa_intraprocedural_devirtualization): New function.
20006 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
20007 (ipa_intraprocedural_devirtualization): Likewise.
20008 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
20010 2013-04-19 Richard Biener <rguenther@suse.de>
20012 PR tree-optimization/57000
20013 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
20015 2013-04-19 Terry Guo <terry.guo@arm.com>
20017 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
20019 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
20020 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
20021 (cortex_m4_fmacs): Use new reservations.
20022 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
20024 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
20026 PR rtl-optimization/56999
20027 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
20029 (lra_coalesce): Remove split_origin_bitmap and related code.
20030 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
20031 ranges if necessary.
20033 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
20035 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
20037 (ix86_expand_call): Remove clobbered_registers array and use
20038 x86_64_ms_sysv_extra_clobbered_registers instead.
20039 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
20041 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
20043 * config/i386/i386.md (*call_rex64_ms_sysv): Use
20044 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
20045 (*call_value_rex64_ms_sysv): Ditto.
20047 2013-04-18 Cary Coutant <ccoutant@google.com>
20049 * dwarf2out.c (output_pubnames): Check die_perennial_p of
20050 parent instead of die_mark.
20052 2013-04-18 Diego Novillo <dnovillo@google.com>
20054 * gimple.c (create_gimple_tmp): New.
20055 (get_expr_type): New.
20056 (build_assign): New.
20057 (build_type_cast): New.
20058 * gimple.h (enum ssa_mode): Define.
20059 (gimple_seq_set_location): New.
20060 * asan.c (build_check_stmt): Change some gimple_build_* calls
20061 to use build_assign and build_type_cast.
20063 2013-04-18 Richard Biener <rguenther@suse.de>
20065 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
20066 handle negative step. Remove redundant checks.
20067 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
20068 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
20069 for negative step and grouped loads fail to vectorize.
20071 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
20073 * emit-rtl.c (reset_insn_used_flags): New function.
20074 (reset_all_used_flags): Use it.
20075 (verify_insn_sharing): New function.
20076 (verify_rtl_sharing): Fix verification for SEQUENCEs.
20078 2013-04-18 Jakub Jelinek <jakub@redhat.com>
20080 PR tree-optimization/56984
20081 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
20082 and (x >> M) >= N don't register any assertion if N << M is the
20085 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
20087 * lower-subreg.c (resolve_simple_move): If called self-recursive,
20088 do not delete_insn insns that have not yet been emitted, only
20089 unlink them with remove_insn.
20090 * df-scan.c (df_insn_delete): Revert r197492.
20092 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
20094 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
20095 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
20097 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
20099 * config/arm/arm.md (movsicc_insn): Convert define_insn into
20100 define_insn_and_split.
20101 (and_scc,ior_scc,negscc): Likewise.
20102 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
20104 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
20106 * config/arm/arm.c (use_return_insn): Return 0 for targets that
20107 can benefit from using a sequence of LDRD instructions in epilogue
20108 instead of a single LDM instruction.
20110 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
20113 * doc/extend.texi: Fix typo.
20115 2013-04-17 Richard Biener <rguenther@suse.de>
20117 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
20118 (vect_build_slp_tree): ... here.
20119 (vect_build_slp_tree_1): Compute which stmts of the SLP group
20120 match. Remove special-casing of mismatched complex loads.
20121 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
20122 re-try the match with swapped commutative operands.
20123 (vect_supported_load_permutation_p): Remove special-casing of
20124 mismatched complex loads.
20125 (vect_analyze_slp_instance): Adjust.
20127 2013-04-17 Richard Biener <rguenther@suse.de>
20129 PR rtl-optimization/56921
20130 * cfgloop.h (struct loop): Add simple_loop_desc member.
20131 (struct niter_desc): Mark with GTY(()).
20132 (simple_loop_desc): Do not use aux field but simple_loop_desc.
20133 * loop-iv.c (get_simple_loop_desc): Likewise.
20134 (free_simple_loop_desc): Likewise.
20137 2013-04-16 Richard Biener <rguenther@suse.de>
20139 PR rtl-optimization/56921
20140 * loop-init.c (pass_rtl_move_loop_invariants): Add
20141 TODO_do_not_ggc_collect to todo_flags_finish.
20142 (pass_rtl_unswitch): Same.
20143 (pass_rtl_unroll_and_peel_loops): Same.
20144 (pass_rtl_doloop): Same.
20146 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
20148 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
20149 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
20150 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
20152 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
20153 * tree-streamer.c (record_common_node): Adjust reference in comment.
20155 2013-04-17 Terry Guo <terry.guo@arm.com>
20157 * config/arm/cortex-m4.md: Add a new bypass.
20159 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20161 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
20163 (*subs_<optab><mode>_multp2): New pattern.
20164 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
20165 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
20167 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20169 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
20170 (*subs_mul_imm_<mode>): New pattern.
20172 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
20175 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
20176 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
20177 (vsx_movti_32bit): Change j->wa to O->wa.
20179 2013-04-16 Richard Biener <rguenther@suse.de>
20181 PR rtl-optimization/56921
20182 * loop-init.c (pass_rtl_move_loop_invariants): Add
20183 TODO_do_not_ggc_collect to todo_flags_finish.
20184 (pass_rtl_unswitch): Same.
20185 (pass_rtl_unroll_and_peel_loops): Same.
20186 (pass_rtl_doloop): Same.
20188 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
20190 * config/arm/arm.c (emit_multi_reg_push): New declaration
20191 for an existing function.
20192 (arm_emit_strd_push): New function.
20193 (arm_expand_prologue): Used here.
20194 (arm_emit_ldrd_pop): New function.
20195 (arm_expand_epilogue): Used here.
20196 (arm_get_frame_offsets): Update condition.
20197 (arm_emit_multi_reg_pop): Add a special case for load of a single
20198 register with writeback.
20200 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
20202 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
20205 2013-04-16 Richard Biener <rguenther@suse.de>
20207 PR tree-optimization/56756
20208 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
20209 (first_mem_ref_loc): New.
20210 (execute_sm): Place the load temporarily before a previous
20211 access instead of in the latch edge to ensure its SSA dependencies
20212 are defined at points dominating the load.
20214 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
20216 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
20217 correct fix by moving header and footer insn to the footer of
20218 the merged basic block. Clear BB_END of the merged-away block.
20220 PR middle-end/43631
20221 * emit-rtl.c (make_note_raw): New function.
20222 (link_insn_into_chain): New static inline function.
20223 (add_insn): Use it.
20224 (add_insn_before, add_insn_after): Factor insn chain linking code...
20225 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
20226 using link_insn_into_chain.
20227 (note_outside_basic_block_p): New helper function for emit_note_after
20228 and emit_note_before.
20229 (emit_note_after): Use nobb variant of add_insn_after if the note
20230 should not be contained in a basic block.
20231 (emit_note_before): Use nobb variant of add_insn_before if the note
20232 should not be contained in a basic block.
20233 (emit_note_copy): Use make_note_raw.
20234 (emit_note): Likewise.
20235 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
20236 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
20237 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
20238 the moved barrier the tail of the basic block it follows.
20239 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
20241 2013-04-15 Jakub Jelinek <jakub@redhat.com>
20243 PR tree-optimization/56962
20244 * gimple-ssa-strength-reduction.c (record_increment): Only set
20245 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
20246 either rhs1 or rhs2 is equal to c->base_expr.
20248 2013-04-15 Richard Biener <rguenther@suse.de>
20250 PR tree-optimization/56933
20251 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
20253 (GROUP_READ_WRITE_DEPENDENCE): Remove.
20254 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
20255 * tree-vect-data-refs.c (vect_analyze_group_access): Move
20256 dependence check ...
20257 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
20259 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
20260 GROUP_READ_WRITE_DEPENDENCE.
20262 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20264 * emit-rtl.c (reset_all_used_flags): New function.
20265 (verify_rtl_sharing): Call reset_all_used_flags before and after
20266 performing the checks.
20268 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20270 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
20271 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
20272 * config/arm/constraints.md (De): New constraint.
20273 * config/arm/neon.md (anddi3_neon): Delete.
20274 (neon_vand<mode>): Expand to standard anddi3 pattern.
20275 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
20276 Move earlier in the file.
20277 (neon_inv_logic_op2): Likewise.
20278 (arm_anddi_operand_neon): New predicate.
20280 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20282 * configure.ac (gcc_cv_ld_as_needed): Set
20283 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
20284 Use -z ignore, -z record on *-*-solaris2*.
20285 (HAVE_LD_AS_NEEDED): Update comment.
20286 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
20287 * configure: Regenerate.
20288 * config.in: Regenerate.
20289 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
20290 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
20291 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
20292 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
20293 equivalents. Fix markup.
20294 * doc/tm.texi: Regenerate.
20296 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
20298 * config/i386/i386.opt: New option mstack-protector-guard=.
20299 * config/i386/i386-opts.h: Add enum stack_protector_guard.
20300 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
20301 TARGET_SSP_TLS_GUARD.
20302 * config/i386/i386.c (ix86_option_override_internal): Set
20303 ix86_stack_protector_guard.
20304 * config/i386/i386.md (stack_protect_set): Enable for
20305 TARGET_SSP_TLS_GUARD only.
20306 (stack_protect_set_<mode>): Ditto.
20307 (stack_protect_test): Ditto.
20308 (stack_protect_test_<mode>): Ditto.
20309 * doc/invoke.texi (i386 Option): Document.
20311 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
20314 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
20315 (S_MODES): Set H_MODE bit.
20316 (SF_MODES): Set only S_MODE and SF_MODE bits.
20317 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
20318 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
20319 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
20320 <MODE_FLOAT>: Likewise.
20322 2013-04-15 Joey Ye <joey.ye@arm.com>
20324 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
20326 2013-04-15 Joey Ye <joey.ye@arm.com>
20328 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
20330 (thumb_far_jump_used_p): Count instruction size and set
20333 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
20335 * reorg.c (fill_simple_delay_slots): Reindent block of code.
20336 * resource.c (mark_target_live_regs): Reformat conditional block.
20338 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20340 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
20341 notes, they are emitted only just before final.
20342 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
20344 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20346 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
20347 * cfgrtl.c (delete_insn): Call it here instead.
20348 * lra-spills.c (lra_final_code_change): Use delete_insn.
20349 * haifa-sched.c (sched_remove_insn): Likewise.
20350 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
20351 returning to the nop pool.
20352 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
20353 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
20355 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20357 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
20358 * doc/tm.texi: Regenerated.
20360 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
20362 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
20365 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20367 * df-core.c (df_find_def): Compare register numbers.
20368 (df_find_use): Likewise.
20370 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
20373 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
20374 lra_in_progress for return.
20376 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20378 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
20379 define_insn into define_insn_and_split and emit movsicc patterns.
20381 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20383 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
20385 2013-04-12 Richard Biener <rguenther@suse.de>
20387 * tree-pass.h (TODO_do_not_ggc_collect): New.
20388 * passes.c (execute_one_ipa_transform_pass): Honor
20389 TODO_do_not_ggc_collect.
20390 (execute_one_pass): Likewise.
20393 2013-04-10 Richard Biener <rguenther@suse.de>
20395 * passes.c (init_optimization_passes): Remove reload pass.
20396 * ira.c (do_reload): Merge into ...
20398 (rest_of_handle_reload): Remove.
20399 (pass_reload): Likewise.
20400 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20401 of reload for vzeroupper pass placement.
20403 2013-04-12 Jakub Jelinek <jakub@redhat.com>
20405 PR tree-optimization/56918
20406 PR tree-optimization/56920
20407 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
20408 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
20409 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
20410 use wide_mul_with_sign method.
20412 2013-04-12 Richard Biener <rguenther@suse.de>
20414 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
20415 not be considered a gimple constant.
20417 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
20419 * fold-const.c (const_binop): Handle vector shifts by a scalar.
20420 (fold_binary_loc): Call const_binop also for mixed vector-scalar
20423 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
20424 Jakub Jelinek <jakub@redhat.com>
20426 * opts.c: Include diagnostic-color.h.
20427 (common_handle_option): Handle OPT_fdiagnostics_color_.
20428 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
20429 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
20430 (diagnostic-color.o): New.
20431 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
20432 (diagnostic_color_rule): New enum.
20433 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
20434 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
20435 the location string.
20436 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
20437 either NULL, or color kind.
20438 * diagnostic-color.c: New file.
20439 * diagnostic-color.h: New file.
20440 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
20442 * doc/invoke.texi (-fdiagnostics-color): Document.
20443 * pretty-print.h (pp_show_color): Define.
20444 (struct pretty_print_info): Add show_color field.
20445 * diagnostic.c: Include diagnostic-color.h.
20446 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
20447 macros. Colorize error:, warning: etc. strings and also the location
20449 (diagnostic_show_locus): Colorize the caret line.
20450 * pretty-print.c: Include diagnostic-color.h.
20451 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
20452 inside of %< %> quotes or quoted through q format modifier.
20454 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20456 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
20458 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20460 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
20461 code in CC_NZ mode.
20462 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
20465 2013-04-11 Marek Polacek <polacek@redhat.com>
20467 PR tree-optimization/48184
20468 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
20470 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
20472 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
20473 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
20474 (skip_simple_arithmetic): Tidy up.
20475 * tree.h (skip_simple_constant_arithmetic): Declare.
20477 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20479 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
20481 2013-04-11 Richard Biener <rguenther@suse.de>
20483 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20484 generate vector constants.
20486 2013-04-11 Richard Biener <rguenther@suse.de>
20488 PR tree-optimization/56878
20489 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
20490 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
20492 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
20493 Prefer to align the DR with the most invariant base address.
20495 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20497 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
20500 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
20502 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
20503 floating-point vector comparisons against 0.
20505 2013-04-11 Jakub Jelinek <jakub@redhat.com>
20507 PR tree-optimization/56899
20508 * fold-const.c (extract_muldiv_1): Apply distributive law
20509 only if TYPE_OVERFLOW_WRAPS (ctype).
20511 2013-04-11 Bin Cheng <bin.cheng@arm.com>
20514 * ira-costs.c (scan_one_insn): Check whether the source rtx of
20515 loading has side effect.
20517 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
20519 * config/sparc/sparc.c: Include tree-pass.h.
20520 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
20521 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
20522 head of file. Change return type. Split off gate function.
20523 (sparc_gate_work_around_errata): New function.
20524 (pass_work_around_errata): New pass definition.
20525 (insert_pass_work_around_errata) New pass insert definition to
20526 insert pass_work_around_errata just after delayed-branch scheduling.
20527 (sparc_option_override): Insert the pass.
20528 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
20530 2013-04-10 David S. Miller <davem@davemloft.net>
20532 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
20533 or -mcpu=hypersparc.
20535 * target.def (cstore_mode): New hook.
20536 * target.h: Include insn-codes.h
20537 * targhooks.c: Likewise.
20538 (default_cstore_mode): New function.
20539 * targhooks.h: Declare it.
20540 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
20541 * doc/tm.texi: Rebuild.
20542 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
20543 target hook, rather than inspecting the insn_data.
20544 * config/sparc/sparc.c (sparc_cstore_mode): New function.
20545 (TARGET_CSTORE_MODE): Redefine.
20546 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
20548 * config/sparc/predicates.md (cstore_result_operand): New special
20550 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
20551 Use it for operand 0.
20552 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
20553 (*snesi_special): Likewise.
20554 (*snesi_zero): Likewise.
20555 (*seqsi_zero): Likewise.
20556 (*sltu_insn): Likewise.
20557 (*sgeu_insn): Likewise.
20558 (*seqdi_special): Make operand 0 and comparison operation be of
20560 (*snedi_special): Likewise.
20561 (*snedi_special_vis3): Likewise.
20562 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
20563 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
20564 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
20565 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
20566 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
20567 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
20568 (*sltu_extend_sp64): Likewise.
20569 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
20570 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
20571 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
20572 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
20573 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
20575 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
20577 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
20578 (aarch64_start_file): Use the new function.
20580 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20581 Jason Merrill <jason@redhat.com>
20583 * common.opt: Add -gdwarf.
20584 * opts.c (common_handle_option): Handle it.
20585 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
20587 2013-04-10 Richard Biener <rguenther@suse.de>
20589 * passes.c (execute_todo): Do not call ggc_collect conditional here.
20590 (execute_one_ipa_transform_pass): But unconditionally here.
20591 (execute_one_pass): And here.
20592 (init_optimization_passes): Remove reload pass.
20593 * tree-pass.h (TODO_ggc_collect): Remove.
20594 (pass_reload): Likewise.
20595 * ira.c (do_reload): Merge into ...
20597 (rest_of_handle_reload): Remove.
20598 (pass_reload): Likewise.
20599 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20600 of reload for vzeroupper pass placement.
20601 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
20602 and todo_flags_finish of all passes.
20604 2013-04-10 Richard Biener <rguenther@suse.de>
20606 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
20607 first_const_oprnd field, rename first_def_type to first_op_type.
20608 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
20609 (vect_get_and_check_slp_defs): Always use the type of the
20610 operand. Allow mixed vect_external_def, vect_constant_def types.
20611 (vect_get_constant_vectors): Handle mixed vect_external_def,
20612 vect_constant_def types.
20614 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
20616 PR tree-optimization/55524
20617 * tree-ssa-math-opts.c
20618 (convert_mult_to_fma): Don't use an fms construct
20619 when we don't have an fms operation, but fnma, and it looks
20620 likely that we'll be able to use the latter.
20622 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
20624 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
20626 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
20627 inline fail caused by overwritable functions.
20629 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
20631 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
20632 unnecessary bits in the constant power of two case.
20634 2013-04-10 Richard Biener <rguenther@suse.de>
20636 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
20637 broken code swapping operands.
20638 (vect_build_slp_tree): Do not compute load permutations here.
20639 (vect_analyze_slp_instance): Compute load permutations here,
20640 after building the SLP tree.
20642 2013-04-09 Christian Bruel <christian.bruel@st.com>
20644 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
20645 of next/prev_real_insn.
20647 2013-04-09 Jan Hubicka <jh@suse.cz>
20649 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
20650 Drop aliased parameter.
20651 (function_and_variable_visibility): Do not handle alias pairs.
20652 * cgraph.c (varpool_externally_visible_p): Update prototype.
20653 * varpool.c (varpool_add_new_variable): Update.
20655 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20657 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
20659 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
20661 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
20663 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
20665 2013-04-09 Marek Polacek <polacek@redhat.com>
20667 PR tree-optimization/48762
20668 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
20670 2013-04-09 Richard Biener <rguenther@suse.de>
20672 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
20674 (vect_build_slp_tree): Likewise.
20675 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
20676 calculating the cost of a SLP instance.
20677 (vect_analyze_slp_instance): Use it from here, after building
20680 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20682 PR middle-end/56883
20683 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
20684 expand_omp_for_static_chunk): Use simple_p = true in
20685 force_gimple_operand_gsi calls when assigning to addressable decls.
20687 2013-04-09 Jeff Law <law@redhat.com>
20689 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
20690 when the boolean was created by converting a wider object which
20691 had a boolean range.
20693 2013-04-09 Richard Biener <rguenther@suse.de>
20695 * tree-vectorizer.h (slp_void_p): Remove.
20696 (slp_tree): Typedef before _slp_tree declaration.
20697 (struct _slp_tree): Use a vector of slp_tree as children.
20698 (vect_get_place_in_interleaving_chain): Remove.
20699 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
20701 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
20703 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
20704 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
20705 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
20706 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
20707 Use slp_node instead of slp_void_p and adjust.
20709 2013-04-09 Richard Biener <rguenther@suse.de>
20711 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
20712 work that is not necessary.
20714 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20716 PR tree-optimization/56854
20717 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
20718 forward into clobber stmts if it would change MEM_REF lhs into
20721 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
20723 * tree.c (type_hash_lookup, type_hash_add): Make static.
20724 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
20726 2013-04-09 Richard Biener <rguenther@suse.de>
20728 * tree.h (unsave_expr_now): Remove.
20729 * tree-inline.c (mark_local_for_remap_r): Remove.
20730 (unsave_expr_1): Likewise.
20731 (unsave_r): Likewise.
20732 (unsave_expr_now): Likewise.
20733 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
20734 (propagate_tree_value): Likewise.
20736 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
20738 * doc/rtl.texi (sequence): Rewrite documentation to match the
20739 current use of SEQUENCE rtl objects.
20740 * rtl.def (SEQUENCE): Likewise.
20742 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
20743 Update documentation.
20744 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
20745 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
20747 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
20749 2013-04-08 Teresa Johnson <tejohnson@google.com>
20751 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
20752 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
20754 (estimate_edge_size_and_time): Add comment to suggest using rounding
20756 (estimate_node_size_and_time): Ditto.
20757 (remap_edge_change_prob): Use helper rounding divide methods.
20758 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
20759 (gimple_mod_pow2_value_transform): Ditto.
20760 (gimple_mod_subtract_transform): Ditto.
20761 (gimple_ic_transform): Ditto.
20762 (gimple_stringops_transform): Ditto.
20763 * stmt.c (conditional_probability): Ditto.
20764 (emit_case_dispatch_table): Ditto.
20765 * lto-cgraph.c (merge_profile_summaries): Ditto.
20766 * tree-optimize.c (execute_fixup_cfg): Ditto.
20767 * cfgcleanup.c (try_forward_edges): Ditto.
20768 * cfgloopmanip.c (scale_loop_profile): Ditto.
20770 (duplicate_loop_to_header_edge): Ditto.
20771 (lv_adjust_loop_entry_edge): Ditto.
20772 * tree-vect-loop.c (vect_transform_loop): Ditto.
20773 * profile.c (compute_branch_probabilities): Ditto.
20774 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
20775 * lto-streamer-in.c (input_cfg): Ditto.
20776 * gimple-streamer-in.c (input_bb): Ditto.
20777 * ipa-cp.c (update_profiling_info): Ditto.
20778 (update_specialized_profile): Ditto.
20779 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
20780 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
20782 * sched-rgn.c (compute_dom_prob_ps): Ditto.
20783 (compute_trg_info): Ditto.
20784 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
20785 (purge_dead_edges): Ditto.
20786 * loop-unswitch.c (unswitch_loop): Ditto.
20787 * cgraphclones.c (cgraph_clone_edge): Ditto.
20788 (cgraph_clone_node): Ditto.
20789 * tree-inline.c (copy_bb): Ditto.
20790 (copy_edges_for_bb): Ditto.
20791 (initialize_cfun): Ditto.
20792 (copy_cfg_body): Ditto.
20793 (expand_call_inline): Ditto.
20795 2013-04-08 Kai Tietz <ktietz@redhat.com>
20797 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
20798 TARGET_CYGWIN64 by TARGET_64BIT.
20800 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20802 * config/epiphany/epiphany.md (GPR_1): New constant.
20803 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
20804 * config/epiphany/epiphany.c (gen_compare_reg):
20805 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
20806 is already in place.
20807 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
20808 Don't require being called during rtl expansion; If y operlaps r0,
20810 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
20811 (epiphany_expand_epilogue): Likewise.
20813 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
20814 Don't use CC_FPmode for ORDERED / UNORDERED.
20815 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
20817 * config/epiphany/constraints.md (CnL): New constraint.
20818 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
20819 * config/epiphany/predicates.md (add_operand): Allow 1024.
20821 * config/epiphany/epiphany.md (logical_op): New code iterator.
20822 (op_mnc): New code attribute.
20823 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
20824 (mov_f+1, mov_f+2): New peephole2 patterns.
20826 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
20827 (cstoresi4): Also allow re-use of zero result when doing a NE
20828 comparison to a non-zero operand.
20829 Use (clobber (scratch)) for first insn if the gpr output is not needed.
20831 * config/epiphany/epiphany.md (<insn_opname>v2si3):
20832 Use gen_addsi3_i / gen_subsi3_i.
20834 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20838 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
20839 contain anything but clobbers, at most one __builtin_stack_restore,
20840 optionally debug stmts and final resx, and if it has at least one
20841 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
20842 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
20843 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
20844 which isn't defaut definition, remove them.
20845 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
20846 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
20847 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
20848 with MEM_REF LHS with SSA_NAME address.
20850 2013-04-08 Jeff Law <law@redhat.com>
20852 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
20854 2013-04-08 Richard Biener <rguenther@suse.de>
20856 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
20858 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
20859 determined vector type.
20860 (vect_analyze_data_refs): Likewise.
20861 (vect_get_new_vect_var): Adjust.
20862 (vect_create_destination_var): Preserve SSA name versions.
20863 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
20864 not dump anything here.
20866 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20868 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
20869 Add member lr_slot_known.
20870 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
20872 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
20873 Remove code that sets lr_slot_offset according to what a previous
20874 version of epiphany_emit_save_restore used to do.
20875 (epiphany_emit_save_restore): When doing an lr save or restore,
20876 set/verify lr_slot_known and lr_slot_offset.
20878 2013-04-08 Xinyu Qi <xyqi@marvell.com>
20881 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
20884 2013-04-08 Richard Biener <rguenther@suse.de>
20886 * alias.c (find_base_term): Fix thinko in previous change.
20888 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20890 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
20891 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
20892 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
20893 if possible to compute val.
20894 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
20895 For QImode integers don't require anything about precision. Use
20896 const_with_all_bytes_same to find out if the constant doesn't have
20897 repeated bytes in it.
20899 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20901 * config/s390/s390.c (s390_expand_insv): Only accept insertions
20904 2013-04-08 Marek Polacek <polacek@redhat.com>
20906 PR rtl-optimization/48182
20907 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
20910 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20913 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
20914 nuses, make sure we have a label.
20916 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20919 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
20920 (rs6000_emit_swdiv_low_precision): Remove.
20921 (rs6000_emit_swdiv): Rewrite to handle between one and four
20922 iterations of Newton-Raphson generally; modify required number of
20923 iterations for some cases.
20924 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
20926 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
20928 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
20929 set-but-unused variable.
20931 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
20932 basic blocks of released function bodies garbage-collectable.
20934 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
20935 (struct rtl_opt_pass): Add TODO_df_finish.
20937 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
20939 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20941 * config/arm/constraints.md (q): New constraint.
20942 * config/arm/ldrdstrd.md: New file.
20943 * config/arm/arm.md (ldrdstrd.md) New include.
20944 (arm_movdi): Use "q" instead of "r" constraint
20945 for double-word memory access.
20946 (movdf_soft_insn): Likewise.
20947 * config/arm/vfp.md (movdi_vfp): Likewise.
20948 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
20949 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
20950 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
20951 (mem_ok_for_ldrd_strd): Likewise.
20952 (output_move_double): Update assertion.
20954 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20956 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
20958 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20960 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
20961 define_insn_and_split.
20962 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
20964 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20966 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
20967 define_insn_and_split.
20968 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
20969 (shiftsi3_compare): New pattern.
20970 (rrx): New pattern.
20971 * config/arm/unspecs.md (UNSPEC_RRX): New.
20973 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20975 * config/arm/arm.md (negdi_extendsidi): New pattern.
20976 (negdi_zero_extendsidi): Likewise.
20978 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20980 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
20981 define_insn_and_split.
20982 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
20983 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
20985 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20987 * config/arm/arm.md (arm_subdi3): Convert define_insn into
20988 define_insn_and_split.
20989 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
20990 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
20992 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20994 * config/arm/arm.md (subsi3_carryin): New pattern.
20995 (subsi3_carryin_const): Likewise.
20996 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
20997 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
20999 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21001 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
21003 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21005 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
21006 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
21008 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21010 * config/arm/arm.c (arm_expand_builtin): Change fcode
21011 type to unsigned int.
21013 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21015 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
21017 2013-04-04 Ian Lance Taylor <iant@google.com>
21019 * doc/standards.texi (Standards): The Go frontend supports the Go 1
21022 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
21024 PR middle-end/56729
21025 * df-scan.c (df_insn_delete): Disable failing assert.
21027 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21029 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
21030 New function prototype.
21031 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
21032 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
21033 (arm_builtin_vectorized_function): New function.
21035 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21037 * config/arm/arm_neon_builtins.def: New file.
21038 * config/arm/arm.c (neon_builtin_data): Move contents to
21039 arm_neon_builtins.def.
21040 (enum arm_builtins): Include neon builtin definitions.
21041 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
21042 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
21044 2013-04-04 Marek Polacek <polacek@redhat.com>
21046 PR tree-optimization/48186
21047 * predict.c (maybe_hot_frequency_p): Return false if
21048 HOT_BB_FREQUENCY_FRACTION is 0.
21049 (cgraph_maybe_hot_edge_p): Likewise.
21051 2013-04-04 Richard Biener <rguenther@suse.de>
21053 PR tree-optimization/56826
21054 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
21057 2013-04-04 Richard Biener <rguenther@suse.de>
21059 PR tree-optimization/56213
21060 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
21061 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
21063 2013-04-04 Richard Biener <rguenther@suse.de>
21065 PR tree-optimization/56837
21066 * tree-loop-distribution.c (classify_partition): For non-zero
21067 values require that the value has the same precision as its
21068 mode to be useful as memset value.
21070 2013-04-03 Nick Clifton <nickc@redhat.com>
21072 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
21073 (fmssf4): Use fmsf.s on E3V5 architectures.
21074 (fnmasf4): Use fnmaf.s on E3V5 architectures.
21075 (fnmssf4): Use fnmsf.s on E3V5 architectures.
21077 2013-04-03 Jeff Law <law@redhat.com>
21079 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
21080 (lra-eliminations.o): Likewise.
21082 2013-04-03 Teresa Johnson <tejohnson@google.com>
21084 * gcov-io.c (compute_working_sets): Moved most of body of old
21085 compute_working_sets here from profile.c.
21086 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
21087 (gcov_working_set_t): Moved typedef here from basic-block.h
21088 (compute_working_set): Declare.
21089 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
21090 (get_working_sets): Renamed from compute_working_set,
21091 replace most of body with call to new compute_working_sets.
21092 (get_exec_counts): Replace call to compute_working_sets
21093 to get_working_sets.
21094 * profile.h (get_working_sets): Renamed from compute_working_set.
21095 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
21096 to get_working_sets.
21097 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
21098 * gcov-dump.c (dump_working_sets): New function.
21100 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
21102 * hwint.c (sext_hwi, zext_hwi): New functions.
21103 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
21104 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
21105 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
21106 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
21107 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
21108 (sext_hwi, zext_hwi): New functions.
21110 2013-04-03 Jeff Law <law@redhat.com>
21112 PR tree-optimization/56799
21113 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
21114 back test for widening conversion erroneously dropped in prior change.
21116 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21119 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
21120 instead of next_real_insn.
21122 2013-04-03 Marek Polacek <polacek@redhat.com>
21125 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
21127 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21130 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
21132 (thumb1_output_casesi): Likewise.
21133 (thumb2_output_casesi): Likewise.
21135 2013-04-03 Richard Biener <rguenther@suse.de>
21137 PR tree-optimization/56817
21138 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
21140 (tree_unroll_loops_completely_1): ... new function to manually
21141 walk the loop tree, properly defering outer loops of unrolled
21142 loops to later iterations.
21144 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21146 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
21147 (vectorizable_load): Likewise.
21148 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
21149 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
21151 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21153 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
21156 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
21158 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
21160 2013-04-03 Bin Cheng <bin.cheng@arm.com>
21162 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
21164 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21166 PR tree-optimization/56790
21167 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
21170 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21172 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
21174 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
21175 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
21178 2013-04-03 Jakub Jelinek <jakub@redhat.com>
21181 * tree.h (force_folding_builtin_constant_p): New decl.
21182 * builtins.c (force_folding_builtin_constant_p): New variable.
21183 (fold_builtin_constant_p): Fold immediately also if
21184 force_folding_builtin_constant_p.
21186 2013-04-03 Richard Biener <rguenther@suse.de>
21188 PR tree-optimization/56812
21189 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
21190 DRs of the same interleaving chain are independent.
21192 2013-04-02 Jason Merrill <jason@redhat.com>
21194 * gdbinit.in (pbb): Use debug fn.
21196 2013-04-02 Lawrence Crowl <crowl@google.com>
21198 * sese.h (struct ivtype_map_elt_s): Remove unused.
21199 (extern debug_ivtype_map): Remove unused.
21200 (extern eq_ivtype_map_elts): Remove unused.
21201 * sese.c (debug_ivtype_map): Removed unused.
21202 (debug_ivtype_map_1): Removed unused.
21203 (debug_ivtype_elt): Remove unused.
21204 (eq_ivtype_map_elts): Remove unused.
21207 2013-04-02 Kai Tietz <ktietz@redhat.com>
21210 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
21211 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
21212 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
21214 (legitimize_pe_coff_symbol): Likewise.
21215 (is_imported_p): New helper-function.
21216 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
21217 for Windows x64 targets.
21218 (ix86_expand_prologue): Optimize for pe-coff targets.
21219 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
21220 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
21221 medium/large code-model.
21222 (legitimize_pic_address): Likewise.
21223 (legitimize_tls_address): Likewise.
21224 (ix86_expand_call): Likewise.
21225 (x86_output_mi_thunk): Likewise.
21226 (get_dllimport_decl): Add new beimport argument.
21227 (construct_plt_address): Don't assert for x64 pe-coff targets.
21228 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
21230 (SYMBOL_FLAG_STUBVAR): New macro.
21231 (SYMBOL_REF_STUBVAR_P): Likewise.
21232 * config/i386/winnt.c (stub_list): New structure.
21233 (stub_head): New local variable.
21234 (i386_pe_record_stub): New function.
21235 (i386_pe_file_end): Emit refptr-stubs.
21237 2013-04-02 Jakub Jelinek <jakub@redhat.com>
21239 PR rtl-optimization/56745
21240 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
21241 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
21244 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
21245 and both of them are MEM_REFs, just compare first argument for
21246 equality and attempt to deal even with differing offsets.
21249 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
21250 of gimple_clobber_p to be MEM_REF.
21251 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
21252 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
21253 after gimplification.
21254 * asan.c (get_mem_ref_of_assignment): Don't instrument
21255 gimple_clobber_p stmts.
21256 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
21257 gimple_clobber_p stmt if they have MEM_REF lhs and
21258 are dead because of another gimple_clobber_p stmt.
21259 * tree-ssa-live.c (clear_unused_block_pointer): Treat
21260 gimple_clobber_p stmts like debug stmts.
21261 (remove_unused_locals): Remove clobbers with MEM_REF lhs
21262 that refer to unused VAR_DECLs or uninitialized values.
21263 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
21264 gimple_clobber_p stmts if they refer to removed parameters.
21265 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
21268 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
21270 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
21271 using SWI48 mode attribute.
21273 2013-04-02 Wei Mi <wmi@google.com>
21275 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
21276 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
21277 *<rotate_insn><mode>3_mask in i386.md.
21279 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
21281 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
21283 2013-04-02 Richard Biener <rguenther@suse.de>
21285 PR tree-optimization/56778
21286 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
21287 Runtime alias tests are not supported for gather loads.
21288 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
21289 stmts referenced from SSA operands before updating SSA form.
21291 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
21292 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21294 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
21295 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
21296 * config/arm/cortex-a53.md: New file.
21297 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
21298 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
21299 * config/arm/arm.c (arm_issue_rate): Likewise.
21300 * config/arm/arm-tune.md: Regenerate
21301 * config/arm/arm-tables.opt: Regenerate.
21302 * config/arm/arm-cores.def: Add cortex-a53.
21304 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
21306 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
21309 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21311 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
21312 scalar load/store operations using B/H registers.
21313 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
21315 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21317 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
21319 * config/aarch64/aarch64.c
21320 (aarch64_simd_scalar_immediate_valid_for_move): New.
21321 * config/aarch64/aarch64-protos.h
21322 (aarch64_simd_scalar_immediate_valid_for_move): New.
21323 * config/aarch64/constraints.md (Dh, Dq): New.
21324 * config/aarch64/iterators.md (hq): New.
21326 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
21328 * reorg.c (get_branch_condition): Deal with conditional returns.
21329 (fill_simple_delay_slots): Remove dead code dealing with jumps.
21331 2013-04-01 Wei Mi <wmi@google.com>
21333 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
21334 Truncate operand 2 using %b asm operand modifier.
21335 (*<shift_insn><mode>3_mask): Ditto.
21336 (*<rotate_insn><mode>3_mask): Ditto.
21338 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
21340 PR middle-end/56798
21341 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
21343 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
21345 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
21347 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
21349 2013-03-30 Lawrence Crowl <crowl@google.com>
21351 * dse.c (clear_alias_sets): Remove never set.
21352 (disqualified_clear_alias_sets): Remove never set.
21353 (clear_alias_mode_pool): Remove never set.
21354 (dse_step0): Remove condition that is never true.
21355 (canon_address): Remove condition that is never true.
21356 (dse_step7): Remove condition that is never true.
21357 (rest_of_handle_dse): Remove condition that is never true.
21358 (rest_of_handle_dse::did_global): Remove never read from above.
21359 (dse_step2_spill): Remove never called from above.
21360 (dse_step5_spill): Remove never called from above.
21362 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
21364 * doc/md.texi (Standard Names) <casesi>: Update documentation for
21365 JUMP_TABLE_DATA changes.
21366 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
21367 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
21368 (Insns) <jump_table_data>: New entry.
21369 * doc/tm.texi: Regenerate.
21371 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
21373 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
21374 for table jump at the end of a basic block using tablejump_p.
21375 * targhooks.c (default_invalid_within_doloop): Likewise.
21376 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
21377 target hook implementation that is identical to the default hook.
21378 (rs6000_invalid_within_doloop): Remove.
21380 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
21381 unused variable from tablejump_p call.
21383 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
21384 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
21385 (INSN_DELETED_P): Likewise.
21386 (emit_jump_table_data): New prototype.
21387 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
21388 after 4th as unused.
21389 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
21390 * sched-vis.c (print_insn): Likewise.
21391 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
21392 insn for compatibility with back ends that use next_active_insn to
21393 identify jump table data.
21394 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
21395 (remove_insn): Likewise.
21396 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
21398 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
21399 (emit_jump_table_data): New function.
21401 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
21402 basic block, a JUMP_TABLE_DATA never is.
21403 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
21404 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
21405 off from code handling real insns.
21406 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
21407 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
21408 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
21409 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
21410 is not a NONDEBUG_INSN_P.
21411 * ira-costs.c (scan_one_insn): Likewise.
21412 * jump.c (mark_all_labels): Likewise.
21413 (mark_jump_label_1): Likewise.
21414 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21415 * lra.c (get_insn_freq): Expect all insns reaching here to be in
21417 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
21418 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
21419 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
21420 JUMP_TABLE_DATA_P insns.
21421 (calculate_elim_costs_all_insns): Likewise.
21422 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
21423 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
21424 (delete_output_reload): Code style fixups.
21425 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
21426 insn flags on this non-insn.
21427 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
21428 as scheduling barriers, for pre-change compatibility.
21429 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
21430 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
21432 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
21433 redundant JUMP_TABLE_DATA_P test.
21434 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
21435 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21436 (frv_for_each_packet): Likewise.
21437 * config/i386/i386.c (min_insn_size): Likewise.
21438 (ix86_avoid_jump_mispredicts): Likewise.
21439 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21440 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
21441 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21442 (mips16_insn_length): Robustify.
21443 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
21444 (mips16_split_long_branches): Likewise.
21445 * config/pa/pa.c (pa_combine_instructions): Likewise.
21446 * config/rs6000/rs6000.c (get_next_active_insn): Treat
21447 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
21448 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
21449 as contributing to pool range lengths.
21450 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
21451 Remove redundant JUMP_TABLE_DATA_P test.
21452 (sh_loop_align): Likewise.
21453 (split_branches): Likewise.
21454 (sh_insn_length_adjustment): Likewise.
21455 * config/spu/spu.c (get_branch_target): Likewise.
21457 2013-03-29 Jan Hubicka <jh@suse.cz>
21459 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
21460 gcov streaming; stream hot bb threshold to ltrans.
21461 * predict.c (get_hot_bb_threshold): Break out from ....
21462 (maybe_hot_count_p): ... here.
21463 (set_hot_bb_threshold): New function.
21464 * lto-section-in.c (lto_section_name): Add profile.
21465 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
21466 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
21467 and data-streamer.h
21468 (histogram_entry): New structure.
21469 (histogram, histogram_pool): New global vars.
21470 (histogram_hash): New structure.
21471 (histogram_hash::hash): New method.
21472 (histogram_hash::equal): Likewise.
21473 (account_time_size): New function.
21474 (cmp_counts): New function.
21475 (dump_histogram): New function.
21476 (ipa_profile_generate_summary): New function.
21477 (ipa_profile_write_summary): New function.
21478 (ipa_profile_read_summary): New function.
21479 (ipa_profile): Decide on threshold.
21480 (pass_ipa_profile): Add ipa_profile_write_summary and
21481 ipa_profile_read_summary.
21482 * Makefile.in (ipa.o): Update dependencies.
21483 * lto-streamer.h (LTO_section_ipa_profile): New section.
21485 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
21487 * tree.h (VAR_P): New.
21489 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
21492 * doc/invoke.texi ([-fwhole-program]): Fix typo.
21494 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
21496 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
21497 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21498 (control_flow_insn_p): Likewise.
21499 * cfgrtl.c (duplicate_insn_chain): Likewise.
21500 * final.c (get_attr_length_1): Likewise.
21501 (shorten_branches): Likewise.
21502 (final_scan_insn): Likewise.
21503 * function.c (instantiate_virtual_regs): Likewise.
21504 * gcse.c (insert_insn_end_basic_block): Likewise.
21505 * ira-costs.c (scan_one_insn): Likewise.
21506 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21507 * lra.c (check_rtl): Likewise.
21508 * reload1.c (elimination_costs_in_insn): Likewise.
21509 * reorg.c (follow_jumps): Likewise.
21511 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
21512 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21513 (thumb_far_jump_used_p): Likewise.
21514 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
21515 (workaround_speculation): Likewise.
21516 (add_sched_insns_for_speculation): Likewise.
21517 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
21518 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21519 (frv_for_each_packet): Likewise.
21520 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
21521 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
21522 (final_emit_insn_group_barriers): Likewise.
21523 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21524 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21525 (mips16_insn_length): Likewise.
21526 * config/pa/pa.c (pa_reorg): Likewise.
21527 (pa_combine_instructions): Likewise.
21528 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
21529 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
21530 (sh_reorg): Likewise.
21531 (split_branches): Likewise.
21532 * config/spu/spu.c (get_branch_target): Likewise.
21534 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
21537 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
21539 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
21540 Fix declaration name.
21542 2013-03-28 Lawrence Crowl <crowl@google.com>
21544 * graphds.h (struct graph.indicies): Remove unused.
21545 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
21546 (SCOP_ORIGINAL_PDDRS): Remove unused.
21547 * sese.h (extern insert_loop_close_phis): Removed unused.
21548 (extern insert_guard_phis): Removed unused.
21549 (extern ivtype_map_elt_info): Removed unused.
21550 (new_ivtype_map_elt): Removed unused.
21551 * sese.c (ivtype_map_elt_info): Removed unused.
21553 2013-03-28 Lawrence Crowl <crowl@google.com>
21555 * Makefile.in: Add several missing include dependences.
21557 (test-dump.o): New. This object is not added to any executable,
21558 but is present for ad-hoc testing.
21560 (debug (const bitmap_head_def &)): New.
21561 (debug (const bitmap_head_def *)): New.
21563 (extern debug (const bitmap_head_def &)): New.
21564 (extern debug (const bitmap_head_def *)): New.
21566 (debug (edge_def &)): New.
21567 (debug (edge_def *)): New.
21569 (debug (basic_block_def &)): New.
21570 (debug (basic_block_def *)): New.
21572 (dump_node (const_tree, int, FILE *)): Correct source file.
21574 (debug (die_struct &)): New.
21575 (debug (die_struct *)): New.
21577 (extern debug (die_struct &)): New.
21578 (extern debug (die_struct *)): New.
21579 * gimple-pretty-print.c
21580 (debug (gimple_statement_d &)): New.
21581 (debug (gimple_statement_d *)): New.
21582 * gimple-pretty-print.h
21583 (extern debug (gimple_statement_d &)): New.
21584 (extern debug (gimple_statement_d *)): New.
21586 (debug (ira_allocno_copy &)): New.
21587 (debug (ira_allocno_copy *)): New.
21588 (debug (ira_allocno &)): New.
21589 (debug (ira_allocno *)): New.
21591 (extern debug (ira_allocno_copy &)): New.
21592 (extern debug (ira_allocno_copy *)): New.
21593 (extern debug (ira_allocno &)): New.
21594 (extern debug (ira_allocno *)): New.
21596 (debug (live_range &)): New.
21597 (debug (live_range *)): New.
21599 (debug (lra_live_range &)): New.
21600 (debug (lra_live_range *)): New.
21602 (debug (lra_live_range &)): New.
21603 (debug (lra_live_range *)): New.
21605 (debug (omega_pb_d &)): New.
21606 (debug (omega_pb_d *)): New.
21608 (extern debug (omega_pb_d &)): New.
21609 (extern debug (omega_pb_d *)): New.
21611 (debug (const rtx_def &)): New.
21612 (debug (const rtx_def *)): New.
21614 (debug_tree (tree): Move within file.
21615 (debug_raw (const tree_node &)): New.
21616 (debug_raw (const tree_node *)): New.
21617 (dump_tree_via_hooks (const tree_node *, int)): New.
21618 (debug (const tree_node &)): New.
21619 (debug (const tree_node *)): New.
21620 (debug_verbose (const tree_node &)): New.
21621 (debug_verbose (const tree_node *)): New.
21622 (debug_head (const tree_node &)): New.
21623 (debug_head (const tree_node *)): New.
21624 (debug_body (const tree_node &)): New.
21625 (debug_body (const tree_node *)): New.
21626 (debug_vec_tree (tree): Move and reimplement in terms of dump.
21627 (debug (vec<tree, va_gc> &)): New.
21628 (debug (vec<tree, va_gc> *)): New.
21630 (extern debug (const rtx_def &)): New.
21631 (extern debug (const rtx_def *)): New.
21633 (debug_raw (simple_bitmap_def &)): New.
21634 (debug_raw (simple_bitmap_def *)): New.
21635 (debug (simple_bitmap_def &)): New.
21636 (debug (simple_bitmap_def *)): New.
21638 (extern debug (simple_bitmap_def &)): New.
21639 (extern debug (simple_bitmap_def *)): New.
21640 (extern debug_raw (simple_bitmap_def &)): New.
21641 (extern debug_raw (simple_bitmap_def *)): New.
21643 (debug (vinsn_def &)): New.
21644 (debug (vinsn_def *)): New.
21645 (debug_verbose (vinsn_def &)): New.
21646 (debug_verbose (vinsn_def *)): New.
21647 (debug (expr_def &)): New.
21648 (debug (expr_def *)): New.
21649 (debug_verbose (expr_def &)): New.
21650 (debug_verbose (expr_def *)): New.
21651 (debug (vec<rtx> &)): New.
21652 (debug (vec<rtx> *)): New.
21654 (extern debug (vinsn_def &)): New.
21655 (extern debug (vinsn_def *)): New.
21656 (extern debug_verbose (vinsn_def &)): New.
21657 (extern debug_verbose (vinsn_def *)): New.
21658 (extern debug (expr_def &)): New.
21659 (extern debug (expr_def *)): New.
21660 (extern debug_verbose (expr_def &)): New.
21661 (extern debug_verbose (expr_def *)): New.
21662 (extern debug (vec<rtx> &)): New.
21663 (extern debug (vec<rtx> *)): New.
21665 (_list_iter_cond_expr): Make inline instead of static.
21667 (debug (sreal &)): New.
21668 (debug (sreal *)): New.
21670 (extern debug (sreal &)): New.
21671 (extern debug (sreal *)): New.
21673 (extern debug_raw (const tree_node &)): New.
21674 (extern debug_raw (const tree_node *)): New.
21675 (extern debug (const tree_node &)): New.
21676 (extern debug (const tree_node *)): New.
21677 (extern debug_verbose (const tree_node &)): New.
21678 (extern debug_verbose (const tree_node *)): New.
21679 (extern debug_head (const tree_node &)): New.
21680 (extern debug_head (const tree_node *)): New.
21681 (extern debug_body (const tree_node &)): New.
21682 (extern debug_body (const tree_node *)): New.
21683 (extern debug (vec<tree, va_gc> &)): New.
21684 (extern debug (vec<tree, va_gc> *)): New.
21686 (debug (struct loop &)): New.
21687 (debug (struct loop *)): New.
21688 (debug_verbose (struct loop &)): New.
21689 (debug_verbose (struct loop *)): New.
21690 * tree-dump.c: Add header dependence.
21692 (extern debug (struct loop &)): New.
21693 (extern debug (struct loop *)): New.
21694 (extern debug_verbose (struct loop &)): New.
21695 (extern debug_verbose (struct loop *)): New.
21697 (debug (data_reference &)): New.
21698 (debug (data_reference *)): New.
21699 (debug (vec<data_reference_p> &)): New.
21700 (debug (vec<data_reference_p> *)): New.
21701 (debug (vec<ddr_p> &)): New.
21702 (debug (vec<ddr_p> *)): New.
21704 (extern debug (data_reference &)): New.
21705 (extern debug (data_reference *)): New.
21706 (extern debug (vec<data_reference_p> &)): New.
21707 (extern debug (vec<data_reference_p> *)): New.
21708 (extern debug (vec<ddr_p> &)): New.
21709 (extern debug (vec<ddr_p> *)): New.
21711 (debug (pt_solution &)): New.
21712 (debug (pt_solution *)): New.
21714 (extern debug (pt_solution &)): New.
21715 (extern debug (pt_solution *)): New.
21717 (debug (_var_map &)): New.
21718 (debug (_var_map *)): New.
21719 (debug (tree_live_info_d &)): New.
21720 (debug (tree_live_info_d *)): New.
21722 (extern debug (_var_map &)): New.
21723 (extern debug (_var_map *)): New.
21724 (extern debug (tree_live_info_d &)): New.
21725 (extern debug (tree_live_info_d *)): New.
21727 2013-03-28 Jan Hubicka <jh@suse.cz>
21729 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
21731 2013-03-28 Ian Bolton <ian.bolton@arm.com>
21733 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
21734 record only when desired or required.
21736 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
21738 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
21739 *vec_extractv2di_1_rex64. Use x64 isa attribute.
21741 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21743 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
21744 (*andsi3_compare0_uxtw): New pattern.
21745 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
21746 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
21748 2013-03-28 Jan Hubicka <jh@suse.cz>
21750 * data-streamer-in.c (streamer_read_gcov_count): New function.
21751 * gimple-streamer-out.c: Include value-prof.h.
21752 (output_gimple_stmt): Output histogram.
21753 (output_bb): Use streamer_write_gcov_count.
21754 * value-prof.c: Include data-streamer.h
21755 (dump_histogram_value): Add HIST_TYPE_MAX.
21756 (stream_out_histogram_value): New function.
21757 (stream_in_histogram_value): New function.
21758 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
21759 (stream_out_histogram_value, stream_in_histogram_value): Declare.
21760 * data-streamer-out.c (streamer_write_gcov_count): New function.
21761 (streamer_write_gcov_count_stream): New function.
21762 * lto-cgraph.c (lto_output_edge): Update counter streaming.
21763 (lto_output_node): Likewise.
21764 (input_node, input_edge): Likewise.
21765 * lto-streamer-out.c (output_cfg): Update streaming.
21766 * lto-streamer-in.c (input_cfg): Likewise.
21767 * data-streamer.h (streamer_write_gcov_count,
21768 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
21769 * gimple-streamer-in.c: Include value-prof.h
21770 (input_gimple_stmt): Input histograms.
21771 (input_bb): Update profile streaming.
21773 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21775 * genmodes.c (emit_max_int): New function.
21776 (emit_insn_modes_h): Added call to emit_max_function.
21777 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
21779 * machmode.def: Fixed comment.
21781 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21783 * combine.c (try_combine): Removed useless assert.
21784 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
21786 2013-03-28 Marek Polacek <polacek@redhat.com>
21787 Richard Biener <rguenther@suse.de>
21789 PR tree-optimization/56695
21790 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
21791 build signed result of a vector comparison.
21792 * tree-cfg.c (verify_gimple_comparison): Check that a result
21793 of a vector comparison has signed type.
21795 2013-03-28 Richard Biener <rguenther@suse.de>
21797 PR tree-optimization/37021
21798 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
21799 do not restrict gaps between groups.
21800 * tree-vect-stmts.c (vectorizable_load): Properly account for
21801 a gap between groups.
21803 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
21805 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
21806 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
21809 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
21811 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
21812 * doc/extend.texi (Named Address Spaces): Ditto.
21813 (Variable Attributes): Ditto.
21815 2013-03-27 Kai Tietz <ktietz@redhat.com>
21817 * config.build: Add support for cygwin x64 target.
21818 * config.gcc: Likewise.
21819 * config.host: Likewise.
21820 * configure.ac: Likewise
21821 * configure: Regenerated.
21823 2013-03-27 Kai Tietz <ktietz@redhat.com>
21825 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
21826 * config/i386/t-cygwin-w64: New file.
21827 * config/i386/cygwin-w64.h: New file.
21828 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
21829 and add support for x64-cygwin target.
21830 (CPP_SPEC): Likewise.
21831 (CXX_WRAP_SPEC_LIST): Undefine before define.
21832 (LIBGCJ_SONAME): Use 15 as version.
21834 2013-03-27 Richard Biener <rguenther@suse.de>
21836 PR tree-optimization/56716
21837 * tree-ssa-structalias.c (perform_var_substitution): Adjust
21838 dumping for ref nodes.
21840 2013-03-27 Martin Jambor <mjambor@suse.cz>
21842 PR tree-optimization/55334
21843 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
21844 restricted pointers to arrays.
21846 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
21848 * Makefile.in (.SUFFIXES): Add .cc.
21849 (.c.o): Apply same recipe for implicit rule .cc.o.
21851 2013-03-27 Richard Biener <rguenther@suse.de>
21853 PR tree-optimization/37021
21854 * tree-vect-data-refs.c (vect_check_strided_load): Allow
21855 REALPART/IMAGPART_EXPRs around the supported refs.
21856 * tree-ssa-structalias.c (find_func_aliases): Assume that
21857 floating-point values are not used to transfer pointers.
21859 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
21861 * target.def (TARGET_HAS_IFUNC_P): New target hook.
21862 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
21863 * doc/tm.texi: Regenerate.
21864 * targhooks.h (default_has_ifunc_p): New.
21865 * targhooks.c (default_has_ifunc_p): Ditto.
21866 * config/linux-protos.h: New file.
21867 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
21868 hook for linux which disables support of indirect functions in android.
21869 * config/linux-android.c: New file.
21870 * config/t-linux-android.c: Ditto.
21871 * config.gcc: Added new object file linux-android.o.
21872 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
21873 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
21874 * varasm.c (do_assemble_alias): Likewise.
21875 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
21876 doesn't support indirect functions.
21877 * configure: Regenerate.
21879 2013-03-27 Bin Cheng <bin.cheng@arm.com>
21882 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
21883 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
21886 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21888 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
21890 2013-03-27 Terry Guo <terry.guo@arm.com>
21892 * config/arm/arm-cores.def: Added core cortex-r7.
21893 * config/arm/arm-tune.md: Regenerated.
21894 * config/arm/arm-tables.opt: Regenerated.
21895 * doc/invoke.texi: Added entry for core cortex-r7.
21897 2013-03-27 Walter Lee <walt@tilera.com>
21899 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
21900 double-decrement of next_scratch_regno.
21902 2013-03-27 Walter Lee <walt@tilera.com>
21904 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
21906 (insn_v1mulus): Ditto.
21907 (insn_v2muls): Ditto.
21909 2013-03-27 Walter Lee <walt@tilera.com>
21911 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
21912 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
21914 2013-03-27 Walter Lee <walt@tilera.com>
21916 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
21917 (*sibcall_value): Ditto.
21919 2013-03-27 Walter Lee <walt@tilera.com>
21921 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
21922 (insn_mnz_v8qi): ... this ...
21923 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
21925 (insn_v<n>mnz): Replaced by ...
21926 (insn_v1mnz): ... this ...
21927 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
21929 (insn_mz_<mode>): Replaced by ...
21930 (insn_mz_v8qi): ... this ...
21931 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
21933 (insn_v<n>mz): Replaced by ...
21934 (insn_v1mz): ... this ...
21935 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
21938 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
21940 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
21942 2013-03-26 Roland McGrath <mcgrathr@google.com>
21944 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
21945 than fprintf with a non-constant, non-format string.
21947 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21949 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
21950 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
21951 operand 0 predicate.
21952 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
21953 attribute. Use general_x64nomem_operand as operand 1 predicate.
21954 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
21955 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21956 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
21957 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21958 (mov<mode>_insv_1): Remove expander. Merge insn with
21959 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
21960 Use general_x64nomem_operand as operand 1 predicate.
21961 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
21962 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
21963 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21964 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
21965 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
21966 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
21967 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21968 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
21969 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
21970 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
21971 (general_x64nomem_operand): Ditto.
21973 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
21975 * config/rtems.opt: Add -pthread option.
21977 2013-03-26 Richard Biener <rguenther@suse.de>
21979 * alias.c (find_base_term): Avoid redundant and not used recursion.
21980 (base_alias_check): Get the initial base term from the caller.
21981 (true_dependence_1): Compute and pass base terms to base_alias_check.
21982 (write_dependence_p): Likewise.
21983 (may_alias_p): Likewise.
21985 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
21987 * config/aarch64/aarch64.c (aarch64_classify_address): Support
21988 PC-relative load in SI modes and above only.
21990 2013-03-26 Xinyu Qi <xyqi@marvell.com>
21992 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
21993 * config/arm/iwmmxt.md (WCGR0): Update.
21994 (WCGR1, WCGR2, WCGR3): Likewise.
21996 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21998 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
21999 Use x64 and nox64 isa attributes.
22001 2013-03-26 Richard Biener <rguenther@suse.de>
22003 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
22004 alignment computations and rely on get_object_alignment_1
22005 for the !TYPE_P case.
22006 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
22008 2013-03-26 Walter Lee <walt@tilera.com>
22010 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
22011 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
22013 2013-03-25 Jeff Law <law@redhat.com>
22015 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
22016 check for INTEGRAL_TYPE_P that was missing due to checking in
22017 wrong version of prior patch.
22019 2013-03-25 Walter Lee <walt@tilera.com>
22021 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
22022 TILEGX_INSN_SHUFFLEBYTES1.
22023 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
22025 (tilegx_builtins): Ditto.
22026 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
22028 2013-03-25 Walter Lee <walt@tilera.com>
22030 * config/tilegx/tilegx.md (floatsisf2): New pattern.
22031 (floatunssisf2): New pattern.
22032 (floatsidf2): New pattern.
22033 (floatunssidf2): New pattern.
22035 2013-03-25 Walter Lee <walt@tilera.com>
22037 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
22038 tests for constraint J, K, N, P.
22040 2013-03-25 Walter Lee <walt@tilera.com>
22042 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
22043 Use indirect/pcrel encoding.
22044 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
22047 2013-03-25 Steve Ellcey <sellcey@mips.com>
22049 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
22050 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
22051 * config/mips/mips.c (mips_option_override): Set IMADD default.
22052 * config/mips/mips.h (PTF_AVOID_IMADD): New.
22053 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
22054 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
22055 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
22056 * doc/invoke.texi (-mimadd/-mno-imadd): New.
22058 2013-03-25 Jeff Law <law@redhat.com>
22060 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
22061 slightly to avoid creating and folding useless trees. Simplify
22062 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
22064 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
22066 * config/i386/i386.md (*zero_extendsidi2): Merge with
22067 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
22068 * config/i386/predicates.md (x86_64_zext_operand): Rename from
22069 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
22070 targets. Clarify comment.
22072 2013-03-25 Martin Jambor <mjambor@suse.cz>
22074 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
22075 pass-through jump functions differently.
22076 (ipa_read_jump_function): Likewise. Also use setter functions to set
22079 2013-03-25 Martin Jambor <mjambor@suse.cz>
22081 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
22082 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
22084 (ipa_get_indirect_edge_target): New function.
22085 (devirtualization_time_bonus): New parameter known_aggs, pass it to
22086 ipa_get_indirect_edge_target. Update all callers.
22087 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
22088 ipa_get_indirect_edge_target_1 instead of calling
22089 ipa_get_indirect_edge_target.
22090 (create_specialized_node): Pass aggvlas to
22091 ipcp_discover_new_direct_edges.
22093 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22095 * config/arm/arm.md (f_sels, f_seld): New types.
22096 (*cmov<mode>): New pattern.
22097 * config/arm/predicates.md (arm_vsel_comparison_operator): New
22100 2013-03-25 Kai Tietz <ktietz@redhat.com>
22102 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
22103 POSIX-printf for mingw-hosted builds.
22105 2013-03-25 Richard Biener <rguenther@suse.de>
22107 PR middle-end/56694
22108 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
22109 must-not-throw stmt location.
22111 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22113 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
22114 Emit load-acquire versions when acq is true.
22115 (arm_emit_store_exclusive): Add rel parameter.
22116 Emit store-release versions when rel is true.
22117 (arm_split_compare_and_swap): Use acquire-release instructions
22119 of barriers when appropriate.
22120 (arm_split_atomic_op): Likewise.
22121 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
22122 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
22123 (VUNSPEC_SLX): Likewise.
22124 (VUNSPEC_LDA): Likewise.
22125 (VUNSPEC_STL): Likewise.
22126 * config/arm/sync.md (atomic_load<mode>): New pattern.
22127 (atomic_store<mode>): Likewise.
22128 (arm_load_acquire_exclusive<mode>): Likewise.
22129 (arm_load_acquire_exclusivesi): Likewise.
22130 (arm_load_acquire_exclusivedi): Likewise.
22131 (arm_store_release_exclusive<mode>): Likewise.
22133 2013-03-25 Catherine Moore <clm@codesourcery.com>
22135 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
22136 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
22137 * config/mip/predicates.md (lwsp_swsp_operand,
22138 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
22139 sb16_operand, db4_operand, db7_operand, ib3_operand,
22140 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
22141 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
22142 andi16_operand): New predicates.
22143 * config/mips/mips.md (compression): New attribute.
22144 (enabled): New attribute.
22145 (length): Consider compression in computing length.
22146 (shift_compression): New code attribute.
22147 (*add<mode>3): New operands. Record compression.
22148 (sub<mode>3): Likewise.
22149 (one_cmpl<mode>2): Likewise.
22150 (*and<mode>3): Likewise.
22151 (*ior<mode>3): Likewise.
22152 (unnamed pattern for xor): Likewise.
22153 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
22154 (*<optab><mode>3): Likewise.
22155 (*mov<mode>_internal: Likewise.
22156 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
22157 (mips_unsigned_immediate_p): New.
22158 (umips_lwsp_swsp_address_p): New.
22159 (m16_based_address_p): New.
22160 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
22161 (mips_unsigned_immediate_p): New prototype.
22162 (lwsp_swsp_address_p): New prototype.
22163 (m16_based_address_p): New prototype.
22164 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
22165 (mips_signed_immediate_p): New function.
22166 (m16_based_address_p): New function.
22167 (lwsp_swsp_address_p): New function.
22168 (mips_print_operand_punctuation): Recognize short delay slot insns
22169 for microMIPS.add<mode>3"
22171 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22174 * config/arm/iterators.md (v_cmp_result): New mode attribute.
22175 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
22177 2013-03-25 Richard Biener <rguenther@suse.de>
22179 PR tree-optimization/56689
22180 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
22183 2013-03-25 Richard Biener <rguenther@suse.de>
22185 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
22187 (memory_references): Likewise.
22188 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
22189 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
22190 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
22191 (gather_mem_refs_in_loops): Fold into ...
22192 (analyze_memory_references): ... this. Move initialization
22193 to tree_ssa_lim_initialize.
22194 (fill_always_executed_in): Rename to ...
22195 (fill_always_executed_in_1): ... this.
22196 (fill_always_executed_in): Move contains_call computation to
22197 this new function from ...
22198 (tree_ssa_lim_initialize): ... here.
22199 (tree_ssa_lim): Call fill_always_executed_in.
22201 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
22203 * postreload.c (reload_combine): Fix code detecting returns.
22205 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
22207 * function.c (emit_use_return_register_into_block): On cc0 targets,
22208 do not emit the sequence between cc0 setter and user.
22210 2013-03-25 Kai Tietz <ktietz@redhat.com>
22212 * config/i386/predicates.md (local_symbolic_operand): Interpret
22213 dll-imported symbols as none-local.
22215 2013-03-25 Richard Biener <rguenther@suse.de>
22217 * tree-ssa-loop-im.c (struct depend): Remove.
22218 (struct lim_aux_data): Make depends a vec of gimples.
22219 (free_lim_aux_data): Adjust.
22220 (add_dependency): Likewise.
22221 (set_level): Likewise.
22223 2013-03-25 Richard Biener <rguenther@suse.de>
22225 PR middle-end/56434
22226 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
22227 the pointer returned by calls with ECF_MALLOC set.
22229 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
22231 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
22233 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
22235 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
22236 using MMXMODE mode iterator.
22237 (*move<mode>_internal): Merge with *movv2sf_internal and
22238 *movv2sf_internal_rex64 using MMXMODE mode iterator.
22240 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22242 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
22243 (record_last_mem_set_info): Likewise.
22245 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
22246 of XNEWVEC followed by memset.
22247 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
22249 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22251 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
22252 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
22253 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
22254 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
22255 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
22256 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
22257 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
22258 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
22259 BARRIER_P instead of GET_CODE.
22261 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22263 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
22264 inaccuracy in the probing code.
22266 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
22267 (ctrapdi4): Likewise.
22269 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22271 * calls.c (expand_call): Add missing guard to code handling return
22272 of non-BLKmode structures in MSB.
22273 * function.c (expand_function_end): Likewise.
22275 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22277 * combine.c (try_combine): Adjust comment. Do not add the set of
22278 insn #0 if the destination indirectly is set or dies in insn #2.
22279 Tidy up code to distribute a new note.
22281 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22283 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
22284 also for alternatives 16 and 17.
22286 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22288 * config/i386/sse.md (*mov<mode>_internal): Merge with
22289 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
22290 Emit insn template depending on type attribute. Use
22291 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22292 movd instead of movq mnemonic for interunit moves. Rewrite mode
22293 attribute calculation. Remove unit attribute calculation.
22294 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
22295 Set prefix_data16 attribute for DImode ssemov types.
22296 Use Ym instead of y for SSE-MMX conversion alternatives.
22297 Reorder operand constraints.
22299 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
22301 * df.h (df_insn_delete): Adjust prototype.
22302 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
22303 and let it decide whether mark the basic block dirty.
22304 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
22305 * df-scan.c (df_insn_info_delete): New helper function, split
22306 off from df_insn_delete.
22307 (df_scan_free_bb_info): Use it.
22308 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
22310 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
22311 that the insn is actually an insn and it has a non-NULL basic block.
22312 Do not mark basic block dirty if only deleting a DEBUG_INSN.
22314 2013-03-22 Richard Biener <rguenther@suse.de>
22316 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
22318 (mem_ref_alloc): Do not allocate them.
22319 (refs_independent_p): Do not query or maintain a cache.
22321 2013-03-22 Richard Biener <rguenther@suse.de>
22323 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
22324 (gather_mem_refs_in_loops): Do not compute it.
22325 (analyze_memory_references): Do not allocate it.
22326 (tree_ssa_lim_finalize): Do not free it.
22327 (for_all_locs_in_loop): Do not query all_refs_in_loop.
22329 2013-03-22 Richard Biener <rguenther@suse.de>
22331 * is-a.h (as_a): Use gcc_checking_assert.
22333 2013-03-22 Ian Bolton <ian.bolton@arm.com>
22335 * config/aarch64/aarch64.c (aarch64_print_operand): New
22336 format specifier for printing a constant in hex.
22337 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
22338 format specifier for printing second operand.
22340 2013-03-22 Richard Biener <rguenther@suse.de>
22342 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
22344 (gather_mem_refs_in_loops): Perform store accumulation here.
22345 (create_vop_ref_mapping_loop): Remove.
22346 (create_vop_ref_mapping): Likewise.
22347 (analyze_memory_references): Initialize refs_stored_in_loop.
22348 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
22349 (record_indep_loop): Remove.
22350 (record_dep_loop): New function.
22351 (ref_indep_loop_p_1): Adjust to only walk over references
22352 in the loop, not its subloops.
22353 (ref_indep_loop_p): Rename to ...
22354 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
22355 maintaining a more fine-grained cache.
22356 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
22357 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
22359 2013-03-22 Richard Biener <rguenther@suse.de>
22361 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
22362 (struct mem_ref): Make accesses_in_loop a vec of a vec of
22363 aggregate mem_ref_loc.
22364 (free_mem_ref_locs): Inline into ...
22365 (memref_free): ... this and adjust.
22366 (mem_ref_alloc): Adjust.
22367 (mem_ref_locs_alloc): Remove.
22368 (record_mem_ref_loc): Adjust.
22369 (get_all_locs_in_loop): Rewrite into ...
22370 (for_all_locs_in_loop): ... this iterator.
22371 (rewrite_mem_ref_loc): New functor.
22372 (rewrite_mem_refs): Use for_all_locs_in_loop.
22373 (sm_set_flag_if_changed): New functor.
22374 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
22375 (ref_always_accessed): New functor.
22376 (ref_always_accessed_p): Use for_all_locs_in_loop.
22378 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22380 * tree-pass.h (PROP_gimple_lvec): New.
22381 * passes.c (dump_properties): Handle PROP_gimple_lvec.
22382 (init_optimization_passes): Move pass_lower_vector.
22383 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
22385 (pass_lower_vector): Provide PROP_gimple_lvec.
22386 (pass_lower_vector_ssa): Likewise.
22387 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
22389 2013-03-21 Mark Wielaard <mjw@redhat.com>
22391 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
22393 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22395 * config/i386/i386.md (*movdi_internal): Disparage slightly
22396 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
22397 conversion alternatives.
22399 2013-03-21 Jakub Jelinek <jakub@redhat.com>
22401 PR middle-end/48087
22402 * diagnostic.def (DK_WERROR): New kind.
22403 * diagnostic.h (werrorcount): Define.
22404 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
22405 promoted to DK_ERROR, increment DK_WERROR counter instead of
22407 * toplev.c (toplev_main): Call print_ignored_options even if
22408 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
22409 even if just werrorcount is non-zero.
22412 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
22414 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
22415 (string_cst_pool_decl): New function.
22416 (optimize_one_addr_into_implicit_ptr): New function.
22417 (resolve_addr_in_expr): Optimize DWARF location expression
22418 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
22419 which doesn't live in memory, but has DW_AT_location or
22420 DW_AT_const_value, or refers to a string literal, into
22421 DW_OP_GNU_implicit_pointer.
22422 (optimize_location_into_implicit_ptr): New function.
22423 (resolve_addr): If removing DW_AT_location of a variable because
22424 it was DW_OP_addr of address of the variable, but the variable doesn't
22425 live in memory, try to emit const value attribute for the initializer.
22427 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22429 * tree.h (VECTOR_TYPE_P): New macro.
22430 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
22431 TYPE_MODE): Use it.
22432 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
22433 VEC_COND_EXPR cannot be lvalues.
22434 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
22436 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22438 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
22439 Restrict the transformation to equal modes.
22441 2013-03-21 Richard Biener <rguenther@suse.de>
22443 PR tree-optimization/39326
22444 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
22445 (MEM_ANALYZABLE): Adjust.
22446 (record_mem_ref_loc): Move bitmap ops ...
22447 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
22448 unanalyzable refs, do not record locations for it.
22449 (analyze_memory_references): Allocate ref zero as shared
22451 (refs_independent_p): Do not test for unanalyzed mems here.
22452 (ref_indep_loop_p_1): Special-case disambiguation against
22453 the unanalyzed ref.
22454 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
22456 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
22458 * config/arm/arm-protos.h (tune_params): Add
22459 prefer_neon_for_64bits field.
22460 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
22461 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
22462 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
22463 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
22464 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
22465 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
22466 (arm_option_override): Handle -mneon-for-64bits new option.
22467 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
22468 (prefer_neon_for_64bits): Declare new variable.
22469 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
22470 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
22471 (arch_enabled): Handle new arch types. Remove support for onlya8
22473 (one_cmpldi2): Use new arch names.
22474 (zero_extend<mode>di2, extend<mode>di2): Ditto.
22475 * config/arm/arm.opt (mneon-for-64bits): Add option.
22476 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
22477 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
22478 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
22480 * doc/invoke.texi (-mneon-for-64bits): Document.
22482 2013-03-21 Richard Biener <rguenther@suse.de>
22484 PR tree-optimization/39326
22485 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
22486 (sort_bbs_in_loop_postorder_cmp): New function.
22487 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
22489 2013-03-21 Richard Biener <rguenther@suse.de>
22491 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
22492 (vect_insert_into_interleaving_chain): Likewise.
22493 (vect_drs_dependent_in_basic_block): Inline ...
22494 (vect_slp_analyze_data_ref_dependence): ... here. New function,
22496 (vect_analyze_data_ref_dependence): ... here. Simplify.
22497 (vect_check_interleaving): Simplify.
22498 (vect_analyze_data_ref_dependences): Likewise. Split out ...
22499 (vect_slp_analyze_data_ref_dependences): ... this new function.
22500 (dr_group_sort_cmp): New function.
22501 (vect_analyze_data_ref_accesses): Compute data-reference groups
22502 here instead of in vect_analyze_data_ref_dependence. Use
22503 a more efficient algorithm.
22504 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
22505 vect_slp_analyze_data_ref_dependences. Call
22506 vect_analyze_data_ref_accesses earlier.
22507 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
22508 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
22509 (vect_slp_analyze_data_ref_dependences): New prototype.
22511 2013-03-21 Richard Biener <rguenther@suse.de>
22513 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
22514 ref is stored in the loop.
22515 (find_refs_for_sm): Walk only over all stores.
22516 (store_motion_loop): Allocate from lim_bitmap_obstack.
22517 (store_motion): Likewise.
22519 2013-03-21 Richard Biener <rguenther@suse.de>
22521 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
22522 Update virtual SSA form.
22524 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22526 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
22527 * configure: Regenerate.
22528 * config.in: Regenerate.
22529 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
22530 if !HAVE_LD_EH_FRAME_CIEV3.
22532 2013-03-21 Richard Biener <rguenther@suse.de>
22534 * tree-cfg.c (verify_expr_no_block): New function.
22535 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
22536 nor DECL_VALUE_EXPR have locations with associated blocks.
22537 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
22538 (clear_unused_block_pointer): Remove code dealing with
22539 blocks in DECL_DEBUG_EXPR locations.
22541 2013-03-21 Richard Biener <rguenther@suse.de>
22543 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
22544 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
22545 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
22546 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
22547 instead of DECL_DEBUG_EXPR_IS_FROM.
22548 * gimplify.c (gimplify_modify_expr): Likewise.
22549 * tree-cfg.c (verify_expr_location_1): Likewise.
22550 * tree-complex.c (create_one_component_var): Likewise.
22551 * tree-sra.c (create_access_replacement): Likewise.
22552 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
22553 (clear_unused_block_pointer): Likewise.
22554 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
22555 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
22556 * var-tracking.c (var_debug_decl): Likewise.
22557 (track_expr_p): Likewise.
22558 * tree-inline.c (add_local_variables): Likewise. Set
22559 DECL_HAS_DEBUG_EXPR_P after copying it.
22560 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
22561 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
22563 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22566 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
22567 * configure: Regenerate.
22568 * config.in: Regenerate.
22569 * config/i386/i386.md (*movdf_internal): Use
22570 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22571 movd instead of movq mnemonic for interunit moves.
22572 (*movdi_internal): Ditto.
22574 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22576 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
22577 (abd<mode>_3): New pattern.
22578 (aba<mode>_3): New pattern.
22579 (fabd<mode>_3): New pattern.
22581 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22583 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
22584 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
22585 occurrence of REGISTER_PREFIX as its empty string.
22587 2013-03-20 Jeff Law <law@redhat.com>
22589 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
22590 addititional equivalences for equality comparisons between an SSA_NAME
22591 and a constant where the SSA_NAME was set from a widening conversion.
22593 2013-03-20 Walter Lee <walt@tilera.com>
22595 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
22597 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22599 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
22600 depending on type attribute.
22601 (*movti_internal): Ditto.
22602 (*movtf_internal): Ditto.
22603 (*movxf_internal): Ditto.
22604 (*movdf_internal): Ditto.
22605 (*movsf_internal): Ditto.
22607 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22609 * config/i386/i386.md (*movti_internal): Set prefix attribute to
22610 maybe_vex for sselog1 and ssemov types.
22611 (*movdi_internal): Reorder operand constraints.
22612 (*movsi_internal): Ditto. Set prefix attribute to
22613 maybe_vex for sselog1 and ssemov types.
22614 (*movtf_internal): Set prefix attribute to maybe_vex
22615 for sselog1 and ssemov types.
22616 (*movdf_internal): Ditto. Set prefix_data16 attribute for
22617 DImode ssemov types. Reorder operand constraints.
22618 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
22619 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
22620 attribute for SImode ssemov types. Reorder operand constraints.
22622 2013-03-20 Martin Jambor <mjambor@suse.cz>
22624 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
22625 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
22627 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
22629 * config/rs6000/predicates.md (indexed_address, update_address_mem
22630 update_indexed_address_mem): New predicates.
22631 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
22632 attribute for load/store instructions.
22633 * config/rs6000/dfp.md (movsd_store): Likewise.
22634 (movsd_load): Likewise.
22635 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
22636 (unnamed HI->DI extend define_insn): Likewise.
22637 (unnamed SI->DI extend define_insn): Likewise.
22638 (unnamed QI->SI extend define_insn): Likewise.
22639 (unnamed QI->HI extend define_insn): Likewise.
22640 (unnamed HI->SI extend define_insn): Likewise.
22641 (unnamed HI->SI extend define_insn): Likewise.
22642 (extendsfdf2_fpr): Likewise.
22643 (movsi_internal1): Likewise.
22644 (movsi_internal1_single): Likewise.
22645 (movhi_internal): Likewise.
22646 (movqi_internal): Likewise.
22647 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
22648 attribute for load/store instructions.
22649 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
22651 (mov<mode>_softfloat): Likewise.
22652 (mov<mode>_hardfloat32): Likewise.
22653 (mov<mode>_hardfloat64): Likewise.
22654 (mov<mode>_softfloat64): Likewise.
22655 (movdi_internal32): Likewise.
22656 (movdi_internal64): Likewise.
22657 (probe_stack_<mode>): Likewise.
22659 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
22661 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
22662 floating point, and decimal floating point to reload iterator.
22664 * config/rs6000/constraints.md (wl constraint): New constraints to
22665 return FLOAT_REGS if certain options are used to reduce the number
22666 of separate patterns that exist in the file.
22667 (wx constraint): Likewise.
22668 (wz constraint): Likewise.
22670 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22671 -mdebug=reg, print wg, wl, wx, and wz constraints.
22672 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22673 Initialize the reload functions for 64-bit binary/decimal floating
22675 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22676 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22677 create the buffer on the stack to overcome not having a 32-bit
22679 (rs6000_emit_move): Likewise.
22680 (rs6000_secondary_memory_needed_rtx): Likewise.
22681 (rs6000_alloc_sdmode_stack_slot): Likewise.
22682 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22683 via xxlxor, just like DFmode 0.0.
22685 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
22686 define as 1 if we are running on a power7 or newer.
22687 (enum r6000_reg_class_enum): Add new constraints.
22689 * config/rs6000/dfp.md (movsd): Delete, combine with binary
22690 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22691 with other moves by using conditional constraits (wg). Use LFIWZX
22692 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
22693 (movsd splitter): Likewise.
22694 (movsd_hardfloat): Likewise.
22695 (movsd_softfloat): Likewise.
22697 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
22698 binary and decimal floating point moves.
22699 (fmove_ok): New attributes to combine binary and decimal floating
22700 point moves, and to combine power6x (mfpgpr) moves along normal
22702 (real_value_to_target): Likewise.
22703 (f32_lr): Likewise.
22704 (f32_lm): Likewise.
22705 (f32_li): Likewise.
22706 (f32_sr): Likewise.
22707 (f32_sm): Likewise.
22708 (f32_si): Likewise.
22709 (movsf): Combine binary and decimal floating point moves. Combine
22710 power6x (mfpgpr) moves with other moves by using conditional
22711 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
22712 (mov<mode> for SFmode/SDmode); Likewise.
22713 (SFmode/SDmode splitters): Likewise.
22714 (movsf_hardfloat): Likewise.
22715 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
22716 (movsf_softfloat): Likewise.
22717 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
22719 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
22720 wx and wz constraints.
22722 * config/rs6000/constraints.md (wg constraint): New constraint to
22723 return FLOAT_REGS if -mmfpgpr (power6x) was used.
22725 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
22728 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22729 -mdebug=reg, print wg, wl, wx, and wz constraints.
22730 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22731 Initialize the reload functions for 64-bit binary/decimal floating
22733 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22734 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22735 create the buffer on the stack to overcome not having a 32-bit
22737 (rs6000_emit_move): Likewise.
22738 (rs6000_secondary_memory_needed_rtx): Likewise.
22739 (rs6000_alloc_sdmode_stack_slot): Likewise.
22740 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22741 via xxlxor, just like DFmode 0.0.
22743 * config/rs6000/dfp.md (movdd): Delete, combine with binary
22744 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22745 with other moves by using conditional constraits (wg). Use LFIWZX
22746 and STFIWX for loading SDmode on power7.
22747 (movdd splitters): Likewise.
22748 (movdd_hardfloat32): Likewise.
22749 (movdd_softfloat32): Likewise.
22750 (movdd_hardfloat64_mfpgpr): Likewise.
22751 (movdd_hardfloat64): Likewise.
22752 (movdd_softfloat64): Likewise.
22754 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
22755 64-bit binary and decimal floating point moves.
22756 (FMOVE64X): Likewise.
22757 (movdf): Combine 64-bit binary and decimal floating point moves.
22758 Combine power6x (mfpgpr) moves with other moves by using
22759 conditional constraits (wg).
22760 (mov<mode> for DFmode/DDmode): Likewise.
22761 (DFmode/DDmode splitters): Likewise.
22762 (movdf_hardfloat32): Likewise.
22763 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
22764 (movdf_softfloat32): Likewise.
22765 (movdf_hardfloat64_mfpgpr): Likewise.
22766 (movdf_hardfloat64): Likewise.
22767 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
22768 (movdf_softfloat64): Likewise.
22769 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
22770 (reload_<mode>_load): Move to later in the file so they aren't in
22771 the middle of the floating point move insns.
22772 (reload_<mode>_store): Likewise.
22774 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
22777 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
22778 constraint if -mdebug=reg.
22779 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
22780 Enable using dd reload support if needed.
22782 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
22783 binary and decimal floating point moves in rs6000.md.
22784 (movtd_internal): Likewise.
22786 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
22787 decimal floating point moves.
22789 (movtf_internal): Likewise.
22790 (mov<mode>_internal, TDmode/TFmode): Likewise.
22791 (movtf_softfloat): Likewise.
22792 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
22794 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
22795 movdi_internal64, using wg constraint for move direct operations.
22796 (movdi_internal64): Likewise.
22798 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
22799 MODES_TIEABLE_P for selected modes. Print the numerical value of
22800 the various virtual registers. Use GPR/FPR first/last values,
22801 instead of hard coding the register numbers. Print which modes
22802 have reload functions registered.
22803 (rs6000_option_override_internal): If -mdebug=reg, trace the options
22804 settings before/after setting cpu, target and subtarget settings.
22805 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
22806 and for secondary reload failures in rs6000_secondary_reload_inner.
22807 (rs6000_secondary_reload_fail): Likewise.
22808 (rs6000_secondary_reload_inner): Likewise.
22810 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
22811 macros for first/last GPR and FPR registers.
22812 (LAST_GPR_REGNO): Likewise.
22813 (FIRST_FPR_REGNO): Likewise.
22814 (LAST_FPR_REGNO): Likewise.
22816 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
22817 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
22818 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
22819 (vcond<mode><mode>): Likewise.
22820 (vcondu<mode><mode>): Likewise.
22821 (vector_gtu<mode>): Likewise.
22822 (vector_gte<mode>): Likewise.
22823 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
22824 to prevent the compiler from converting DImode operations to TImode.
22825 (ior<mode>3): Likewise.
22826 (and<mode>3): Likewise.
22827 (one_cmpl<mode>2): Likewise.
22828 (nor<mode>3): Likewise.
22829 (andc<mode>3): Likewise.
22831 * config/rs6000/constraints.md (wt constraint): New constraint
22832 that returns VSX_REGS if TImode is allowed in VSX registers.
22834 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
22835 constant under VSX.
22837 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
22838 similar to TImode, but it is restricted to being in the GPRs.
22840 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
22841 TImode to occupy a single VSX register.
22843 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
22844 -mvsx-timode for power7/power8.
22845 (power7 cpu): Likewise.
22846 (power8 cpu): Likewise.
22848 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
22849 sure that TFmode/TDmode take up two registers if they are ever
22850 allowed in the upper VSX registers.
22851 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
22853 (rs6000_init_hard_regno_mode_ok): Likewise.
22854 (rs6000_debug_reg_global): Add debugging for PTImode and wt
22855 constraint. Print if LRA is turned on.
22856 (rs6000_option_override_internal): Give an error if -mvsx-timode
22857 and VSX is not enabled.
22858 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
22859 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
22860 to reg+offset addressing. Use PTImode when checking offset
22861 addresses for validity.
22862 (reg_offset_addressing_ok_p): Likewise.
22863 (rs6000_legitimate_offset_address_p): Likewise.
22864 (rs6000_legitimize_address): Likewise.
22865 (rs6000_legitimize_reload_address): Likewise.
22866 (rs6000_legitimate_address_p): Likewise.
22867 (rs6000_eliminate_indexed_memrefs): Likewise.
22868 (rs6000_emit_move): Likewise.
22869 (rs6000_secondary_reload): Likewise.
22870 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
22871 reloads to fpr registers to continue to use reg+offset addressing,
22872 but 64-bit reloads to altivec registers need reg+reg addressing.
22873 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
22874 it. Treat LO_SUM like a PLUS operation.
22875 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
22876 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
22877 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
22878 registers to share a register with a smaller sized type, since VSX
22879 puts scalars in the upper 64-bits.
22880 (print_operand): Add support for PTImode.
22881 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
22882 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
22883 registers, but don't have arithmetic support.
22884 (rs6000_memory_move_cost): Add test for VSX.
22885 (rs6000_opt_masks): Add -mvsx-timode.
22887 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
22890 (VSr): Use wt constraint for TImode.
22891 (VSv): Drop TImode support.
22892 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
22893 (vsx_movti_64bit): Likewise.
22894 (vsx_movti_32bit): Likewise.
22895 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
22896 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
22897 one '?' on the appropriate output constraint. Do not allow TImode
22898 logical operations on 32-bit systems.
22899 (vsx_ior<mode>3): Likewise.
22900 (vsx_xor<mode>3): Likewise.
22901 (vsx_one_cmpl<mode>2): Likewise.
22902 (vsx_nor<mode>3): Likewise.
22903 (vsx_andc<mode>3): Likewise.
22904 (vsx_concat_<mode>): Likewise.
22905 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
22907 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
22908 OPTION_MASK_VSX_TIMODE.
22909 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
22910 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
22912 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
22913 (TI2 iterator): New iterator for TImode, PTImode.
22914 (wd mode attribute): Add values for vector types.
22915 (movti_string): Replace TI move operations with operations for TImode
22916 and PTImode. Add support for TImode being allowed in VSX registers.
22917 (mov<mode>_string, TImode/PTImode): Likewise.
22918 (movti_ppc64): Likewise.
22919 (mov<mode>_ppc64, TImode/PTImode): Likewise.
22920 (TI mode splitters): Likewise.
22922 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
22925 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
22927 PR tree-optimization/56355
22928 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
22929 Also handle integers with undefined overflow.
22931 2013-03-20 Catherine Moore <clm@codesourcery.com>
22932 Maciej W. Rozycki <macro@codesourcery.com>
22933 Tom de Vries <tom@codesourcery.com>
22934 Nathan Sidwell <nathan@codesourcery.com>
22935 Iain Sandoe <iain@codesourcery.com>
22936 Nathan Froyd <froydnj@codesourcery.com>
22937 Chao-ying Fu <fu@mips.com>
22939 * doc/extend.texi: (micromips, nomicromips, nocompression):
22940 Document new function attributes.
22941 * doc/invoke.texi (minterlink-compressed, mmicromips,
22942 m14k, m14ke, m14kec): Document new options.
22943 (minterlink-mips16): Update documentation.
22944 * doc/md.texi (ZC, ZD): Document new constraints.
22945 * configure.ac (gcc_cv_as_micromips): Check if linker
22946 supports the .set micromips directive.
22947 * configure: Regenerate.
22948 * config.in: Regenerate.
22949 * config/mips/mips-tables.opt: Regenerate.
22950 * config/mips/micromips.md: New file.
22951 * constraints.md (ZC, ZD): New constraints.
22952 * config/mips/predicates.md (movep_src_register): New predicate.
22953 (movep_src_operand): New predicate.
22954 (non_volatile_mem_operand): New predicate.
22955 * config/mips/mips.md (multimem): New type.
22956 (length): Differentiate between 17-bit and 18-bit branch offsets.
22957 (MOVEP1, MOVEP2): New mode iterator.
22958 (mov_<load>l): Use ZC constraint.
22959 (mov_<load>r): Likewise.
22960 (mov_<store>l): Likewise.
22961 (mov_<store>r): Likewise.
22962 (*branch_equality<mode>_inverted): Add microMIPS support.
22963 (*branch_equality<mode>): Likewise.
22964 (*jump_absolute): Likewise.
22965 (indirect_jump_<mode>): Likewise.
22966 (tablejump_<mode>): Likewise.
22967 (<optab>_internal): Likewise.
22968 (sibcall_internal): Likewise.
22969 (sibcall_value_internal): Likewise.
22970 (prefetch): Use constraint ZD.
22971 * config/mips/mips.opt (minterlink-compressed): New option.
22972 (minterlink-mips16): Now an alias for minterlink-compressed.
22973 (mmicromips): New option.
22974 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
22975 (compare_and_swap_12): Likewise.
22976 (sync_add<mode>): Likewise.
22977 (sync_<optab>_12): Likewise.
22978 (sync_old_<optab>_12): Likewise.
22979 (sync_new_<optab>_12): Likewise.
22980 (sync_nand_12): Likewise.
22981 (sync_old_nand_12): Likewise.
22982 (sync_new_nand_12): Likewise.
22983 (sync_sub<mode>): Likewise.
22984 (sync_old_add<mode>): Likewise.
22985 (sync_old_sub<mode>): Likewise.
22986 (sync_new_add<mode>): Likewise.
22987 (sync_new_sub<mode>): Likewise.
22988 (sync_<optab><mode>): Likewise.
22989 (sync_old_<optab><mode>): Likewise.
22990 (sync_new_<optab><mode>): Likewise.
22991 (sync_nand<mode>): Likewise.
22992 (sync_old_nand<mode>): Likewise.
22993 (sync_new_nand<mode>): Likewise.
22994 (sync_lock_test_and_set<mode>): Likewise.
22995 (test_and_set_12): Likewise.
22996 (atomic_compare_and_swap<mode>): Likewise.
22997 (atomic_exchange<mode>_llsc): Likewise.
22998 (atomic_fetch_add<mode>_llsc): Likewise.
22999 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23000 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23001 (umips_save_restore_pattern_p): Likewise.
23002 (umips_load_store_pair_p): Likewise.
23003 (umips_output_load_store_pair): Likewise.
23004 (umips_movep_target_p): Likewise.
23005 (umips_12bit_offset_address_p): Likewise.
23006 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
23007 (mips_base_mips16): Rename this...
23008 (mips_base_compression_flags): ...to this. Update all uses.
23009 (mips_attribute_table): Add micromips, nomicromips and nocompression.
23010 (mips_mips16_decl_p): Delete.
23011 (mips_nomips16_decl_p): Delete.
23012 (mips_get_compress_on_flags): New function.
23013 (mips_get_compress_off_flags): New function.
23014 (mips_get_compress_mode): New function.
23015 (mips_get_compress_on_name): New function.
23016 (mips_get_compress_off_name): New function.
23017 (mips_insert_attributes): Support multiple compression types.
23018 (mips_merge_decl_attributes): Likewise.
23019 (umips_12bit_offset_address_p): New function.
23020 (mips_start_function_definition): Emit .set micromips directive.
23021 (mips_call_may_need_jalx_p): New function.
23022 (mips_function_ok_for_sibcall): Add microMIPS support.
23023 (mips_print_operand_punctuation): Support short delay slots and
23025 (umips_swm_mask, umips_swm_encoding): New.
23026 (umips_build_save_restore): New function.
23027 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
23028 (was_mips16_p): Remove.
23029 (old_compression_mode): New.
23030 (mips_set_compression_mode): New function.
23031 (mips_set_current_function): Add microMIPS support.
23032 (mips_option_override): Likewise.
23033 (umips_save_restore_pattern_p): New function.
23034 (umips_output_save_restore): New function.
23035 (umips_load_store_pair_p_1): New function.
23036 (umips_load_store_pair_p): New function.
23037 (umips_output_load_store_pair_1): New function.
23038 (umips_output_load_store_pair): New function.
23039 (umips_movep_target_p) New function.
23040 (mips_prepare_pch_save): Add microMIPS support.
23041 * config/mips/mips.h (TARGET_COMPRESSION): New.
23042 (TARGET_CPU_CPP_BUILTINS): Update macro
23043 to use new compression flags and to support microMIPS.
23044 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
23045 (MIPS_ARCH_FLOAT_SPEC): Likewise.
23046 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
23047 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
23048 (ASM_SPEC): Support mmicromips and mno-micromips.
23049 (M16STORE_REG_P): New macro.
23050 (MIPS_CALL): Support TARGET_MICROMIPS.
23051 (MICROMIPS_J): New macro.
23052 (mips_base_mips16): Rename this...
23053 (mips_base_compression_flags): ...to this.
23054 (UMIPS_12BIT_OFFSET_P): New macro.
23055 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
23056 (MULTILIB_DIRNAMES): Likewise.
23057 2013-03-20 Richard Biener <rguenther@suse.de>
23059 PR tree-optimization/56661
23060 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
23061 the result does not have to be distinct.
23063 2013-03-20 Richard Biener <rguenther@suse.de>
23065 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
23068 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23069 Steven Bosscher <steven@gcc.gnu.org>
23071 PR rtl-optimization/56605
23072 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
23074 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
23077 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
23078 that require movd instead of movq.
23080 2013-03-20 Richard Biener <rguenther@suse.de>
23082 * tree-ssa-structalias.c (struct variable_info): Add pointer
23083 to the first field of an aggregate with sub-vars. Make
23084 this and the pointer to the next subfield its ID.
23085 (vi_next): New function.
23086 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
23087 storedanything_id, integer_id): Increment by one.
23088 (new_var_info, get_call_vi, lookup_call_clobber_vi,
23089 get_call_clobber_vi): Adjust.
23090 (solution_set_expand): Simplify and speedup.
23091 (solution_set_add): Inline into ...
23092 (set_union_with_increment): ... this. Adjust accordingly.
23093 (do_sd_constraint): Likewise.
23094 (do_ds_constraint): Likewise.
23095 (do_complex_constraint): Simplify.
23096 (build_pred_graph): Adjust.
23097 (solve_graph): Likewise. Simplify and speedup.
23098 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
23099 get_constraint_for_component_ref, get_constraint_for_1,
23100 first_vi_for_offset, first_or_preceding_vi_for_offset,
23101 create_function_info_for, create_variable_info_for_1,
23102 create_variable_info_for, intra_create_variable_infos): Adjust.
23103 (init_base_vars): Push NULL for ID zero.
23104 (compute_points_to_sets): Adjust.
23106 2013-03-20 Richard Biener <rguenther@suse.de>
23108 * cfgloop.c (verify_loop_structure): Streamline and avoid
23109 ICEing on corrupt loop tree.
23110 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
23113 2013-03-20 Richard Biener <rguenther@suse.de>
23115 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
23116 check whether an SSA update is needed.
23118 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
23120 * config/mips/constraints.md (T): Rename to...
23124 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
23125 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
23127 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23129 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
23130 (*subsi3_carryin_uxtw): Likewise.
23132 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23134 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
23135 (*rorsi3_insn_uxtw): Likewise.
23137 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23139 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
23140 (*extrsi5_insn_uxtw): Likewise.
23142 2013-03-19 Richard Biener <rguenther@suse.de>
23144 PR tree-optimization/56273
23145 * passes.c (init_optimization_passes): Move second VRP after DOM.
23147 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
23149 * config/i386/i386.md (*movti_internal): Merge from
23150 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
23151 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
23152 nox64 isa attributes.
23154 2013-03-18 Richard Biener <rguenther@suse.de>
23156 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
23158 (merge_node_constraints): Likewise.
23159 (build_succ_graph): Likewise.
23160 (valid_graph_edge): Inline into single caller.
23161 (unify_nodes): Likewise. Use bitmap_set_bit return value
23163 (scc_visit): Fix formatting and variable use.
23164 (do_sd_constraint): Use gcc_checking_assert.
23165 (do_ds_constraint): Likewise.
23166 (do_complex_constraint): Likewise.
23167 (condense_visit): Likewise. Cleanup.
23168 (dump_pred_graph): New function.
23169 (perform_var_substitution): Dump the pred-graph before
23170 variable substitution.
23171 (find_equivalent_node): Use gcc_checking_assert.
23172 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
23174 2013-03-18 Richard Biener <rguenther@suse.de>
23176 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
23177 Remove cond_expr_stmt_list argument and do not gimplify the
23179 (vect_loop_versioning): Adjust.
23180 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
23181 Cleanup to use less temporaries.
23182 (vect_create_data_ref_ptr): Cleanup.
23184 2013-03-18 Jakub Jelinek <jakub@redhat.com>
23186 PR tree-optimization/56635
23187 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
23188 require types_compatible_p types.
23190 2013-03-18 Nick Clifton <nickc@redhat.com>
23192 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
23193 spurious backslash.
23195 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
23196 Add missing line to comment describing function.
23198 2013-03-18 Richard Biener <rguenther@suse.de>
23200 PR tree-optimization/56210
23201 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
23202 Handle string / character search functions.
23203 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
23205 2013-03-18 Richard Biener <rguenther@suse.de>
23207 PR middle-end/56483
23208 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
23209 and implement properly.
23210 * gimple.h (gimple_cond_single_var_p): Remove.
23212 2013-03-18 Richard Biener <rguenther@suse.de>
23214 * tree-data-ref.h (find_data_references_in_loop): Declare.
23215 * tree-data-ref.c (get_references_in_stmt): Use a stack
23216 vector pre-allocated in the callers.
23217 (find_data_references_in_stmt): Adjust.
23218 (graphite_find_data_references_in_stmt): Likewise.
23219 (create_rdg_vertices): Likewise.
23220 (find_data_references_in_loop): Export.
23221 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
23222 Compute dependences here...
23223 (vect_analyze_data_refs): ...not here. When we encounter
23224 a non-vectorizable data reference in basic-block vectorization
23225 truncate the data reference vector. Do not bother to
23226 fixup data-dependence information for gather loads.
23227 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
23228 of data references, as reported.
23230 2013-03-18 Richard Biener <rguenther@suse.de>
23232 PR tree-optimization/3713
23233 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
23234 has_constants and expr.
23235 (stmt_has_constants): Properly valueize SSA names when deciding
23236 whether the stmt has constants.
23238 2013-03-18 Richard Biener <rguenther@suse.de>
23240 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
23241 whole function when there is nothing to do.
23242 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
23243 * tree-vectorizer.c (vectorize_loops): Update virtual and
23244 loop-closed SSA once.
23245 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
23247 2013-03-18 Richard Biener <rguenther@suse.de>
23249 PR middle-end/56113
23250 * domwalk.c (bb_postorder): New global static.
23251 (cmp_bb_postorder): New function.
23252 (walk_dominator_tree): Replace scheme imposing an order for
23253 visiting dominator sons by one sorting them at the time they
23254 are pushed on the stack.
23256 2013-03-18 Richard Biener <rguenther@suse.de>
23258 PR tree-optimization/39326
23259 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
23260 (struct mem_ref): Replace mem member with ao_ref typed member.
23261 (MEM_ANALYZABLE): Adjust.
23262 (memref_eq): Likewise.
23263 (mem_ref_alloc): Likewise.
23264 (gather_mem_refs_stmt): Likewise.
23265 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
23266 (execute_sm_if_changed_flag_set): Adjust.
23267 (execute_sm): Likewise.
23268 (ref_always_accessed_p): Likewise.
23269 (refs_independent_p): Likewise.
23270 (can_sm_ref_p): Likewise.
23272 2013-03-18 Jakub Jelinek <jakub@redhat.com>
23275 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
23276 return 1 even for !unsignedp.
23278 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23280 * config/i386/i386.md (isa): Add x64 and nox64.
23281 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
23282 (*pushtf): Enable *roF alternative for x64 isa only.
23283 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
23284 mode attribute of integer alternatives to DImode for TARGET_64BIT.
23285 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
23286 (*movtf_internal): Merge from *movtf_internal_rex64 and
23287 *movtf_internal_sse. Use x64 and nox64 isa attributes.
23288 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
23289 nox64 isa attributes.
23290 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
23291 nox64 isa attributes.
23292 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
23294 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23296 * config/alpha/alpha.c (TARGET_LRA_P): New define.
23298 2013-03-17 Jakub Jelinek <jakub@redhat.com>
23301 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
23302 class names. Remove trailing comma after "ALL_REGS".
23304 2013-03-16 Jan Hubicka <jh@suse.cz>
23306 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
23307 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
23308 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
23309 of cgraph_get_create_node.
23310 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
23312 2013-03-16 Jason Merrill <jason@redhat.com>
23315 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
23316 with DW_AT_default_value.
23318 2013-03-16 Jakub Jelinek <jakub@redhat.com>
23320 * BASE-VER: Set to 4.9.0.
23322 2013-03-14 Andi Kleen <ak@linux.intel.com>
23325 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
23326 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
23327 Document _x* TSX intrinsics.
23329 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
23330 David Holsgrove <david.holsgrove@xilinx.com>
23332 * configure.ac: Add MicroBlaze TLS support detection.
23333 * configure: Regenerate.
23334 * config/microblaze/microblaze-protos.h
23335 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
23336 symbol_mentioned_p, label_mentioned_p): Add prototypes.
23337 * config/microblaze/microblaze.c (microblaze_address_type): Add
23338 ADDRESS_TLS and tls_reloc address types.
23339 (microblaze_address_info): Add tls_reloc.
23340 (TARGET_HAVE_TLS): Define.
23341 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
23342 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
23343 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
23344 load_tls_operand, microblaze_call_tls_get_addr,
23345 microblaze_legitimize_tls_address): New functions.
23346 (microblaze_classify_unspec): Handle UNSPEC_TLS.
23347 (get_base_reg): Use microblaze_tls_symbol_p.
23348 (microblaze_classify_address): Handle TLS.
23349 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
23350 label_mentioned_p and microblaze_tls_referenced_p.
23351 (microblaze_legitimize_address): Handle TLS.
23352 (microblaze_address_insns): Handle ADDRESS_TLS.
23353 (pic_address_needs_scratch): Handle TLS.
23354 (print_operand_address): Handle TLS.
23355 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
23356 (microblaze_expand_move): Handle TLS.
23357 (microblaze_legitimate_constant_p): Check
23358 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
23359 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
23360 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
23361 (PIC_OFFSET_TABLE_REGNUM): Set.
23362 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
23363 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
23364 (addsi3, movsi_internal2, movdf_internal): Update constraints
23365 * config/microblaze/predicates.md (arith_plus_operand): Define
23366 (move_operand): Redefine as move_src_operand,
23367 check microblaze_tls_referenced_p.
23369 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23371 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
23372 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
23374 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23376 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
23379 2013-03-14 Jakub Jelinek <jakub@redhat.com>
23381 PR tree-optimization/53265
23382 * common.opt (Waggressive-loop-optimizations): New option.
23383 * tree-ssa-loop-niter.c: Include tree-pass.h.
23384 (do_warn_aggressive_loop_optimizations): New function.
23385 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
23386 if number_of_latch_executions returned constant.
23387 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
23388 early. If number_of_latch_executions returned constant, set
23389 nb_iterations_upper_bound back to it.
23390 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
23392 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
23393 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
23395 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
23396 (MULTILIB_OSDIRNAMES): Set.
23397 * genmultilib: If defaultosdirname doesn't start with :: , set
23398 defaultosdirname2 instead, clear it and emit two . multilib_raw
23399 entries instead of just one.
23401 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
23403 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23404 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
23405 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23406 (SUBTARGET_OVERRIDE_OPTIONS): New.
23408 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
23411 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
23412 (musermode): Convert to Var(TARGET_USERMODE).
23413 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
23414 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
23415 * config/sh/sh.c (sh_option_override): Use
23416 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
23417 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
23419 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
23421 (udivsi3_i4_single, divsi3_i4_single): Use
23422 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
23424 2013-03-13 Dave Korn <dave.korn.cygwin@....>
23426 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
23429 2013-03-13 Richard Biener <rguenther@suse.de>
23431 PR tree-optimization/56608
23432 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
23433 calls when vectorizing basic-blocks.
23435 2013-03-13 Jakub Jelinek <jakub@redhat.com>
23438 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
23441 2013-03-12 Jakub Jelinek <jakub@redhat.com>
23443 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
23445 2013-03-11 Jan Hubicka <jh@suse.cz>
23448 * lto-streamer-out.c (output_symbol_p): Skip references from
23449 constructors of external variables.
23451 2013-03-11 Jan Hubicka <jh@suse.cz>
23453 PR middle-end/56571
23454 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
23456 * emit-rtl.c (verify_rtx_sharing): Likewise.
23457 (copy_insn_1): Likewise.
23458 * rtl.c (copy_rtx): Likewise.
23460 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
23463 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
23464 output_operand_lossage message.
23466 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
23469 * arm.c (shift_op): Validate RTL pattern on the fly.
23470 (arm_print_operand, case 'S'): Don't use shift_operator to validate
23473 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23476 * config/pa/pa.md (call_value): Check for calls to powf and direct to
23477 new call patterns that clobber %fr12.
23478 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
23479 split and postreload patterns.
23480 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
23481 registers %fr12 and %fr12R as call used.
23483 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
23485 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
23486 (canon_address, record_store, replace_read, check_mem_read_rtx,
23487 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
23488 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
23489 rest_of_handle_dse): Likewise.
23491 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
23493 PR middle-end/56524
23494 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
23496 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
23497 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
23498 (save_optabs_if_changed): Replace with...
23499 (init_tree_optimization_optabs): ...this.
23500 * optabs.c (save_optabs_if_changed): Rename to...
23501 (init_tree_optimization_optabs): ...this. Take the optimization node
23502 as argument. Do nothing if the base optabs are already correct.
23503 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
23504 to recompute optabs.
23505 * function.h (function): Remove optabs field.
23506 * function.c (invoke_set_current_function_hook): Call
23507 init_tree_optimization_optabs. Use the result to initialize
23510 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
23512 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
23513 if GTMA_HAS_NO_INSTRUMENTATION.
23514 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
23515 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
23516 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
23517 * gimple-pretty-print.c (dump_gimple_transaction): Handle
23518 GTMA_HAS_NO_INSTRUMENTATION.
23520 2013-03-08 Jakub Jelinek <jakub@redhat.com>
23522 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
23525 2013-03-08 Marek Polacek <polacek@redhat.com>
23526 Jakub Jelinek <jakub@redhat.com>
23528 PR tree-optimization/56478
23529 * predict.c (is_comparison_with_loop_invariant_p): Change the
23530 type of loop_step to tree.
23531 (predict_loops): Adjust.
23532 (predict_iv_comparison): Perform the computations on double_ints.
23534 2013-03-08 Richard Biener <rguenther@suse.de>
23536 PR tree-optimization/56570
23537 * tree-cfg.c (verify_expr_location_1): Verify locations for
23539 * tree-sra.c (create_access_replacement): Strip locations
23540 from DECL_DEBUG_EXPRs.
23542 2013-03-08 Richard Biener <rguenther@suse.de>
23544 * tree-inline.c (expand_call_inline): Do not associate
23545 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
23546 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
23548 2013-03-08 Richard Biener <rguenther@suse.de>
23550 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
23551 or block changes with -Og. Fix for location / block encoding
23552 changes and PHI arguments with locations.
23554 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
23556 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
23558 (struct output_info): Likewise.
23559 (register_overhead): Remove bad gcc_assert.
23560 (bitmap_find_bit): If there is only a single bitmap element, do not
23561 count a miss as a search.
23562 (print_statistics): Update for counter type changes.
23563 (dump_bitmap_statistics): Likewise. Print headers such that they
23564 are properly lined up with the printed counters.
23566 2013-03-07 Jakub Jelinek <jakub@redhat.com>
23568 PR tree-optimization/56559
23569 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
23570 check that it has only a single use.
23572 2013-03-07 Richard Biener <rguenther@suse.de>
23574 * doc/invoke.texi (fwhole-program): Discourage use in combination
23577 2013-03-06 Jakub Jelinek <jakub@redhat.com>
23579 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
23581 PR tree-optimization/56539
23582 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
23583 instead of GSI_CONTINUE_LINKING as last argument to
23584 force_gimple_operand_gsi. Adjust function comment.
23586 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
23589 PR middle-end/56548
23590 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
23591 promoted mode, convert the result back to the original mode.
23593 2013-03-06 Richard Biener <rguenther@suse.de>
23595 PR middle-end/56294
23596 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
23597 (insert_updated_phi_nodes_compare_uids): New function.
23598 (update_ssa): Sort symbols_to_rename after UID before
23599 traversing it to insert PHI nodes.
23601 2013-03-06 Richard Biener <rguenther@suse.de>
23603 PR middle-end/50494
23604 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
23605 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
23608 2013-02-13 Richard Biener <rguenther@suse.de>
23611 * varasm.c (output_constant_def_1): Get the decl representing
23612 the constant as argument.
23613 (output_constant_def): Wrap output_constant_def_1.
23614 (make_decl_rtl): Use output_constant_def_1 with the decl
23615 representing the constant.
23616 (build_constant_desc): Optionally re-use a decl already
23617 representing the constant.
23618 (tree_output_constant_def): Adjust.
23620 2013-03-06 Joey Ye <joey.ye@arm.com>
23623 * gcc.c (convert_white_space): New function.
23624 (main): Handles white space in function name.
23626 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
23629 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
23630 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
23631 to SH_DIV_CALL_TABLE for TARGET_SH2.
23632 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
23634 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
23635 call-table options.
23637 2013-03-05 Sterling Augustine <saugustine@google.com>
23638 Cary Coutant <ccoutant@google.com>
23641 * dwarf2out.c (resolve_addr): Don't call
23642 remove_loc_list_addr_table_entries a second time for the same
23645 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23648 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
23649 (avoid_complex_debug_insns): New function.
23650 (expand_debug_locations): Call it.
23652 PR rtl-optimization/56484
23653 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
23654 lifetimes of hard registers on small register class machines.
23656 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23658 * config/microblaze/microblaze-protos.h: Rename
23659 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
23660 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
23662 (microblaze_fast_interrupt_function_p): New function.
23663 (microblaze_is_interrupt_handler): Rename to
23664 microblaze_is_interrupt_variant and add fast_interrupt check.
23665 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
23666 (save_restore_insns): Likewise.
23667 (compute_frame_size): Likewise.
23668 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
23669 (microblaze_globalize_label): Likewise.
23670 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
23671 * config/microblaze/microblaze.md: Use wrapper
23672 microblaze_is_interrupt_variant.
23674 2013-03-05 Kai Tietz <ktietz@redhat.com>
23676 * sdbout.c (sdbout_one_type): Switch to current function's section
23677 supporting cold/hot.
23679 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23681 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
23684 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23686 PR middle-end/56461
23687 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
23688 if VALGRIND_GET_VBITS is defined, temporarily make object
23689 memory all defined, and restore previous valgrind addressability
23690 and definability afterwards. Free this_object at the end.
23692 PR middle-end/56461
23693 * lra.c (lra): Call lra_clear_live_ranges if live_p,
23694 right before calling lra_create_live_ranges, also call it
23695 when clearing live_p. Only call lra_clear_live_ranges
23696 at the end if live_p.
23698 PR middle-end/56461
23699 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
23701 2013-03-05 Richard Biener <rguenther@suse.de>
23703 PR tree-optimization/56521
23704 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
23707 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
23710 * except.h (remove_unreachable_eh_regions): New prototype.
23711 * except.c (remove_eh_handler_splicer): New function, split out
23712 of remove_eh_handler.
23713 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
23714 warning about running it on many EH regions one at a time.
23715 (remove_unreachable_eh_regions_worker): New function, walk the
23716 EH tree in depth-first order and remove non-marked regions.
23717 (remove_unreachable_eh_regions): New function.
23718 * tree-eh.c (mark_reachable_handlers): New function, split out
23719 from remove_unreachable_handlers.
23720 (remove_unreachable_handlers): Use mark_reachable_handlers and
23721 remove_unreachable_eh_regions.
23722 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
23723 and remove_unreachable_eh_regions.
23725 2013-03-05 Richard Biener <rguenther@suse.de>
23727 PR middle-end/56525
23728 * loop-init.c (fix_loop_structure): Remove loops in two stages,
23729 not freeing them until the end.
23731 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23733 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
23735 2013-03-05 Richard Biener <rguenther@suse.de>
23737 PR tree-optimization/56270
23738 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
23739 of loads after scheduling an SLP instance.
23741 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23743 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
23745 (check_gcc_parallelize): Run guality.exp as a separate job from
23746 vect.exp with unsorted.exp and $(dg_target_exps) separately from
23747 struct-layout-1.exp with stackalign.exp.
23749 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
23751 PR middle-end/56461
23752 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
23753 load_index sbitmap even if some bit in it isn't set.
23755 PR middle-end/56461
23756 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
23757 (discover_iteration_bound_by_body_walk): Change queues to
23758 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
23759 spelling in comment. Call safe_push on queues[bound_index] directly.
23760 Release queues[queue_index] in every iteration unconditionally.
23761 Release bounds vector.
23763 PR middle-end/56461
23764 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
23765 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
23766 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
23769 2013-03-05 Richard Biener <rguenther@suse.de>
23772 * tree-inline.c (remap_blocks_to_null): New function.
23773 (expand_call_inline): When expanding a call stmt without
23774 an associated block inline remap all callee blocks to NULL.
23776 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23778 PR rtl-optimization/56494
23779 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
23780 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
23781 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
23783 PR middle-end/56461
23784 * sel-sched-ir.c (free_sched_pools): Release
23785 succs_info_pool.stack[succs_info_pool.max_top] vectors too
23786 if succs_info_pool.max_top isn't -1.
23789 * opts.c (opts_obstack, opts_concat): Moved to...
23790 * opts-common.c (opts_obstack, opts_concat): ... here.
23792 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23794 PR middle-end/56461
23795 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
23797 2013-03-04 Martin Jambor <mjambor@suse.cz>
23799 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
23800 all appropriate places.
23802 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
23804 PR tree-optimization/56424
23805 * ipa-split.c (split_function): Do not set the RSO flag if result is
23806 not by reference and its type is a register type.
23808 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
23810 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
23811 (microblaze_legitimate_pic_operand): Likewise
23812 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
23813 new function microblaze_legitimate_pic_operand
23814 * config/microblaze/microblaze-protos.h
23815 (microblaze_legitimate_pic_operand): Declare.
23817 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23819 * config/microblaze/predicates.md (call_insn_simple_operand):
23820 New predicate for supported rtx code types.
23821 * config/microblaze/microblaze.md (call_internal1): Use
23822 call_insn_simple_operand predicate.
23824 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23826 PR middle-end/56461
23827 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
23828 partitions.ordered_remove.
23830 PR middle-end/56461
23831 * tree-vect-stmts.c (vectorizable_conversion): Don't call
23832 vec_oprnds0.create (1) for modifier == NONE.
23834 PR middle-end/56461
23835 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
23836 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
23837 vec_oprnds1 right before pushing anything to it for
23840 PR middle-end/56461
23841 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
23842 set nbbs to 0 instead of having separate code path.
23843 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
23844 instead of false as last argument if returning NULL.
23846 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
23848 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
23849 the attribute is now called "target" instead of "option".
23850 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
23851 * doc/tm.texi.in (Target Attributes): Likewise document the correct
23852 attribute/pragma name for TARGET_OPTION_VALID_P and
23853 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
23854 * doc/tm.texi: Regenerated.
23856 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
23858 * config/microblaze/microblaze.c:
23859 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
23860 * config/microblaze/microblaze.h: Add -mxl-reorder to
23862 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
23863 instructions emitted if TARGET_REORDER.
23864 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
23865 or 0 for -m/-mno case, but initialises as 2 to detect default use case
23868 2013-03-01 Xinliang David Li <davidxl@google.com>
23870 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
23873 2013-03-01 Jakub Jelinek <jakub@redhat.com>
23875 PR middle-end/56461
23876 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
23877 vector even when returning true. Fix up function comment formatting.
23879 PR middle-end/56461
23880 * ira-build.c (ira_loop_nodes_count): New variable.
23881 (create_loop_tree_nodes): Initialize it.
23882 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
23884 PR middle-end/56461
23885 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
23886 method on dr_chain and result_chain.
23887 * tree-vect-stmts.c (vectorizable_store): Only call
23888 result_chain.create if j == 0.
23890 PR middle-end/56461
23891 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
23892 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
23893 before overwriting it.
23895 2013-03-01 Tobias Burnus <burnus@net-b.de>
23897 * doc/extended.texi (C Extensions): Change order in @menu
23899 (Other MIPS Built-in Functions): Move last MIPS entry before
23900 "picoChip Built-in Functions".
23901 (SH Built-in Functions): Move after RX Built-in Functions.
23902 * doc/gcc.texi (Introduction): Change order in @menu
23904 * doc/md.texi (Constraints): Ditto.
23905 * gty.texi (Type Information): Ditto.
23906 (User-provided marking routines for template types): Make
23908 * doc/invoke.texi (AArch64 Options): Move before
23909 "Adapteva Epiphany Options".
23911 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
23912 Jakub Jelinek <jakub@redhat.com>
23915 * asan.c (gate_asan): Lookup no_sanitize_address instead of
23916 no_address_safety_analysis attribute.
23917 * doc/extend.texi (no_address_safety_attribute): Rename to
23918 no_sanitize_address attribute, mention no_address_safety_analysis
23919 attribute as deprecated alias.
23921 2013-02-28 Jakub Jelinek <jakub@redhat.com>
23923 PR middle-end/56461
23924 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
23925 type to vec<vec<tree> > *.
23926 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
23927 to be vec<tree> instead of vec<tree> *, set vec_defs
23928 to vNULL and call vec_defs.create (number_of_vects), adjust other
23930 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
23931 vectorizable_condition): Adjust vect_get_slp_defs callers.
23933 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23935 * config/aarch64/aarch64.c
23936 (aarch64_float_const_representable): Remove unused variable.
23938 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23940 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
23942 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23944 * config/aarch64/aarch64-builtins.c
23945 (aarch64_init_simd_builtins): Make static.
23947 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23949 * config/aarch64/aarch64.c
23950 (aarch64_simd_make_constant): Make static.
23952 2013-02-28 Martin Jambor <mjambor@suse.cz>
23954 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
23955 with no initialization to the RHS of debug statements.
23957 2013-02-28 Martin Jambor <mjambor@suse.cz>
23959 PR tree-optimization/56294
23960 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
23962 (get_access_replacement): Do not call create_access_replacement.
23963 Assert a replacement exists.
23964 (get_repl_default_def_ssa_name): Create the replacement declaration
23967 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23969 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
23970 final_end_function.
23972 2013-02-28 Marek Polacek <polacek@redhat.com>
23974 PR rtl-optimization/56466
23975 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
23976 if we're changing a loop.
23977 (peel_loops_completely): Likewise.
23979 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
23982 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
23984 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23987 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
23988 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
23989 INTX_FTYPE_FX, FX_FTYPE_INTX.
23990 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
23992 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23994 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
23995 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
23996 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
23997 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
23998 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
23999 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
24000 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
24001 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
24002 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
24003 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
24004 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
24005 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
24006 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
24007 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
24008 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
24009 (avrxmega6): Increase max flash segments from 5 to 6.
24010 * config/avr/t-multilib: Regenerate.
24011 * config/avr/avr-tables.opt: Regenerate.
24012 * doc/avr-mmcu.texi: Regenerate.
24014 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
24016 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
24017 (avr_device_to_arch): Rename to avr_device_to_ld.
24018 (avr_device_to_as): New prototype.
24019 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
24020 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
24021 * config/avr/driver-avr.c (avr_device_to_as): New.
24022 (avr_device_to_arch): Rename to avr_device_to_ld.
24024 2013-02-27 Jakub Jelinek <jakub@redhat.com>
24026 PR middle-end/56461
24027 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
24028 method on dr_chain and result_chain.
24030 PR middle-end/56461
24031 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
24032 pointer_set_destroy on not_executed_last_iteration.
24034 PR middle-end/56461
24035 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
24037 PR middle-end/56461
24038 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
24039 FOR_EACH_DEFINED_FUNCTION when freeing state.
24041 PR middle-end/56461
24042 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
24044 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
24047 PR middle-end/56461
24048 * ipa-cp.c (decide_whether_version_node): Call vec_free on
24049 known_aggs[i].items and release known_aggs vector.
24051 PR middle-end/56461
24052 * ipa-reference.c (propagate): Free node_info even for alias nodes.
24054 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24056 * config/microblaze/microblaze.c (microblaze_emit_compare):
24057 Use xor for EQ/NE comparisions.
24058 * config/microblaze/microblaze.md (cstoresf4): Add constraints
24059 (cbranchsf4): Adjust operator to comparison_operator.
24061 2013-02-27 Jakub Jelinek <jakub@redhat.com>
24063 PR middle-end/56461
24064 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
24066 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
24067 vec_safe_push, always update *slot.
24068 (redirect_edge_var_map_clear): Use vec_free.
24069 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
24070 (free_var_map_entry): Use vec_free.
24071 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
24072 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
24074 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
24076 PR middle-end/45472
24077 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
24078 when the may_trap_p bit of the exprs being merged differs.
24079 Reorder tests for speculativeness in the logical and operator.
24081 2013-02-27 Jakub Jelinek <jakub@redhat.com>
24083 * incpath.c (add_standard_paths): Use reconcat instead of concat
24084 where appropriate and avoid leaking memory.
24086 * opts.h: Include obstack.h.
24087 (opts_concat): New prototype.
24088 (opts_obstack): New declaration.
24089 * opts.c (opts_concat): New function.
24090 (opts_obstack): New variable.
24091 (init_options_struct): Call gcc_init_obstack on opts_obstack.
24092 (finish_options): Use opts_concat instead of concat
24093 and XOBNEWVEC instead of XNEWVEC.
24094 * opts-common.c (generate_canonical_option, decode_cmdline_option,
24095 generate_option): Likewise.
24096 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
24097 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
24100 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
24101 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
24103 2013-02-26 Jakub Jelinek <jakub@redhat.com>
24105 PR middle-end/56461
24106 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
24108 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
24110 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
24111 (arm_block_move_unaligned_straight): Likewise.
24112 (arm_adjust_block_mem): Likewise.
24114 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
24117 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
24118 temp, cond and label.
24119 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
24122 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
24123 * config/c6x/c6x.h (dbx_register_map): Update declaration.
24126 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
24127 of prologue/epilogue functions.
24130 * config/tilegx/tilegx.c (tilegx_expand_prologue):
24131 Remove unused variable cfa_offset.
24132 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
24135 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
24136 type promotion to unsigned.
24139 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
24140 for HOST_WIDE_INT of 32 bit / same size as int.
24141 (arm_block_move_unaligned_straight): Likewise.
24142 (arm_adjust_block_mem): Likewise.
24145 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
24148 2013-02-26 Marek Polacek <polacek@redhat.com>
24150 PR tree-optimization/56426
24151 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
24153 2013-02-26 Richard Biener <rguenther@suse.de>
24156 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
24157 unused variable loops.
24159 2013-02-26 Jakub Jelinek <jakub@redhat.com>
24161 PR tree-optimization/56448
24162 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
24163 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
24164 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
24165 later operands of the references, or even first operand for
24166 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
24168 PR tree-optimization/56443
24169 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
24170 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
24171 to type_for_mode langhook.
24173 2013-02-25 Matt Turner <mattst88@gmail.com>
24175 * doc/invoke.texi: Document r4700.
24177 2013-02-25 Richard Biener <rguenther@suse.de>
24179 PR tree-optimization/56175
24180 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
24182 (simplify_bitwise_binary): ... here. Also guard the conversion
24183 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
24185 2013-02-25 Catherine Moore <clm@codesourcery.com>
24188 2013-02-24 Catherine Moore <clm@codesourcery.com>
24189 Maciej W. Rozycki <macro@codesourcery.com>
24190 Tom de Vries <tom@codesourcery.com>
24191 Nathan Sidwell <nathan@codesourcery.com>
24192 Iain Sandoe <iain@codesourcery.com>
24193 Nathan Froyd <froydnj@codesourcery.com>
24194 Chao-ying Fu <fu@mips.com>
24196 * doc/extend.texi: (micromips, nomicromips, nocompression):
24197 Document new function attributes.
24198 * doc/invoke.texi (minterlink-compressed, mmicromips,
24199 m14k, m14ke, m14kec): Document new options.
24200 (minterlink-mips16): Update documentation.
24201 * doc/md.texi (ZC, ZD): Document new constraints.
24202 * configure.ac (gcc_cv_as_micromips): Check if linker
24203 supports the .set micromips directive.
24204 * configure: Regenerate.
24205 * config.in: Regenerate.
24206 * config/mips/mips-tables.opt: Regenerate.
24207 * config/mips/micromips.md: New file.
24208 * constraints.md (ZC, AD): New constraints.
24209 * config/mips/predicates.md (movep_src_register): New predicate.
24210 (movep_src_operand): New predicate.
24211 (non_volatile_mem_operand): New predicate.
24212 * config/mips/mips.md (multimem): New type.
24213 (length): Differentiate between 17-bit and 18-bit branch offsets.
24214 (MOVEP1, MOVEP2): New mode iterator.
24215 (mov_<load>l): Use ZC constraint.
24216 (mov_<load>r): Likewise.
24217 (mov_<store>l): Likewise.
24218 (mov_<store>r): Likewise.
24219 (*branch_equality<mode>_inverted): Add microMIPS support.
24220 (*branch_equality<mode>): Likewise.
24221 (*jump_absolute): Likewise.
24222 (indirect_jump_<mode>): Likewise.
24223 (tablejump_<mode>): Likewise.
24224 (<optab>_internal): Likewise.
24225 (sibcall_internal): Likewise.
24226 (sibcall_value_internal): Likewise.
24227 (prefetch): Use constraint ZD.
24228 * config/mips/mips.opt (minterlink-compressed): New option.
24229 (minterlink-mips16): Now an alias for minterlink-compressed.
24230 (mmicromips): New option.
24231 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24232 (compare_and_swap_12): Likewise.
24233 (sync_add<mode>): Likewise.
24234 (sync_<optab>_12): Likewise.
24235 (sync_old_<optab>_12): Likewise.
24236 (sync_new_<optab>_12): Likewise.
24237 (sync_nand_12): Likewise.
24238 (sync_old_nand_12): Likewise.
24239 (sync_new_nand_12): Likewise.
24240 (sync_sub<mode>): Likewise.
24241 (sync_old_add<mode>): Likewise.
24242 (sync_old_sub<mode>): Likewise.
24243 (sync_new_add<mode>): Likewise.
24244 (sync_new_sub<mode>): Likewise.
24245 (sync_<optab><mode>): Likewise.
24246 (sync_old_<optab><mode>): Likewise.
24247 (sync_new_<optab><mode>): Likewise.
24248 (sync_nand<mode>): Likewise.
24249 (sync_old_nand<mode>): Likewise.
24250 (sync_new_nand<mode>): Likewise.
24251 (sync_lock_test_and_set<mode>): Likewise.
24252 (test_and_set_12): Likewise.
24253 (atomic_compare_and_swap<mode>): Likewise.
24254 (atomic_exchange<mode>_llsc): Likewise.
24255 (atomic_fetch_add<mode>_llsc): Likewise.
24256 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24257 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24258 (umips_save_restore_pattern_p): Likewise.
24259 (umips_load_store_pair_p): Likewise.
24260 (umips_output_load_store_pair): Likewise.
24261 (umips_movep_target_p): Likewise.
24262 (umips_12bit_offset_address_p): Likewise.
24263 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24264 (mips_base_mips16): Rename this...
24265 (mips_base_compression_flags): ...to this. Update all uses.
24266 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24267 (mips_mips16_decl_p): Delete.
24268 (mips_nomips16_decl_p): Delete.
24269 (mips_get_compress_on_flags): New function.
24270 (mips_get_compress_off_flags): New function.
24271 (mips_get_compress_mode): New function.
24272 (mips_get_compress_on_name): New function.
24273 (mips_get_compress_off_name): New function.
24274 (mips_insert_attributes): Support multiple compression types.
24275 (mips_merge_decl_attributes): Likewise.
24276 (umips_12bit_offset_address_p): New function.
24277 (mips_start_function_definition): Emit .set micromips directive.
24278 (mips_call_may_need_jalx_p): New function.
24279 (mips_function_ok_for_sibcall): Add microMIPS support.
24280 (mips_print_operand_punctuation): Support short delay slots and
24282 (umips_swm_mask, umips_swm_encoding): New.
24283 (umips_build_save_restore): New function.
24284 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24285 (was_mips16_p): Remove.
24286 (old_compression_mode): New.
24287 (mips_set_compression_mode): New function.
24288 (mips_set_current_function): Add microMIPS support.
24289 (mips_option_override): Likewise.
24290 (umips_save_restore_pattern_p): New function.
24291 (umips_output_save_restore): New function.
24292 (umips_load_store_pair_p_1): New function.
24293 (umips_load_store_pair_p): New function.
24294 (umips_output_load_store_pair_1): New function.
24295 (umips_output_load_store_pair): New function.
24296 (umips_movep_target_p) New function.
24297 (mips_prepare_pch_save): Add microMIPS support.
24298 * config/mips/mips.h (TARGET_COMPRESSION): New.
24299 (TARGET_CPU_CPP_BUILTINS): Update macro
24300 to use new compression flags and to support microMIPS.
24301 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24302 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24303 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24304 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24305 (ASM_SPEC): Support mmicromips and mno-micromips.
24306 (M16STORE_REG_P): New macro.
24307 (MIPS_CALL): Support TARGET_MICROMIPS.
24308 (MICROMIPS_J): New macro.
24309 (mips_base_mips16): Rename this...
24310 (mips_base_compression_flags): ...to this.
24311 (UMIPS_12BIT_OFFSET_P): New macro.
24312 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24313 (MULTILIB_DIRNAMES): Likewise.
24315 2013-02-25 Tom de Vries <tom@codesourcery.com>
24317 PR rtl-optimization/56131
24318 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
24319 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
24320 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
24322 2013-02-25 Tobias Burnus <burnus@net-b.de>
24324 * doc/invoke.texi (-fsanitize=): Move from optimization
24325 to debugging options.
24327 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
24329 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
24331 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
24332 Alexander Monakov <amonakov@ispras.ru>
24334 PR middle-end/56077
24335 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
24336 flush pending lists also on non-jumps. Adjust comment.
24338 2013-02-24 Catherine Moore <clm@codesourcery.com>
24339 Maciej W. Rozycki <macro@codesourcery.com>
24340 Tom de Vries <tom@codesourcery.com>
24341 Nathan Sidwell <nathan@codesourcery.com>
24342 Iain Sandoe <iain@codesourcery.com>
24343 Nathan Froyd <froydnj@codesourcery.com>
24344 Chao-ying Fu <fu@mips.com>
24346 * doc/extend.texi: (micromips, nomicromips, nocompression):
24347 Document new function attributes.
24348 * doc/invoke.texi (minterlink-compressed, mmicromips,
24349 m14k, m14ke, m14kec): Document new options.
24350 (minterlink-mips16): Update documentation.
24351 * doc/md.texi (ZC, ZD): Document new constraints.
24352 * configure.ac (gcc_cv_as_micromips): Check if linker
24353 supports the .set micromips directive.
24354 * configure: Regenerate.
24355 * config.in: Regenerate.
24356 * config/mips/mips-tables.opt: Regenerate.
24357 * config/mips/micromips.md: New file.
24358 * constraints.md (ZC, AD): New constraints.
24359 * config/mips/predicates.md (movep_src_register): New predicate.
24360 (movep_src_operand): New predicate.
24361 (non_volatile_mem_operand): New predicate.
24362 * config/mips/mips.md (multimem): New type.
24363 (length): Differentiate between 17-bit and 18-bit branch offsets.
24364 (MOVEP1, MOVEP2): New mode iterator.
24365 (mov_<load>l): Use ZC constraint.
24366 (mov_<load>r): Likewise.
24367 (mov_<store>l): Likewise.
24368 (mov_<store>r): Likewise.
24369 (*branch_equality<mode>_inverted): Add microMIPS support.
24370 (*branch_equality<mode>): Likewise.
24371 (*jump_absolute): Likewise.
24372 (indirect_jump_<mode>): Likewise.
24373 (tablejump_<mode>): Likewise.
24374 (<optab>_internal): Likewise.
24375 (sibcall_internal): Likewise.
24376 (sibcall_value_internal): Likewise.
24377 (prefetch): Use constraint ZD.
24378 * config/mips/mips.opt (minterlink-compressed): New option.
24379 (minterlink-mips16): Now an alias for minterlink-compressed.
24380 (mmicromips): New option.
24381 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24382 (compare_and_swap_12): Likewise.
24383 (sync_add<mode>): Likewise.
24384 (sync_<optab>_12): Likewise.
24385 (sync_old_<optab>_12): Likewise.
24386 (sync_new_<optab>_12): Likewise.
24387 (sync_nand_12): Likewise.
24388 (sync_old_nand_12): Likewise.
24389 (sync_new_nand_12): Likewise.
24390 (sync_sub<mode>): Likewise.
24391 (sync_old_add<mode>): Likewise.
24392 (sync_old_sub<mode>): Likewise.
24393 (sync_new_add<mode>): Likewise.
24394 (sync_new_sub<mode>): Likewise.
24395 (sync_<optab><mode>): Likewise.
24396 (sync_old_<optab><mode>): Likewise.
24397 (sync_new_<optab><mode>): Likewise.
24398 (sync_nand<mode>): Likewise.
24399 (sync_old_nand<mode>): Likewise.
24400 (sync_new_nand<mode>): Likewise.
24401 (sync_lock_test_and_set<mode>): Likewise.
24402 (test_and_set_12): Likewise.
24403 (atomic_compare_and_swap<mode>): Likewise.
24404 (atomic_exchange<mode>_llsc): Likewise.
24405 (atomic_fetch_add<mode>_llsc): Likewise.
24406 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24407 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24408 (umips_save_restore_pattern_p): Likewise.
24409 (umips_load_store_pair_p): Likewise.
24410 (umips_output_load_store_pair): Likewise.
24411 (umips_movep_target_p): Likewise.
24412 (umips_12bit_offset_address_p): Likewise.
24413 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24414 (mips_base_mips16): Rename this...
24415 (mips_base_compression_flags): ...to this. Update all uses.
24416 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24417 (mips_mips16_decl_p): Delete.
24418 (mips_nomips16_decl_p): Delete.
24419 (mips_get_compress_on_flags): New function.
24420 (mips_get_compress_off_flags): New function.
24421 (mips_get_compress_mode): New function.
24422 (mips_get_compress_on_name): New function.
24423 (mips_get_compress_off_name): New function.
24424 (mips_insert_attributes): Support multiple compression types.
24425 (mips_merge_decl_attributes): Likewise.
24426 (umips_12bit_offset_address_p): New function.
24427 (mips_start_function_definition): Emit .set micromips directive.
24428 (mips_call_may_need_jalx_p): New function.
24429 (mips_function_ok_for_sibcall): Add microMIPS support.
24430 (mips_print_operand_punctuation): Support short delay slots and
24432 (umips_swm_mask, umips_swm_encoding): New.
24433 (umips_build_save_restore): New function.
24434 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24435 (was_mips16_p): Remove.
24436 (old_compression_mode): New.
24437 (mips_set_compression_mode): New function.
24438 (mips_set_current_function): Add microMIPS support.
24439 (mips_option_override): Likewise.
24440 (umips_save_restore_pattern_p): New function.
24441 (umips_output_save_restore): New function.
24442 (umips_load_store_pair_p_1): New function.
24443 (umips_load_store_pair_p): New function.
24444 (umips_output_load_store_pair_1): New function.
24445 (umips_output_load_store_pair): New function.
24446 (umips_movep_target_p) New function.
24447 (mips_prepare_pch_save): Add microMIPS support.
24448 * config/mips/mips.h (TARGET_COMPRESSION): New.
24449 (TARGET_CPU_CPP_BUILTINS): Update macro
24450 to use new compression flags and to support microMIPS.
24451 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24452 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24453 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24454 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24455 (ASM_SPEC): Support mmicromips and mno-micromips.
24456 (M16STORE_REG_P): New macro.
24457 (MIPS_CALL): Support TARGET_MICROMIPS.
24458 (MICROMIPS_J): New macro.
24459 (mips_base_mips16): Rename this...
24460 (mips_base_compression_flags): ...to this.
24461 (UMIPS_12BIT_OFFSET_P): New macro.
24462 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24463 (MULTILIB_DIRNAMES): Likewise.
24465 2013-02-24 Jakub Jelinek <jakub@redhat.com>
24468 * target-globals.c (save_target_globals): For init_reg_sets and
24469 target_reinit remporarily set this_fn_optabs to this_target_optabs.
24471 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
24473 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
24474 * config/aarch64/t-aarch64
24475 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
24477 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
24479 PR inline-asm/56148
24480 * lra-constraints.c (process_alt_operands): Reload operand
24481 conflicting with earlier clobber only if no more other conflicting
24484 2013-02-22 Jakub Jelinek <jakub@redhat.com>
24487 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
24488 if not linking a shared library.
24490 2013-02-22 Seth LaForge <sethml@google.com>
24492 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
24494 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
24496 * config/arm/arm.md (split for extendsidi): Update condition.
24497 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
24498 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
24499 (qhs_zextenddi_cstr): Likewise.
24501 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24503 PR middle-end/56420
24504 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
24505 avoid signed wrapping.
24506 (expand_mult): Handle properly multiplication by
24507 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
24508 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
24509 in the compiler if coeff is HOST_WIDE_INT_MIN.
24510 (expand_divmod): Don't make ext_op1 static, change it's type to
24511 uhwi. Avoid undefined behavior in -INTVAL (op1).
24513 PR rtl-optimization/50339
24514 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
24516 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
24517 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
24518 into splitting_ashiftrt field.
24519 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
24521 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
24524 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
24526 PR middle-end/56108
24527 * trans-mem.c (execute_tm_mark): Do not expand transactions that
24528 are sure to go irrevocable.
24530 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
24532 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
24533 scalars are valid operands.
24535 2013-02-21 Martin Jambor <mjambor@suse.cz>
24537 PR tree-optimization/56310
24538 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
24539 only matching indices and non-negative final offsets.
24540 (intersect_aggregates_with_edge): Pass src_idx to
24541 agg_replacements_to_vector. Pass src_idx insstead of index to
24542 intersect_with_agg_replacements.
24544 2013-02-21 Martin Jambor <mjambor@suse.cz>
24546 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
24547 instead of hard-wired defaults.
24549 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
24551 * doc/invoke.texi (MIPS Options): Update documentation of the
24552 floating-point multiply-accumulate instruction restrictions.
24554 2013-02-21 Kostya Serebryany <kcc@google.com>
24556 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
24557 asan_shadow_offset on x86_64 linux.
24559 2013-02-21 Richard Biener <rguenther@suse.de>
24561 PR tree-optimization/56415
24563 2013-02-11 Richard Biener <rguenther@suse.de>
24565 PR tree-optimization/56273
24566 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24569 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24572 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
24575 PR inline-asm/56405
24576 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
24577 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
24579 2013-02-20 Jan Hubicka <jh@suse.cz>
24581 PR tree-optimization/56265
24582 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
24583 when target is referenced for first time.
24585 2013-02-20 Richard Biener <rguenther@suse.de>
24587 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
24588 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
24589 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
24590 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
24591 not return anything.
24592 (rename_ssa_copies): Do not remove unused locals.
24593 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
24594 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
24595 * passes.c (execute_function_todo): Do not schedule unused locals
24596 removal if cleanup_tree_cfg did something.
24597 * tree-ssa-live.c (remove_unused_locals): Dump statistics
24598 about the number of removed locals.
24600 2013-02-20 Richard Biener <rguenther@suse.de>
24602 PR tree-optimization/56398
24603 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
24605 2013-02-20 Martin Jambor <mjambor@suse.cz>
24607 PR tree-optimization/55334
24608 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
24609 restricted pointers to arrays.
24611 2013-02-20 Richard Biener <rguenther@suse.de>
24612 Jakub Jelinek <jakub@redhat.com>
24614 PR tree-optimization/56396
24615 * tree-ssa-ccp.c (n_const_val): New static variable.
24616 (get_value): Return NULL for SSA names we don't have a lattice
24618 (ccp_initialize): Initialize n_const_val.
24619 * tree-ssa-copy.c (n_copy_of): New static variable.
24620 (init_copy_prop): Initialize n_copy_of.
24621 (get_value): Return NULL_TREE for SSA names we don't have a
24624 2013-02-20 Martin Jambor <mjambor@suse.cz>
24626 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
24628 2013-02-20 Richard Biener <rguenther@suse.de>
24630 * genpreds.c (write_lookup_constraint): Do not compare first
24631 letter of the constraint again.
24633 2013-02-20 Richard Biener <rguenther@suse.de>
24635 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
24637 (get_use_iv_cost): Terminate hashtable walk when coming across
24640 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
24642 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
24643 reassociation for avx2 targets.
24645 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24647 * config/microblaze/microblaze.c: microblaze_has_clz = 0
24648 Add version check for v8.10.a to enable microblaze_has_clz
24649 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
24650 version and TARGET_PATTERN_COMPARE check
24651 * config/microblaze/microblaze.md: New clzsi2 instruction
24653 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24655 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
24656 function before branching.
24658 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24660 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
24662 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
24664 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24666 PR middle-end/55889
24667 * sel-sched.c: Include ira.h.
24668 (implicit_clobber_conflict_p): New function.
24669 (moveup_expr): Use it.
24670 * Makefile.in (sel-sched.o): Depend on ira.h.
24672 2013-02-19 Richard Biener <rguenther@suse.de>
24674 PR tree-optimization/56384
24675 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
24676 (vn_hash_type): Split out from ...
24677 (vn_hash_constant_with_type): ... here.
24678 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
24679 (vn_phi_eq): Compare types from vn_phi_s structure.
24680 (vn_phi_lookup): Populate vn_phi_s type.
24681 (vn_phi_insert): Likewise.
24683 2013-02-19 Jakub Jelinek <jakub@redhat.com>
24685 PR tree-optimization/56350
24686 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
24687 if haven't found reduction or nested cycle operand, rather than
24688 asserting we must find it.
24690 PR tree-optimization/56381
24691 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
24694 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
24695 Jakub Jelinek <jakub@redhat.com>
24698 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
24699 (swap_optab_enable): Same.
24700 (init_all_optabs): Use argument instead of global.
24701 * tree.h (struct tree_optimization_option): New field target_optabs.
24702 * expr.h (init_all_optabs): Add argument to prototype.
24703 (TREE_OPTIMIZATION_OPTABS): New.
24704 (save_optabs_if_changed): Protoize.
24705 * optabs.h: Declare this_fn_optabs.
24706 * optabs.c (save_optabs_if_changed): New.
24707 Declare this_fn_optabs.
24708 (init_optabs): Add argument to init_all_optabs() call.
24709 * function.c (invoke_set_current_function_hook): Handle per
24711 * function.h (struct function): New field optabs.
24712 * config/mips/mips.c (mips_set_mips16_mode): Handle when
24713 optimization_current_node has changed.
24714 * target-globals.h (save_target_globals_default_opts): Protoize.
24715 * target-globals.c (save_target_globals_default_opts): New.
24717 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24720 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
24721 registers %fr12 and %fr12R as call used.
24724 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
24725 and HImode, require all displacements to be an integer multiple of
24727 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
24728 only allow QImode and HImode when reload is in progress and strict is
24729 true. Likewise for symbolic addresses. Use base14_operand to check
24730 displacements in REG+BASE addresses.
24732 2013-02-18 Richard Biener <rguenther@suse.de>
24734 PR tree-optimization/56366
24735 * tree-vect-loop.c (get_initial_def_for_induction): Properly
24736 handle sign-conversion of outer-loop initial induction value.
24738 2013-02-18 Richard Biener <rguenther@suse.de>
24740 PR middle-end/56349
24741 * cfghooks.c (merge_blocks): If we merge a latch into another
24742 block adjust references to it.
24743 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
24744 (verify_loop_structure): Verify that a recorded latch is in fact
24747 2013-02-18 Richard Biener <rguenther@suse.de>
24749 PR tree-optimization/56321
24750 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
24751 order SSA name release and virtual operand unlinking.
24753 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24755 * config/microblaze/microblaze.md (save_stack_block): Define.
24756 (restore_stack_block): Likewise.
24758 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24760 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
24761 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
24762 * config/microblaze/microblaze.c (microblaze_option_override):
24763 Bail out early for PIC modes when target does not support PIC.
24765 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24767 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
24768 Replace with a microblaze version.
24769 (microblaze_trampoline_init): Adapt for microblaze.
24770 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
24773 2013-02-16 Jakub Jelinek <jakub@redhat.com>
24774 Dodji Seketeli <dodji@redhat.com>
24777 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
24778 (instrument_mem_region_access): Do not forget to always put
24779 instrumentation of the of 'base' and 'base + len' in a "if (len !=
24780 0) statement, even for cases where either 'base' or 'base + len'
24781 are not instrumented -- because they have been previously
24782 instrumented. Simplify the logic by putting all the statements
24783 instrument 'base + len' inside a sequence, and then insert that
24784 sequence right before the current insertion point. Then, to
24785 instrument 'base + len', just get an iterator on that statement.
24786 And do not forget to update the pointer to iterator the function
24787 received as argument.
24789 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
24791 PR rtl-optimization/56348
24792 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
24794 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
24796 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
24797 (clean_graph_dump_file): Pass base to start_graph_dump.
24799 2013-02-14 Richard Henderson <rth@redhat.com>
24802 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
24804 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
24806 * collect2-aix.h: Define F_LOADONLY.
24808 2013-02-14 Richard Biener <rguenther@suse.de>
24811 * varasm.c (output_constant_def_1): Get the decl representing
24812 the constant as argument.
24813 (output_constant_def): Wrap output_constant_def_1.
24814 (make_decl_rtl): Use output_constant_def_1 with the decl
24815 representing the constant.
24816 (build_constant_desc): Optionally re-use a decl already
24817 representing the constant.
24818 (tree_output_constant_def): Adjust.
24820 2013-02-14 Dodji Seketeli <dodji@redhat.com>
24823 * asan.c (instrument_builtin_call): Really put the length of the
24824 second source argument into src1_len.
24826 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24828 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
24829 argument. If it is false, don't create edge from then_bb to
24831 (insert_if_then_before_iter): Pass true to it.
24832 (build_check_stmt): Pass false to it.
24833 (transform_statements): Flush hash table only on extended basic
24834 block boundaries, rather than at the beginning of every bb.
24835 Don't flush hash table on nonfreeing_call_p calls.
24836 * tree-flow.h (nonfreeing_call_p): New prototype.
24837 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
24839 2013-02-13 David S. Miller <davem@davemloft.net>
24841 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
24843 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
24846 * ira.c (max_regno_before_ira): Move from ...
24848 (fix_reg_equiv_init): Use max_regno_before_ira instead of
24851 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24853 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
24855 2013-02-13 Richard Biener <rguenther@suse.de>
24858 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
24859 globals in MEM_REFs.
24861 2013-02-13 Richard Biener <rguenther@suse.de>
24863 * loop-init.c (loop_optimizer_init): Clear loop state when
24864 re-initializing preserved loops.
24865 * loop-unswitch.c (unswitch_single_loop): Return whether
24866 we unswitched the loop. Do not verify loop state here.
24867 (unswitch_loops): When we unswitched a loop discover new loops.
24869 2013-02-13 Kostya Serebryany <kcc@google.com>
24871 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
24873 * sanitizer.def: Rename __asan_init to __asan_init_v1.
24875 2013-02-12 Dodji Seketeli <dodji@redhat.com>
24877 Avoid instrumenting duplicated memory access in the same basic block
24878 * Makefile.in (asan.o): Add new dependency on hash-table.h
24879 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
24880 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
24881 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
24882 (free_mem_ref_resources, has_mem_ref_been_instrumented)
24883 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
24884 (get_mem_ref_of_assignment): New functions.
24885 (get_mem_refs_of_builtin_call): Extract from
24886 instrument_builtin_call and tweak a little bit to make it fit with
24888 (instrument_builtin_call): Use the new
24889 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
24890 of is_gimple_builtin_call.
24891 (instrument_derefs, instrument_mem_region_access): Insert the
24892 instrumented memory reference into the hash table.
24893 (maybe_instrument_assignment): Renamed instrument_assignment into
24894 this, and change it to advance the iterator when instrumentation
24895 actually happened and return true in that case. This makes it
24896 homogeneous with maybe_instrument_assignment, and thus give a
24897 chance to callers to be more 'regular'.
24898 (transform_statements): Clear the memory reference hash table
24899 whenever we enter a new BB, when we cross a function call, or when
24900 we are done transforming statements. Use
24901 maybe_instrument_assignment instead of instrumentation. No more
24902 need to special case maybe_instrument_assignment and advance the
24903 iterator after calling it; it's now handled just like
24904 maybe_instrument_call. Update comment.
24906 2013-02-13 Richard Biener <rguenther@suse.de>
24908 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
24909 Fix loop discovery code.
24911 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
24913 PR inline-asm/56148
24914 * lra-constraints.c (process_alt_operands): Match early clobber
24915 operand with itself. Check conflicts with earlyclobber only if
24916 the operand is not reloaded. Prefer to reload conflicting operand
24917 if earlyclobber and matching operands are the same.
24919 2013-02-12 Richard Biener <rguenther@suse.de>
24922 * lto-streamer-out.c (write_symbol): Do not output symbols
24923 for hard register variables.
24925 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
24928 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
24929 (umulsidi3_insn, mulsidi3_insn): New insns.
24931 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
24933 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
24934 (struct tune_params): Add vec_costs field.
24935 * config/arm/arm.c (arm_builtin_vectorization_cost)
24936 (arm_add_stmt_cost): New functions.
24937 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
24938 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
24939 (arm_default_vec_cost): New struct of type cpu_vec_costs.
24940 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
24941 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
24942 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
24943 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
24945 2013-02-12 Richard Biener <rguenther@suse.de>
24948 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
24949 decls again if possible.
24951 2013-02-12 Richard Biener <rguenther@suse.de>
24953 PR middle-end/56288
24954 * tree-ssa.c (verify_ssa_name): Fix check, move
24955 SSA_NAME_IN_FREE_LIST check up.
24957 2013-02-12 Jakub Jelinek <jakub@redhat.com>
24958 Steven Bosscher <steven@gcc.gnu.org>
24960 PR rtl-optimization/56151
24961 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
24962 equal to op0 or op1, and last_insn pattern is CODE operation
24963 with MEM dest and one of the operands matches that MEM.
24965 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
24967 * doc/extend.texi: Document Function Multiversioning and "default"
24968 parameter string to target attribute.
24969 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
24970 target attribute parameter is "default".
24971 (ix86_compare_version_priority): Remove checks for target attribute.
24972 (ix86_mangle_function_version_assembler_name): Change error to sorry.
24973 Remove check for target attribute equal to NULL. Add assert.
24974 (ix86_generate_version_dispatcher_body): Change error to sorry.
24976 2013-02-11 Iain Sandoe <iain@codesourcery.com>
24977 Jack Howarth <howarth@bromo.med.uc.edu>
24978 Patrick Marlier <patrick.marlier@gmail.com>
24981 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
24982 define ENDFILE_SPEC as TM_DESTRUCTOR.
24983 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
24985 2013-02-11 Alexander Potapenko <glider@google.com>
24986 Jack Howarth <howarth@bromo.med.uc.edu>
24987 Jakub Jelinek <jakub@redhat.com>
24990 * config/darwin.c (cdtor_record): Rename ctor_record.
24991 (sort_cdtor_records): Rename sort_ctor_records.
24992 (finalize_dtors): New routine to sort destructors by
24993 priority before use in assemble_integer.
24994 (machopic_asm_out_destructor): Use finalize_dtors if needed.
24996 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
24998 PR rtl-optimization/56275
24999 * simplify-rtx.c (avoid_constant_pool_reference): Check that
25000 offset is non-negative and less than cmode size before
25001 calling simplify_subreg.
25003 2013-02-11 Richard Biener <rguenther@suse.de>
25005 PR tree-optimization/56264
25006 * cfgloop.h (fix_loop_structure): Adjust prototype.
25007 * loop-init.c (fix_loop_structure): Return the number of
25008 newly discovered loops.
25009 * tree-cfgcleanup.c (repair_loop_structures): When new loops
25010 are discovered, do a full loop-closed SSA rewrite.
25012 2013-02-11 Richard Biener <rguenther@suse.de>
25014 PR tree-optimization/56273
25015 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
25017 (check_array_ref): Fix missing newline in dumps.
25018 (search_for_addr_array): Likewise.
25020 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
25022 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
25024 2013-02-09 Jakub Jelinek <jakub@redhat.com>
25027 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
25029 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
25031 PR rtl-optimization/56246
25032 * lra-constraints.c (simplify_operand_subreg): Try to reuse
25034 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
25035 constraints are satisfied.
25037 2013-02-08 Jeff Law <law@redhat.com>
25040 * emit-rtl.c (reg_is_parm_p): New function.
25041 * regs.h (reg_is_parm_p): New prototype.
25042 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
25043 callee-clobbered registers.
25045 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
25048 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
25049 If there is no implicit builtin declaration, just return NULL.
25051 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
25053 * config/i386/sse.md (FMAMODEM): New mode iterator.
25054 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
25055 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
25057 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
25059 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
25060 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
25061 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
25063 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25065 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
25066 (microblaze*-*-elf): Likewise.
25067 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
25069 * config/microblaze/microblaze-c.c: Add builtin defines for
25070 _LITTLE_ENDIAN and _BIG_ENDIAN.
25071 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
25072 add to TARGET_DEFAULT flags.
25073 Expand ASM_SPEC and LINK_SPEC.
25074 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
25075 * config/microblaze/microblaze.md: Update extendsidi2 and
25076 movdi_internal instructions to use low-order / high-order reg
25078 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
25079 options and inversemask / mask of LITTLE_ENDIAN.
25080 * config/microblaze/t-microblaze: Expand multilib options to
25081 include mlittle-endian (le) and update exceptions patterns.
25083 2013-02-08 Jakub Jelinek <jakub@redhat.com>
25085 PR rtl-optimization/56195
25086 * lra-constraints.c (get_reload_reg): Don't reuse regs
25087 if they have smaller mode than requested, if they have
25088 wider mode than requested, try to return a SUBREG.
25090 PR tree-optimization/56250
25091 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
25092 if type is unsigned and code isn't MULT_EXPR.
25094 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
25096 PR tree-optimization/56064
25097 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
25098 bits according to mode.
25099 * fixed-value.h (fixed_from_double_int)
25100 (const_fixed_from_double_int): Adjust comments.
25102 2013-02-08 Richard Biener <rguenther@suse.de>
25105 * lto-streamer.h (struct data_in): Remove current_file, current_line
25106 and current_col members.
25107 * lto-streamer-out.c (lto_output_location): Stream changed bits
25108 en-block for efficiency.
25109 * lto-streamer-in.c (clear_line_info): Remove.
25110 (lto_input_location): Cache current file, line and column
25111 globally via local statics. Read changed bits en-block.
25112 (input_function): Do not call clear_line_info.
25113 (lto_read_body): Likewise.
25114 (lto_input_toplevel_asms): Likewise.
25116 2013-02-08 Michael Matz <matz@suse.de>
25118 PR tree-optimization/52448
25119 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
25120 (nt_call_phase): New static.
25121 (add_or_mark_expr): Only mark accesses with newer phase than any
25123 (nonfreeing_call_p): New.
25124 (nt_init_block): Update nt_call_phase, mark blocks as visited.
25125 (nt_fini_block): Keep blocks marked as visited.
25126 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
25128 2013-02-08 Richard Biener <rguenther@suse.de>
25130 * ira.c (ira): Free broken dominator information.
25132 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
25134 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
25136 2013-02-08 Marek Polacek <polacek@redhat.com>
25138 * cfgloop.c (verify_loop_structure): Add more checking of headers.
25140 2013-02-08 Richard Biener <rguenther@suse.de>
25142 PR middle-end/56181
25143 * cfgloop.h (flow_loops_find): Adjust.
25144 (bb_loop_header_p): Declare.
25145 * cfgloop.c (bb_loop_header_p): New function split out from ...
25146 (flow_loops_find): ... here. Adjust function signature,
25147 support incremental loop structure update.
25148 (verify_loop_structure): Cleanup. Verify a loop is a loop.
25149 * cfgloopmanip.c (fix_loop_structure): Move ...
25150 * loop-init.c (fix_loop_structure): ... here.
25151 (apply_loop_flags): Split out from ...
25152 (loop_optimizer_init): ... here.
25153 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
25154 in incremental mode, only remove dead loops here.
25156 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
25159 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
25160 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
25161 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
25162 (*round<mode>3.libgcc): New insns for fixed-modes.
25163 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
25164 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
25165 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
25166 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
25167 implementations. Define to __builtin_avr_absFX,
25168 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
25169 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
25170 __builtin_avr_countlsFX, respectively.
25171 * config/avr/avr-c.c (target.h): Include it.
25172 (enum avr_builtin_id): New enum.
25173 (avr_resolve_overloaded_builtin): New static function.
25174 (avr_register_target_pragmas): Use it to set
25175 targetm.resolve_overloaded_builtin.
25176 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
25177 tree nodes used by DEF_BUILTIN.
25178 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
25179 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
25180 <AVR_BUILTIN_xxBITS>: Same.
25182 2013-02-08 Richard Biener <rguenther@suse.de>
25184 * cfgloop.c (verify_loop_structure): Properly handle
25185 a loop exiting to another loop header.
25186 * ira-int.h (ira_loops): Remove.
25187 * ira.c (ira_loops): Remove.
25188 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
25189 (do_reload): Use loop_optimizer_finalize.
25190 * ira-build.c (create_loop_tree_nodes): Use get_loops and
25191 number_of_loops to access the loop tree.
25192 (more_one_region_p): Likewise.
25193 (finish_loop_tree_nodes): Likewise.
25194 (rebuild_regno_allocno_maps): Likewise.
25195 (mark_loops_for_removal): Likewise.
25196 (mark_all_loops_for_removal): Likewise.
25197 (remove_unnecessary_regions): Likewise.
25198 (ira_build): Likewise.
25199 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
25201 2013-02-08 Richard Biener <rguenther@suse.de>
25203 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
25204 * ipa-pure-const.c (analyze_function): Avoid calling
25205 mark_irreducible_loops twice.
25206 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
25208 2013-02-07 David S. Miller <davem@davemloft.net>
25210 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
25212 * var-tracking.c (vt_add_function_parameter): Test the presence of
25213 HAVE_window_save properly and do not remap argument registers when
25214 we have a leaf function.
25216 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
25219 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
25221 * config/i386/i386.c (ix86_print_operand): Ditto.
25223 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
25225 * lra-constraints.c (process_alt_operands): Fix recently added comment.
25227 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
25229 PR rtl-optimization/56225
25230 * lra-constraints.c (process_alt_operands): Check that reload hard
25231 reg can hold value for strict_low_part.
25233 2013-02-07 Jakub Jelinek <jakub@redhat.com>
25236 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
25237 dwarf2out_end_function.
25238 (in_first_function_p, maybe_at_text_label_p,
25239 first_loclabel_num_not_at_text_label): New variables.
25240 (dwarf2out_var_location): In the first function find out
25241 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
25242 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
25245 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
25247 PR rtl-optimization/56178
25248 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
25249 SUBREG of a register. Tidy up related block of code.
25250 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
25251 note if the source is a register or a SUBREG of a register.
25253 2013-02-07 Jakub Jelinek <jakub@redhat.com>
25256 * config/rs6000/rs6000.md (ptrm): New mode attr.
25257 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
25258 call_value_indirect_aix<pttrsize>,
25259 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
25262 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25264 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
25265 if -bnortl. Convert to strcmp and strncmp.
25267 2013-02-07 Alan Modra <amodra@gmail.com>
25270 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
25271 addresses won't wrap when offsetting.
25272 (rs6000_secondary_reload): Provide secondary reloads needed for
25273 wrapping LO_SUM addresses.
25275 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
25277 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
25278 MACH, just __MACH__.
25280 2013-02-06 Richard Biener <rguenther@suse.de>
25282 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
25283 instead of calling fix_loop_structure.
25285 2013-02-06 Jakub Jelinek <jakub@redhat.com>
25287 PR middle-end/56217
25288 * omp-low.c (use_pointer_for_field): Return false if
25289 lower_send_shared_vars doesn't generate any copy-out code.
25291 2013-02-06 Tom de Vries <tom@codesourcery.com>
25293 PR rtl-optimization/56131
25294 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
25295 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
25296 of the label is NULL. Add comment.
25298 2013-02-05 Jakub Jelinek <jakub@redhat.com>
25300 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
25303 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
25304 (STATIC_LIBTSAN_LIBS): Likewise.
25305 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
25306 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
25307 is defined, don't add anything else beyond that.
25308 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
25309 (LINK_COMMAND_SPEC): Use them.
25311 PR tree-optimization/56205
25312 * tree-stdarg.c (check_all_va_list_escapes): Return true if
25313 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
25314 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
25316 2013-02-05 Richard Biener <rguenther@suse.de>
25318 PR tree-optimization/53342
25319 PR tree-optimization/53185
25320 * tree-vectorizer.h (vect_check_strided_load): Remove.
25321 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
25322 not disallow peeling for vectorized strided loads.
25323 (vect_check_strided_load): Make static and simplify.
25324 (vect_analyze_data_refs): Adjust.
25325 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
25326 correctly when vectorizing strided loads.
25328 2013-02-05 Richard Biener <rguenther@suse.de>
25330 * doc/install.texi: Refer to ISL, not PPL.
25332 2013-02-05 Jan Hubicka <jh@suse.cz>
25334 PR tree-optimization/55789
25335 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
25337 2013-02-05 Jan Hubicka <jh@suse.cz>
25339 PR tree-optimization/55789
25340 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
25341 the dead call anyway.
25343 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
25346 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
25348 2013-02-04 Alexander Potapenko <glider@google.com>
25349 Jack Howarth <howarth@bromo.med.uc.edu>
25350 Jakub Jelinek <jakub@redhat.com>
25353 * config/darwin.c (sort_ctor_records): Stabilized qsort
25354 on constructor priority by using original position.
25355 (finalize_ctors): New routine to sort constructors by
25356 priority before use in assemble_integer.
25357 (machopic_asm_out_constructor): Use finalize_ctors if needed.
25359 2013-02-04 Jakub Jelinek <jakub@redhat.com>
25362 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
25363 about visibility on artificial decls.
25364 * config/sol2.c (solaris_assemble_visibility): Likewise.
25366 2013-02-04 Kai Tietz <ktietz@redhat.com>
25369 * config/i386/i386.c (function_value_ms_64): Add additional valtype
25370 argument and improve checking of return-argument types for 16-byte
25372 (ix86_function_value_1): Add additional valtype argument on call
25373 of function_value_64.
25374 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
25375 handling infunction_value_64 function.
25377 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
25379 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
25381 2013-02-04 Richard Biener <rguenther@suse.de>
25383 PR tree-optimization/56188
25384 * tree-ssa-structalias.c (label_visit): Consider case with
25385 initially non-empty points-to set.
25386 (perform_var_substitution): Dump node mapping and clean up.
25388 2013-02-04 Richard Guenther <rguenther@suse.de>
25391 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
25392 node prevail as last resort.
25393 (lto_symtab_merge_decls): Remove guard on LTRANS here.
25394 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
25396 2013-02-04 Richard Biener <rguenther@suse.de>
25398 PR tree-optimization/56113
25399 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
25401 (equiv_class_lookup_or_add): ... this.
25402 (label_visit): Adjust and fix error in previous patch.
25403 (perform_var_substitution): Adjust.
25405 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
25407 * config/sh/divtab.c: Fix formatting and comments throughout the file.
25408 * config/sh/sh4-300.md: Likewise.
25409 * config/sh/sh4a.md: Likewise.
25410 * config/sh/constraints.md: Likewise.
25411 * config/sh/sh.md: Likewise.
25412 * config/sh/netbsd-elf.h: Likewise.
25413 * config/sh/predicates.md: Likewise.
25414 * config/sh/sh-protos.h: Likewise.
25415 * config/sh/ushmedia.h: Likewise.
25416 * config/sh/linux.h: Likewise.
25417 * config/sh/sh.c: Likewise.
25418 * config/sh/superh.h: Likewise.
25419 * config/sh/elf.h: Likewise.
25420 * config/sh/sh4.md: Likewise.
25421 * config/sh/sh.h: Likewise.
25423 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25425 * config/pa/constraints.md: Adjust unused letters. Change "T"
25426 constraint to match_test floating_point_store_memory_operand().
25427 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
25428 (base14_operand): New.
25429 (floating_point_store_memory_operand): New.
25430 (integer_store_memory_operand): Revise to use base14_operand and
25431 reg_plus_base_memory_operand.
25432 (move_dest_operand): Allow symbolic_memory_operands.
25433 (symbolic_memory_operand): Check for LO_SOM.
25434 (symbolic_operand): Change default case to break.
25435 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
25436 CONST_DOUBLE values to be reloaded by putting them into memory when
25437 the destination is a floating point register.
25438 (movdf): Remove code to handle CONST_DOUBLE.
25440 (reload_indf_r1): New.
25441 (reload_insf_r1): New.
25442 Consistently use "Q" and "T" constraints with integer and floating
25443 point move instructions, respectively.
25444 (movdi): Remove FAIL.
25445 Change predicate for source operand unamed DImode move from
25446 general_operand to move_src_operand.
25447 (umulsidi3): Change predicate for destination operand to
25449 Likewise for similar unamed patterns.
25450 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
25451 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
25452 (hppa_legitimize_address): Simplify mask calculation.
25453 (pa_emit_move_sequence): Revised handling of secondary reloads from
25454 REG+D addresses for floating point loads and stores. Directly handle
25455 loading CONST0_RTX (mode) to a floating point register.
25456 (pa_secondary_reload): Handle reloading DF and SFmode constant values
25457 to floating point registers. Don't restrict secondary reloads to
25458 floating point registers to integer modes. Revise some comments and
25460 (TARGET_LEGITIMATE_ADDRESS_P): Define.
25461 (pa_legitimate_address_p): New.
25462 (pa_legitimize_reload_address): New.
25463 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
25464 (STRICT_REG_OK_FOR_BASE_P): New.
25465 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
25466 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
25468 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
25469 Andrew Dixie <andrewd@gentrack.com>
25471 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
25474 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
25476 * expmed.c (extract_bit_field_1): Pass the full width of the
25477 structure to get_best_reg_extraction_insn.
25479 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25482 * configure.ac (use_cxa_atexit): Add AIX.
25483 * configure: Regenerate.
25485 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
25487 2013-02-01 Jakub Jelinek <jakub@redhat.com>
25490 * final.c (need_profile_function): New variable.
25491 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
25492 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
25493 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
25494 notes, targetm.asm_out.function_prologue doesn't emit anything,
25495 HAVE_prologue and profiler should be emitted before prologue,
25496 set need_profile_function instead of emitting it.
25497 (final_scan_insn): If need_profile_function, emit
25498 profile_function on the first NOTE_INSN_BASIC_BLOCK or
25499 NOTE_INSN_FUNCTION_BEG note.
25501 2013-02-01 Richard Henderson <rth@redhat.com>
25503 * config/rs6000/rs6000.md (smulditi3): New.
25506 * config/alpha/alpha.md (umulditi3): New.
25508 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25510 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
25511 (ASM_OUTPUT_ALIGNED_LOCAL): New.
25513 2013-02-01 Richard Biener <rguenther@suse.de>
25515 PR tree-optimization/56113
25516 * tree-ssa-structalias.c (label_visit): Reduce work for
25517 single-predecessor nodes.
25519 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
25521 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
25522 range isn't testing for zero.
25524 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
25526 PR middle-end/56113
25527 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
25529 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
25530 Nick Clifton <nickc@redhat.com>
25532 * config/v850/constraints.md (Q): Define as a memory constraint.
25533 * config/v850/predicates.md (label_ref_operand): New predicate.
25534 (e3v5_shift_operand): New predicate.
25535 (ior_operator): New predicate.
25536 * config/v850/t-v850: Add e3v5 multilib.
25537 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
25538 (v850_gen_movdi): Prototype.
25539 * config/v850/v850.c: Add support for e3v5 architecture.
25540 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
25542 (construct_save_jarl): Add e3v5 long JARL support.
25543 (v850_adjust_insn_length): New function. Adjust length of call
25544 insns when using e3v5 instructions.
25545 (v850_gen_movdi): New function: Generate instructions to move a
25547 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
25548 (CPP_SPEC): Define __v850e3v5__ as appropriate.
25549 (TARGET_USE_FPU): Enable for e3v5.
25550 (CONST_OK_FOR_W): New macro.
25551 (ADJUST_INSN_LENGTH): Define.
25552 * config/v850/v850.md (UNSPEC_LOOP): Define.
25553 (attr cpu): Add v850e3v5.
25554 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
25555 (movdi): New pattern.
25556 (movdi_internal): New pattern.
25557 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
25558 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
25559 (cstoresf4): Likewise.
25560 (cstoredf4): Likewise.
25561 (insv): New pattern.
25562 (rotlso3_a): New pattern.
25563 (rotlsi3_b): New pattern
25564 (rotlsi3_v850e3v5): New pattern.
25565 (doloop_begin): New pattern.
25566 (fix_loop_counter): New pattern.
25567 (doloop_end): New pattern.
25568 (branch_normal): Add e3v5 long branch support.
25569 (branch_invert): Likewise.
25570 (branch_z_normal): Likewise.
25571 (branch_z_invert): Likewise.
25572 (branch_nz_normal): Likewise.
25573 (branch_nz_invert): Likewise.
25574 (call_internal_short): Add e3v5 register-indirect JARL support.
25575 (call_internal_long): Likewise.
25576 (call_value_internal_short): Likewise.
25577 (call_value_internal_long): Likewise.
25578 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
25579 (mloop): New option.
25580 * config.gcc: Add support for configuring v840e3v5 target.
25581 * doc/invoke.texi: Document new v850 specific command line options.
25583 2013-01-31 Paul Koning <ni1d@arrl.net>
25587 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
25588 children if parent is a class.
25589 (prune_unused_types_prune): Don't add DW_AT_declaration.
25591 2013-01-31 Richard Biener <rguenther@suse.de>
25593 PR tree-optimization/56157
25594 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
25595 match up operand with SLP child.
25597 2013-01-31 Jason Merrill <jason@redhat.com>
25600 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
25601 parameters the first time.
25602 (gen_scheduled_generic_parms_dies): Check completeness here.
25604 2013-01-31 Richard Biener <rguenther@suse.de>
25606 PR middle-end/53073
25607 * common.opt (faggressive-loop-optimizations): New flag,
25608 enabled by default.
25609 * doc/invoke.texi (faggressive-loop-optimizations): Document.
25610 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
25611 infer_loop_bounds_from_undefined by it.
25613 2013-01-31 Richard Biener <rguenther@suse.de>
25615 PR tree-optimization/56150
25616 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
25617 visit virtual operands.
25618 (find_uses_to_rename_bb): Likewise.
25620 2013-01-31 Richard Biener <rguenther@suse.de>
25622 PR tree-optimization/56150
25623 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
25624 mixed store non-store stmts.
25626 2013-01-30 Jakub Jelinek <jakub@redhat.com>
25629 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
25630 LIBASAN_EARLY_SPEC is defined.
25631 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
25632 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
25634 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
25637 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
25638 invalid args instead of ICEing on it.
25639 (ix86_valid_target_attribute_tree): Return error_mark_node if
25640 ix86_valid_target_attribute_inner_p failed.
25641 (ix86_valid_target_attribute_p): Return false only if
25642 ix86_valid_target_attribute_tree returned error_mark_node. Allow
25643 target("default") attribute.
25644 (sorted_attr_string): Change argument from const char * to tree,
25645 merge in all target attribute arguments rather than just one.
25646 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
25647 instead of free. Avoid using strcat.
25648 (ix86_mangle_function_version_assembler_name): Mangle
25649 target("default") as if no target attribute is present. Adjust
25650 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
25651 instead of xmalloc and XDELETEVEC instead of free.
25652 (ix86_function_versions): Don't return true if one of the decls
25653 doesn't have target attribute. If they don't and one of the decls
25654 is DECL_FUNCTION_VERSIONED, report an error. Adjust
25655 sorted_attr_string caller. Use XDELETEVEC instead of free.
25656 (ix86_supports_function_versions): Remove.
25657 (make_name): Fix up formatting.
25658 (make_dispatcher_decl): Remove resolver_name and its initialization.
25659 Avoid leaking memory.
25660 (is_function_default_version): Return true if there is
25661 target("default") attribute rather than no target attribute at all.
25662 (make_resolver_func): Avoid leaking memory.
25663 (ix86_generate_version_dispatcher_body): Likewise.
25664 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
25665 * target.def (supports_function_versions): Remove.
25666 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
25667 * doc/tm.texi: Regenerated.
25669 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
25671 PR rtl-optimization/56144
25672 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
25673 for values with side effects.
25675 2013-01-30 Richard Biener <rguenther@suse.de>
25677 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
25678 (sparseset_pop): Likewise.
25679 * cfganal.c (compute_idf): Likewise. Increase work-stack size
25680 to be able to use quick_push in the worker loop.
25682 2013-01-30 Marek Polacek <polacek@redhat.com>
25684 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
25686 2013-01-30 Richard Biener <rguenther@suse.de>
25689 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
25691 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
25693 PR tree-optimization/56064
25694 * fixed-value.c (fixed_from_double_int): New function.
25695 * fixed-value.h (fixed_from_double_int): New prototype.
25696 (const_fixed_from_double_int): New static inline function.
25697 * fold-const.c (native_interpret_fixed): New static function.
25698 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
25699 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
25700 (native_encode_fixed): New static function.
25701 (native_encode_expr) <FIXED_CST>: Use it.
25702 (native_interpret_int): Move double_int worker code to...
25703 * double-int.c (double_int::from_buffer): ...this new static method.
25704 * double-int.h (double_int::from_buffer): Prototype it.
25706 2013-01-30 Richard Biener <rguenther@suse.de>
25708 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
25709 New pointer-map and obstack.
25710 (init_alias_vars): Allocate pointer-map and obstack.
25711 (delete_points_to_sets): Free them.
25712 (find_what_var_points_to): Cache result.
25713 (find_what_p_points_to): Adjust for changed interface of
25714 find_what_var_points_to.
25715 (compute_points_to_sets): Likewise.
25716 (ipa_pta_execute): Likewise.
25718 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25720 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
25721 * configure: Regenerate.
25722 * config.in: Regenerate.
25723 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
25724 #nobits/#progbits if supported.
25726 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
25729 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
25730 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
25731 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
25733 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25735 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
25736 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
25738 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25740 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
25741 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
25743 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25745 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
25747 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
25748 * config/arm/cortex-a7.md: New bypasses using
25749 arm_mac_accumulator_is_result.
25751 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25753 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
25754 (cortex_a7_neon_mla): Likewise.
25755 (cortex_a7_fpfmad): New reservation.
25756 (cortex_a7_fpmacs): Use ffmas and update required units.
25757 (cortex_a7_fpmuld): Update required units and latency.
25758 (cortex_a7_fpmacd): Likewise.
25759 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
25760 (cortex_a7_neon). Likewise.
25761 (bypass) Update participating units.
25763 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25765 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
25766 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
25768 * config/arm/vfp11.md (vfp_farith): Use ffmas.
25769 (vfp_fmul): Use ffmad.
25770 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
25771 (cortex_r4_fmacd): Use ffmad.
25772 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
25773 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
25774 (cortex_a9_fmacd): Use ffmad.
25775 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
25776 (cortex_a8_vfp_macd): Use ffmad.
25777 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
25778 (cortex_a5_fpmacd): Use ffmad.
25779 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
25780 (cortex_a15_vfp_macd): Use ffmad.
25781 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
25783 2013-01-29 Jason Merrill <jason@redhat.com>
25786 * varasm.c (default_assemble_visibility): Don't warn about
25787 visibility on artificial decls.
25789 2013-01-29 Richard Biener <rguenther@suse.de>
25791 PR tree-optimization/56113
25792 * tree-ssa-structalias.c (equiv_class_lookup): Also return
25794 (label_visit): Free duplicate bitmaps and record the leader instead.
25795 (perform_var_substitution): Adjust.
25797 2013-01-29 Richard Biener <rguenther@suse.de>
25799 PR tree-optimization/55270
25800 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
25801 the CFG, schedule loops for fixup.
25803 2013-01-29 Nick Clifton <nickc@redhat.com>
25805 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
25808 2013-01-28 Leif Ekblad <leif@rdos.net>
25810 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
25811 * config/i386/i386.h (TARGET_RDOS): New macro.
25812 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
25813 * config/i386/i386.c (ix86_option_override_internal): For 64bit
25814 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
25815 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
25816 DEFAULT_LARGE_SECTION_THRESHOLD.
25817 * config/i386/i386.md (R14_REG, R15_REG): New constants.
25818 * config/i386/rdos.h: New file.
25819 * config/i386/rdos64.h: New file.
25821 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
25824 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
25827 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25829 PR rtl-optimization/56117
25830 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
25831 call cselib_lookup_from_insn on the MEM before calling
25832 add_insn_mem_dependence.
25834 2013-01-28 Richard Biener <rguenther@suse.de>
25836 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
25837 to a stmt that didn't have one.
25838 (copy_phis_for_bb): Likewise for PHI arguments.
25839 (copy_debug_stmt): Likewise for debug stmts.
25841 2013-01-28 Richard Biener <rguenther@suse.de>
25843 PR tree-optimization/56034
25844 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
25845 (partition_builtin_p): Adjust.
25846 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
25847 it is the last partition.
25848 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
25849 up the vertex for the definition.
25850 (classify_partition): Classify whether a partition is a
25851 PKIND_REDUCTION, thus has uses outside of the loop.
25852 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
25853 Merge all PKIND_REDUCTION partitions into the last partition.
25854 (tree_loop_distribution): Seed partitions from reductions as well.
25856 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25858 PR tree-optimization/56125
25859 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
25860 pow(x,c) into sqrt(x) * powi(x, n/2) or
25861 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
25862 optimizing for size.
25863 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
25864 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
25867 PR tree-optimization/56094
25868 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
25869 to UNKNOWN_LOCATION while gimplifying expr.
25871 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
25874 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
25875 operand 0 in movabs insn template for -masm=intel asm alternative.
25876 (*movabs<mode>_2): Ditto for operand 1.
25878 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
25881 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
25884 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25886 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
25889 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25891 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
25892 Undef to avoid warning.
25894 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25896 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
25897 * configure: Regenerate.
25899 2013-01-25 Jakub Jelinek <jakub@redhat.com>
25901 PR tree-optimization/56098
25902 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
25903 for stmts with volatile ops.
25904 (cond_store_replacement): Don't optimize if assign has volatile ops.
25905 (cond_if_else_store_replacement_1): Don't optimize if either
25906 then_assign or else_assign have volatile ops.
25907 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
25910 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25912 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
25914 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25916 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
25917 missing ':' in asm example.
25919 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
25921 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
25922 entries into lane and laneq entries.
25923 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
25924 Remove AdvSIMD scalar modes.
25925 (aarch64_sq<r>dmulh_laneq<mode>): New.
25926 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
25928 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
25929 builtin implementations to relfect changes in RTL in aarch64-simd.md.
25930 * config/aarch64/iterators.md (VCOND): New.
25933 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25936 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
25937 Add NULL LIBNAME argument to existing definitions.
25938 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
25939 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
25940 * config/avr/avr.c (DEF_BUILTIN): Same.
25941 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
25942 (avr_expand_builtin): Expand to a vanilla call if a libgcc
25943 implementation is available (DECL_ASSEMBLER_NAME is set).
25944 (avr_fold_absfx): New static function.
25945 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
25946 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
25947 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
25948 AVR_BUILTIN_ABSLLK.
25949 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
25950 (abshk, absk, abslk, absllk): Provide as static inline functions.
25952 2013-01-25 Marek Polacek <polacek@redhat.com>
25954 PR tree-optimization/56035
25955 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
25957 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
25959 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
25960 (*movtf_internal_rex64): Add (!o,C) alternative
25961 (*movxf_internal_rex64): Ditto.
25962 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
25964 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
25966 * doc/invoke.texi: fix typo.
25967 * doc/objc.texi: fix typo.
25969 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
25971 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
25972 for the first two alternatives.
25974 2013-01-24 Diego Novillo <dnovillo@google.com>
25976 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
25977 (ggc-zone.o): Remove.
25978 * configure.ac: Remove option --with-gc.
25979 * configure: Re-generate.
25980 * doc/install.texi: Remove documentation for --with-gc.
25981 * gengtype.c (write_enum_defn): Remove. Update all users.
25982 (write_Types_process_field): Remove generation of gt_e_* argument.
25983 (output_type_enum): Remove. Update all users.
25984 (write_enum_defn): Remove. Update all users.
25985 (enum alloc_zone): Remove. Update all users.
25986 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
25987 * ggc-common.c (ggc_splay_alloc): Remove first argument.
25988 Update all callers.
25989 (struct ptr_data): Remove field TYPE. Update all users.
25990 (gt_pch_note_object): Remove argument TYPE. Update all users.
25991 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
25993 * ggc-none.c (ggc_alloc_typed_stat): Remove.
25994 (struct alloc_zone): Remove.
25995 (ggc_internal_alloc_zone_stat): Remove.
25996 (ggc_internal_cleared_alloc_zone_stat): Remove.
25997 * ggc-page.c (ggc_alloc_typed_stat): Remove.
25998 (ggc_pch_count_object): Remove last argument. Update all users.
25999 (ggc_pch_alloc_object): Remove last argument. Update all users.
26000 (struct alloc_zone): Remove.
26001 * ggc-zone.c: Remove.
26002 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
26003 (struct alloc_zone): Remove.
26004 (ggc_alloc_typed_stat): Remove.
26005 (ggc_alloc_typed): Remove.
26006 (ggc_splay_alloc): Remove first argument.
26007 (rtl_zone): Remove. Update all users.
26008 (tree_zone): Remove. Update all users.
26009 (tree_id_zone): Remove. Update all users.
26010 (ggc_internal_zone_alloc_stat): Remove. Update all users.
26011 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
26012 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
26013 * tree-ssanames.c: Remove references to zone allocator in comments.
26015 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
26017 * config/avr/avr.c (avr_out_fract): Make register numbers that
26018 might be outside of source operand signed.
26020 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
26022 * config/i386/constraints.md (Yf): New constraint.
26023 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
26024 of f constraint to conditionaly disable x87 register preferences.
26025 (*movdf_internal): Ditto.
26026 (*movsf_internal): Ditto.
26028 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
26030 PR inline-asm/55934
26031 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
26032 that have operands with impossible constraints.
26033 Add a FIXME for a speed-up opportunity.
26034 * lra-constraints.c (process_alt_operands): Verify that a class
26035 selected from constraints on asms is valid for the operand mode.
26036 (curr_insn_transform): Remove incorrect comment.
26038 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
26040 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
26041 TOC operand is a valid symbol ref in the constant pool.
26043 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26045 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
26047 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
26050 * config/avr/stdfix.h: New file.
26051 * t-avr (stdfix-gcc.h): New rule to build it.
26052 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
26054 2013-01-23 Kostya Serebryany <kcc@google.com>
26056 * config/darwin.h: remove dependency on
26057 CoreFoundation (asan on Mac OS).
26059 2013-01-23 Jakub Jelinek <jakub@redhat.com>
26062 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
26063 instead of cmpdi_operand for first comparison operand.
26064 Don't assert that comparison operands aren't both constants.
26066 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
26068 * doc/install.texi (Downloading the Source): Update references to
26069 downloading separate components.
26071 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
26073 * doc/extend.texi (__int128): Improve grammar.
26075 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
26078 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
26079 alternative to (o,r).
26080 (*movdi_internal_rex64): Remove (!o,n) alternative.
26081 (DImode immediate->memory splitter): Remove.
26082 (DImode immediate->memory peephole2): Remove.
26083 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
26084 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
26085 alternative to (!o,*r).
26086 (*movtf_internal_sse): New pattern.
26087 (*movxf_internal_rex64): New pattern.
26088 (*movxf_internal): Disable for TARGET_64BIT.
26089 (*movdf_internal_rex64): Remove (!o,F) alternative.
26091 2013-01-22 Jakub Jelinek <jakub@redhat.com>
26093 PR middle-end/56074
26094 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
26095 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
26096 * tree-vect-loop-manip.c (find_loop_location): Also ignore
26097 stmt locations where LOCATION_LOCUS of the stmt location is
26098 UNKNOWN_LOCATION or BUILTINS_LOCATION.
26101 * config/i386/i386.md (UNSPEC_STOS): New.
26102 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
26103 *strsetqi_1): Add UNSPEC_STOS.
26105 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
26108 * doc/invoke.texi: Remove left over -Wsynth example.
26110 2013-01-21 Jakub Jelinek <jakub@redhat.com>
26112 PR tree-optimization/56051
26113 * fold-const.c (fold_binary_loc): Don't fold
26114 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
26115 a narrowing conversion, or widening conversion from signed
26118 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
26120 PR rtl-optimization/56023
26121 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
26122 dependent on debug instruction.
26124 2013-01-21 Martin Jambor <mjambor@suse.cz>
26126 PR middle-end/56022
26127 * function.c (allocate_struct_function): Call
26128 invoke_set_current_function_hook earlier.
26130 2013-01-21 Jakub Jelinek <jakub@redhat.com>
26132 * reload1.c (init_reload): Only initialize reload_obstack
26133 during the first call.
26135 2013-01-21 Marek Polacek <polacek@redhat.com>
26137 * cfgloop.c (verify_loop_structure): Fix up grammar.
26139 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
26141 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
26142 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
26144 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26147 * config/arm/marvell-pj4.md: Update copyright year.
26148 Fix up use of alu to alu_reg and simple_alu_imm.
26150 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
26152 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
26154 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
26157 * lra-constraints.c (curr_insn_transform): Don't reuse original
26158 insn for secondary memory move when memory mode should be different.
26160 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26162 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
26163 atomic_storedi_1): New patterns.
26165 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26167 btver2 pipeline descriptions.
26168 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
26170 * config/i386/i386.md (btver2_decode): New type attributes.
26171 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
26173 * config/i386/btver2.md: New file describing btver2 pipelines.
26175 2013-01-19 Andrew Pinski <apinski@cavium.com>
26177 PR tree-optimization/52631
26178 * tree-ssa-sccvn (visit_use): Before looking up the original
26179 statement, try looking up the simplified expression.
26181 2013-01-19 Anthony Green <green@moxielogic.com>
26183 * config/moxie/moxie.c (moxie_expand_prologue): Set
26184 current_function_static_stack_size.
26186 2013-01-18 Jakub Jelinek <jakub@redhat.com>
26188 PR tree-optimization/56029
26189 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
26190 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
26192 2013-01-18 Sharad Singhai <singhai@google.com>
26194 PR tree-optimization/55995
26195 * dumpfile.c (dump_loc): Print location only if available.
26196 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
26198 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
26201 * lra-constraints.c (curr_insn_transform): Reuse original insn for
26202 secondary memory move.
26203 (inherit_reload_reg): Use rclass instead of cl for
26204 check_secondary_memory_needed_p.
26206 2013-01-18 Jakub Jelinek <jakub@redhat.com>
26208 PR middle-end/56015
26209 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
26210 the case where writing real complex part of target modifies op1.
26212 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
26214 * config/aarch64/aarch64-simd.md
26215 (aarch64_vcond_internal<mode>): Handle unordered cases.
26216 * config/aarch64/iterators.md (v_cmp_result): New.
26218 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
26219 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26221 * config/arm/marvell-pj4.md: New file.
26222 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
26223 * config/arm/arm.md (generic_sched): Add marvell_pj4.
26224 (generic_vfp): Likewise.
26225 * config/arm/arm-cores.def: Add marvell-pj4.
26226 * config/arm/arm-tune.md: Regenerate.
26227 * config/arm/arm-tables.opt: Regenerate.
26228 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
26229 * doc/invoke.texi: Document marvell-pj4.
26231 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
26233 * config/aarch64/arm_neon.h: Map scalar types to standard types.
26235 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26240 * var-tracking.c (negative_power_of_two_p): New.
26241 (global_get_addr_cache, local_get_addr_cache): New.
26242 (get_addr_from_global_cache, get_addr_from_local_cache): New.
26243 (vt_canonicalize_addr): Rewrite using the above. Adjust the
26245 (vt_stack_offset_p): Remove.
26246 (vt_canon_true_dep): Always canonicalize loc's address.
26247 (clobber_overlapping_mems): Make sure we have a MEM.
26248 (local_get_addr_clear_given_value): New.
26249 (val_reset): Clear local cached entries.
26250 (compute_bb_dataflow): Create and release the local cache.
26251 Disable duplicate MEMs clobbering.
26252 (emit_notes_in_bb): Clobber MEMs likewise.
26253 (vt_emit_notes): Create and release the local cache.
26254 (vt_initialize, vt_finalize): Create and release the global
26255 cache, respectively.
26256 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
26258 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26260 PR libmudflap/53359
26261 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
26262 not found in the symtab.
26264 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26267 PR rtl-optimization/55547
26268 PR rtl-optimization/53827
26271 * alias.c (offset_overlap_p): New, factored out of...
26272 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
26273 the conservative special case for symbolic constants. Don't
26274 adjust zero sizes on alignment.
26276 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
26278 PR rtl-optimization/52573
26279 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
26280 REG_UNUSED for the same register.
26282 2013-01-17 Richard Biener <rguenther@suse.de>
26283 Marek Polacek <polacek@redhat.com>
26285 PR rtl-optimization/55833
26286 * loop-unswitch.c (unswitch_loops): Move loop verification...
26287 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
26288 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
26289 Set it to true when we're removing a loop from hierarchy tree in
26290 an irreducible region.
26291 (fix_bb_placements): Adjust caller.
26292 (fix_loop_placements): Likewise.
26294 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
26296 * config/avr/builtins.def (DEF_BUILTIN): Factor out
26297 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
26298 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
26299 Remove ID. Adjust comments.
26300 * config/avr/avr-c.c (avr_builtin_name): Remove.
26301 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
26302 * config/avr/avr.c (avr_tolower): New static function.
26303 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
26304 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
26305 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
26308 2013-01-17 Jan Hubicka <jh@suse.cz>
26310 PR tree-optimization/55273
26311 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
26313 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
26316 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
26317 store through atomic_store<mode>_1.
26318 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
26320 2013-01-17 Martin Jambor <mjambor@suse.cz>
26322 PR tree-optimizations/55264
26323 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
26324 for virtual methods.
26325 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
26326 virtual methods before inlining is over.
26327 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
26329 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
26332 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
26334 PR rtl-optimization/56005
26335 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
26336 pending reads for prefetch.
26338 2013-01-16 Ian Bolton <ian.bolton@arm.com>
26340 * config/aarch64/aarch64.md
26341 (*cstoresi_neg_uxtw): New pattern.
26342 (*cmovsi_insn_uxtw): New pattern.
26343 (*<optab>si3_uxtw): New pattern.
26344 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
26345 (*<optab>si3_insn_uxtw): New pattern.
26346 (*bswapsi2_uxtw): New pattern.
26348 2013-01-16 Richard Biener <rguenther@suse.de>
26350 * tree-inline.c (tree_function_versioning): Remove set but
26351 never used variable.
26353 2013-01-16 Richard Biener <rguenther@suse.de>
26355 PR tree-optimization/55964
26356 * tree-flow.h (rename_variables_in_loop): Remove.
26357 (rename_variables_in_bb): Likewise.
26358 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
26359 (copy_loop_before): Adjust and delete update-ssa status.
26360 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
26361 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
26362 (rename_variables_in_loop): Remove.
26363 (slpeel_update_phis_for_duplicate_loop): Likewise.
26364 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
26365 use available cfg machinery instead of duplicating it.
26366 Update PHI nodes and perform poor-mans SSA update here.
26367 (slpeel_tree_peel_loop_to_edge): Adjust.
26369 2013-01-16 Richard Biener <rguenther@suse.de>
26371 PR tree-optimization/54767
26372 PR tree-optimization/53465
26373 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
26374 (vrp_visit_phi_node): For PHI arguments coming via backedges
26375 drop all symbolical range information.
26376 (execute_vrp): Compute backedges.
26378 2013-01-16 Richard Biener <rguenther@suse.de>
26380 * doc/install.texi: Update CLooG and ISL requirements to
26383 2013-01-16 Christian Bruel <christian.bruel@st.com>
26386 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
26387 (broken_move): Handle UNSPECV_SP_SWITCH_B.
26388 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
26390 2013-01-16 DJ Delorie <dj@redhat.com>
26392 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
26393 (UNSPECV_SP_SWITCH_E): New.
26394 (sp_switch_1): Change to an unspec.
26395 (sp_switch_2): Change to an unspec. Don't use post-inc when we
26398 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
26400 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
26401 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
26402 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
26403 (expand_mem_thread_fence): Ditto.
26404 (expand_mem_signal_fence): Ditto.
26405 (expand_atomic_load): Ditto.
26406 (expand_atomic_store): Ditto.
26408 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
26410 PR rtl-optimization/55547
26411 PR rtl-optimization/53827
26414 * alias.c (memrefs_conflict_p): Set sizes to negative after
26417 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26420 * function.c (thread_prologue_and_epilogue_insns): Always
26421 add crtl->drap_reg to set_up_by_prologue.set, even if
26422 stack_realign_drap is false.
26424 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
26426 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
26427 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
26428 *call): Fix indention.
26430 2013-01-15 Tom de Vries <tom@codesourcery.com>
26433 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
26436 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
26438 PR rtl-optimization/55153
26439 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
26441 2013-01-15 Martin Jambor <mjambor@suse.cz>
26443 PR tree-optimization/55920
26444 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
26445 accesses as grp_to_be_debug_replaced.
26447 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26449 PR tree-optimization/55920
26450 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
26451 there is non-useless type conversion needed from debug rhs to lhs,
26452 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
26454 2013-01-15 Joseph Myers <joseph@codesourcery.com>
26455 Mikael Pettersson <mikpe@it.uu.se>
26458 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
26460 (ASM_OUTPUT_CASE_LABEL): Remove.
26461 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
26462 * final.c (shorten_branches): Update alignment of labels before
26463 jump tables if CASE_VECTOR_SHORTEN_MODE.
26465 2013-01-15 Richard Biener <rguenther@suse.de>
26468 * system.h: Do not include gmp.h for building host tools.
26470 2013-01-15 Richard Biener <rguenther@suse.de>
26472 PR middle-end/55882
26473 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
26474 account for bitpos when computing alignment.
26476 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
26478 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
26479 (ix86_target_macros_internal): Likewise.
26481 * config/i386/i386.c (m_CORE2I7): Removed.
26482 (m_CORE_HASWELL): New macro.
26483 (m_CORE_ALL): Likewise.
26484 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
26485 (initial_ix86_arch_features): Likewise.
26486 (processor_target_table): Initializations for Core avx2.
26487 (cpu_names): New names "core-avx2".
26488 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
26489 PROCESSOR_CORE_HASWELL.
26490 (ix86_issue_rate): New case.
26491 (ia32_multipass_dfa_lookahead): Likewise.
26492 (ix86_sched_init_global): Likewise.
26494 * config/i386/i386.h (TARGET_HASWELL): New macro.
26495 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
26496 (processor_type): New PROCESSOR_HASWELL.
26498 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26500 PR tree-optimization/55955
26501 * tree-vect-loop.c (vectorizable_reduction): Give up early on
26502 *SHIFT_EXPR and *ROTATE_EXPR codes.
26504 PR tree-optimization/48766
26505 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
26506 -ftrapv disable -fwrapv.
26508 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26511 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
26512 etc. to 1 and not to __flash.
26513 Use LL suffix for __INT24_MAX__ with -mint8.
26514 Use ULL suffix for __UINT24_MAX__ with -mint8.
26516 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26518 * config/avr/avr-arch.h
26519 (struct base_arch_s): Use typedef avr_arch_t instead.
26520 (struct arch_info_s): Use typedef avr_arch_info_t instead.
26521 (struct mcu_type_s): Use typedef avr_mcu_t instead.
26522 * config/avr/avr.c: Same.
26523 * config/avr/avr-devices.c: Same.
26524 * config/avr/driver-avr.c: Same.
26525 * config/avr/gen-avr-mmcu-texi.c: Same.
26526 * config/avr/avr-mcus.def: Adjust comment.
26528 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
26530 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
26531 * config/aarch64/iterators.md (VALLDI): New.
26533 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
26534 Andi Kleen <ak@linux.intel.com>
26537 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
26538 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
26541 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26543 * config/avr/avr-stdint.h: Remove trailing blanks.
26544 * config/avr/avr-log.h: Same.
26545 * config/avr/avr-arch.h: Same.
26546 * config/avr/avr-devices.c: Same.
26547 * config/avr/avr-dimode.md: Same.
26548 * config/avr/predicates.md: Same.
26549 * config/avr/avr-c.c: Same. And fix typo.
26551 * config/avr/avr-protos.h: Same. And:
26552 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
26553 (init_cumulative_args): Rename to avr_init_cumulative_args.
26554 (expand_prologue): Rename to avr_expand_prologue.
26555 (expand_epilogue): Rename to avr_expand_epilogue.
26556 (adjust_insn_length): Rename to avr_adjust_insn_length.
26557 (notice_update_cc): Rename to avr_notice_update_cc.
26558 (final_prescan_insn): Rename to avr_final_prescan_insn.
26559 * config/avr/avr.c: Same.
26560 * config/avr/avr.h: Same.
26561 * config/avr/avr.md: Remove trailing blanks.
26562 (prologue): Use avr_expand_prologue.
26563 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
26565 2013-01-14 Richard Biener <rguenther@suse.de>
26567 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
26568 verify_location, collect_subblocks): New functions.
26569 (verify_gimple_in_cfg): Verify that locations only reference
26570 BLOCKs in the functions BLOCK tree.
26572 2013-01-14 Richard Biener <rguenther@suse.de>
26574 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
26576 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
26578 (insert_out_of_ssa_copy_on_edge): Likewise.
26579 (rewrite_close_phi_out_of_ssa): Likewise.
26580 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
26582 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
26583 propagated constants.
26584 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
26587 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26589 * config/avr/avr-modes.def: Add GPL copyright notice.
26591 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
26593 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
26594 MEMMODEL_MASK to determine memory model.
26595 (atomic_store<mode>): Ditto from operands[2].
26596 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
26598 2013-01-13 Jakub Jelinek <jakub@redhat.com>
26601 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
26602 (fold_gimple_assign): Don't call unshare_expr here.
26603 (fold_ctor_reference): Call unshare_expr.
26605 2013-01-13 Terry Guo <terry.guo@arm.com>
26607 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
26608 * doc/fragments.texi: Document MULTILIB_REUSE.
26609 * gcc.c (multilib_reuse): New internal spec.
26610 (set_multilib_dir): Also search multilib from multilib_reuse.
26611 * genmultilib (tmpmultilib3): Refactor code.
26612 (tmpmultilib4): Ditto.
26613 (multilib_reuse): New multilib argument.
26615 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
26617 * Makefile.in: Update copyright.
26619 2013-01-12 Tom de Vries <tom@codesourcery.com>
26621 PR middle-end/55890
26622 * calls.c (expand_call): Check if arg_nr is valid.
26624 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
26626 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
26627 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
26628 documentation. Add missing '__' in front of
26629 __builtin_ia32_packssdw256.
26631 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26634 * config/s390/s390.c (s390_preferred_reload_class): Do not return
26635 NO_REGS for larl operands.
26636 (s390_reload_larl_operand): Use s390_load_address instead of
26639 2013-01-11 Richard Biener <rguenther@suse.de>
26641 * tree-cfg.c (verify_node_sharing_1): Split out from ...
26642 (verify_node_sharing): ... here.
26643 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
26645 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
26647 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
26648 Substitute TREECHECKING.
26649 * configure: Regenerate.
26650 * Makefile.in (TREECHECKING): New.
26652 2013-01-11 Richard Guenther <rguenther@suse.de>
26654 PR tree-optimization/44061
26655 * tree-vrp.c (extract_range_basic): Compute zero as
26656 value-range for __builtin_constant_p of function parameters.
26658 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
26660 Update copyright years.
26662 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
26664 PR rtl-optimization/55672
26665 * lra-eliminations.c (mark_not_eliminable): Permit addition with
26666 const to be eliminable.
26668 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
26670 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
26671 * configure: Regenerate.
26673 2013-01-10 Richard Biener <rguenther@suse.de>
26675 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
26677 2013-01-10 Richard Biener <rguenther@suse.de>
26680 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
26681 locations for virtual PHI arguments.
26682 (rewrite_update_phi_arguments): Likewise.
26684 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
26686 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
26689 2013-01-10 Jakub Jelinek <jakub@redhat.com>
26691 PR tree-optimization/55921
26692 * tree-complex.c (expand_complex_asm): New function.
26693 (expand_complex_operations_1): Call it for GIMPLE_ASM.
26695 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26698 * config/s390/s390.c (s390_symref_operand_p)
26699 (s390_loadrelative_operand_p): Merge the two functions.
26700 (s390_check_qrst_address, print_operand_address): Add parameters
26701 to s390_loadrelative_operand_p invokation.
26702 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
26703 (s390_reload_larl_operand, s390_secondary_reload): Use
26704 s390_loadrelative_operand_p instead of s390_symref_operand_p.
26705 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
26707 2013-01-09 Mike Stump <mikestump@comcast.net>
26709 * dse.c (record_store): Remove unnecessary assert.
26711 2013-01-09 Jan Hubicka <jh@suse.cz>
26713 PR tree-optimization/55569
26714 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
26715 * cfgloop.h (scale_loop_profile): Likewise.
26717 2013-01-09 Jan Hubicka <jh@suse.cz>
26720 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
26722 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
26724 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
26726 PR middle-end/55114
26727 * expr.h (maybe_emit_group_store): Declare.
26728 * expr.c (maybe_emit_group_store): New function.
26729 * builtins.c (expand_builtin_int_roundingfn): Call it.
26730 (expand_builtin_int_roundingfn_2): Likewise.
26732 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
26734 PR rtl-optimization/55829
26735 * lra-constraints.c (match_reload): Add code for absent output.
26736 (curr_insn_transform): Add code for reloads of matched inputs
26739 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
26741 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
26742 attribute of movddup insn to DF.
26743 (*vec_interleave_lowv2df): Ditto.
26744 (vec_dupv2df): Ditto.
26746 2013-01-09 Jan Hubicka <jh@suse.cz>
26748 PR tree-optimiation/55875
26749 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
26750 EVERY_ITERATION parameter.
26751 (number_of_iterations_exit): Check if exit is executed every iteration.
26752 (idx_infer_loop_bounds): Similarly here.
26753 (n_of_executions_at_most): Simplify
26754 to only test for cases where statement is dominated by the
26755 particular bound; handle correctly the "postdominance" test.
26756 (scev_probably_wraps_p): Use max loop iterations info
26757 as a global bound first.
26759 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
26760 Nick Clifton <nickc@redhat.com>
26762 * config/v850/v850.md (cbranchsf4): New pattern.
26763 (cstoresf4): New pattern.
26764 (cbranchdf4): New pattern.
26765 (cstoredf4): New pattern.
26766 (movsicc): Disallow floating point comparisons.
26767 (cmpsf_le_insn): Fix order of operators.
26768 (cmpsf_lt_insn): Likewise.
26769 (cmpsf_eq_insn): Likewise.
26770 (cmpdf_le_insn): Likewise.
26771 (cmpdf_lt_insn): Likewise.
26772 (cmpdf_eq_insn): Likewise.
26773 (cmpsf_ge_insn): Use LE comparison.
26774 (cmpdf_ge_insn): Likewise.
26775 (cmpsf_gt_insn): Use LT comparison.
26776 (cmpdf_gt_insn): Likewise.
26777 (cmpsf_ne_insn): Delete pattern.
26778 (cmpdf_ne_insn): Delete pattern.
26779 * config/v850/v850.c (v850_gen_float_compare): Use
26780 gen_cmpdf_eq_insn for NE comparison.
26781 (v850_float_z_comparison_operator)
26782 (v850_float_nz_comparison_operator): Move from here ...
26783 * config/v850/predicates.md: ... to here. Move GT and GE
26784 comparisons into v850_float_z_comparison_operator.
26785 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
26787 (v850_float_nz_comparison_operator): Likewise.
26789 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26791 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
26792 with calls to gen_insvsi/gen_insvdi.
26794 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26796 * config/i386/i386.c (initial_ix86_tune_features): Set up
26797 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
26799 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
26800 Jakub Jelinek <jakub@redhat.com>
26802 PR tree-optimization/48189
26803 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
26804 If nitercst is 0, don't predict the exit edge.
26806 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26808 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
26809 in asm_fprintf with reg_names.
26810 (aarch64_print_operand_address): Likewise.
26811 (aarch64_return_addr): Likewise.
26812 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
26814 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26816 * config/pa/pa.h (VAL_U6_BITS_P): Define.
26817 (INT_U6_BITS): Likewise.
26818 * config/pa/predicates.md (uint6_operand): New predicate.
26819 (shift5_operand, shift6_operand): Likewise.
26820 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
26822 (lshrdi3): Use shift6_operand.
26823 (shrpsi4, shrpdi4): New insn patterns.
26824 (extzv): Delete expander.
26825 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
26826 predicates in unamed zero extract patterns. Tighten common constraint.
26827 (extv): Delete expander.
26828 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
26829 predicates in unamed sign extract patterns. Tighten common constraint.
26830 (insv): Delete expander.
26831 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
26832 predicates in unamed insert patterns. Tighten common constraint.
26833 Change uint32_operand predicate to uint6_operand predicate in unamed
26834 DImode pattern to insert constant values of type 1...1xxxx.
26836 2013-01-04 Jan Hubicka <jh@suse.cz>
26838 PR tree-optimization/55823
26839 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
26842 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26843 Uros Bizjak <ubizjak@gmail.com>
26845 PR rtl-optimization/55845
26846 * df-problems.c (can_move_insns_across): Stop scanning at
26847 volatile_insn_p source instruction or give up if
26848 across_from .. across_to range contains any volatile_insn_p
26851 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26853 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
26854 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
26856 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
26857 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
26859 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26862 * asan.c (asan_clear_shadow): New function.
26863 (asan_emit_stack_protection): Use it.
26865 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26867 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
26868 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
26869 with tab instead of space.
26871 2013-01-08 Nick Clifton <nickc@redhat.com>
26873 * config/rl78/rl78.c (rl78_expand_prologue): Always select
26874 register bank 0 at the start of an interrupt handler.
26875 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
26878 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26880 * config/aarch64/aarch64-simd.md
26881 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
26882 (aarch64_simd_bsl): Likewise.
26883 (aarch64_vcond_internal<mode>): Likewise.
26884 (vcond<mode><mode>): Likewise.
26885 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
26886 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
26888 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26890 * config/aarch64/aarch64-builtins.c
26891 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
26893 2013-01-08 Martin Jambor <mjambor@suse.cz>
26896 * tree-sra.c (analyze_access_subtree): Return true also after
26897 potentially creating a debug-only replacement.
26899 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26901 PR middle-end/55890
26902 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
26904 PR tree-optimization/54120
26905 * tree-vrp.c (range_fits_type_p): Don't allow
26906 src_precision < precision from signed vr to unsigned_p
26907 if vr->min or vr->max is negative.
26908 (simplify_float_conversion_using_ranges): Test can_float_p
26909 against CODE_FOR_nothing.
26911 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26912 Richard Biener <rguenther@suse.de>
26914 PR middle-end/55851
26915 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
26916 types instead of just INTEGER_TYPE types.
26918 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
26920 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
26923 2013-01-07 Steve Ellcey <sellcey@mips.com>
26926 * config/mips/mips.opt: Change mad to mmad to match documentation.
26928 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26931 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
26932 .progmemx.data now.
26934 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26937 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
26938 (avr_addrspace_t): Add .section_name field.
26939 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
26941 (avr_addrspace): Same. Initialize .section_name. Remove last
26942 NULL entry. Put __memx into .progmemx.data.
26943 (progmem_section_prefix): Remove.
26944 (avr_asm_init_sections): No need to initialize progmem_section.
26945 (avr_asm_named_section): Use avr_addrspace[].section_name to get
26946 section name prefix.
26947 (avr_asm_select_section): Ditto. And use get_unnamed_section to
26948 retrieve the progmem section.
26949 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
26950 boundary to run over avr_addrspace[].
26951 (avr_register_target_pragmas): Ditto.
26953 2013-01-06 Jakub Jelinek <jakub@redhat.com>
26955 * varasm.c (output_constant_def_contents): For asan_protect_global
26956 protected strings, adjust DECL_ALIGN if needed, before testing for
26958 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
26959 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
26961 (output_object_block): For asan protected decls, emit asan padding
26962 after their contents.
26963 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
26964 (asan_finish_file): Test it here instead.
26966 2013-01-07 Nick Clifton <nickc@redhat.com>
26967 Matthias Klose <doko@debian.org>
26968 Doug Kwan <dougkwan@google.com>
26969 H.J. Lu <hongjiu.lu@intel.com>
26972 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
26974 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
26976 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
26978 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
26980 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
26982 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26985 * doc/install.texi (Cross-Compiler-Specific Options): Document
26988 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
26990 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
26991 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
26992 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
26993 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
26994 vqmovun_high_s64): Fix source operand number and update copyright.
26996 2013-01-07 Richard Biener <rguenther@suse.de>
26998 PR middle-end/55890
26999 * gimple.h (gimple_call_builtin_p): New overload.
27000 * gimple.c (validate_call): New function.
27001 (gimple_call_builtin_p): Likewise.
27002 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
27003 Use gimple_call_builtin_p.
27004 (find_func_clobbers): Likewise.
27005 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
27006 (strlen_optimize_stmt): Likewise.
27008 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
27010 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
27011 (vld1q_dup_*): Likewise.
27012 (vld1_*): Likewise.
27013 (vld1q_*): Likewise.
27014 (vld1_lane_*): Likewise.
27015 (vld1q_lane_*): Likewise.
27017 2013-01-07 Richard Biener <rguenther@suse.de>
27019 * lto-streamer.h (LTO_minor_version): Bump to 2.
27021 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
27023 * config/aarch64/aarch64-protos.h
27024 (aarch64_const_double_zero_rtx_p): Rename to...
27025 (aarch64_float_const_zero_rtx_p): ...this.
27026 (aarch64_float_const_representable_p): New.
27027 (aarch64_output_simd_mov_immediate): Likewise.
27028 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
27029 move immediate case.
27030 * config/aarch64/aarch64.c
27031 (aarch64_const_double_zero_rtx_p): Rename to...
27032 (aarch64_float_const_zero_rtx_p): ...this.
27033 (aarch64_print_operand): Allow printing of new constants.
27034 (aarch64_valid_floating_const): New.
27035 (aarch64_legitimate_constant_p): Check for valid floating-point
27037 (aarch64_simd_valid_immediate): Likewise.
27038 (aarch64_vect_float_const_representable_p): New.
27039 (aarch64_float_const_representable_p): Likewise.
27040 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
27041 (aarch64_output_simd_mov_immediate): New.
27042 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
27043 (*movdf_aarch64): Likewise.
27044 * config/aarch64/constraints.md (Ufc): New.
27045 (Y): call aarch64_float_const_zero_rtx.
27046 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
27048 2013-01-07 Richard Biener <rguenther@suse.de>
27050 PR tree-optimization/55888
27051 PR tree-optimization/55862
27052 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
27053 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
27054 not if it is contained therein.
27056 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27058 * config/avr/t-avr: Typo.
27060 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27063 * config/avr/t-avr: Don't automatically rebuild
27064 $(srcdir)/config/avr/t-multilib
27065 $(srcdir)/config/avr/avr-tables.opt
27066 $(srcdir)/doc/avr-mmcu.texi
27067 (avr-mcus): New phony target to build them on request.
27068 (s-avr-mlib, s-avr-mmcu-texi): Remove.
27069 * avr/avr-mcus.def: Adjust comments.
27071 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
27073 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
27075 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
27077 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
27079 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
27081 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
27083 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
27085 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
27086 to generate profiling.
27087 * config/rs6000/aix64.h (LIB_SPEC): Same.
27089 2013-01-04 Andrew Pinski <apinski@cavium.com>
27091 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
27093 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
27095 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
27097 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
27099 (ix86_expand_move): Ditto.
27100 (ix86_zero_extend_to_Pmode): Ditto.
27101 (ix86_expand_call): Ditto.
27102 (ix86_expand_special_args_builtin): Ditto.
27103 (ix86_expand_builtin): Ditto.
27105 2013-01-04 Richard Biener <rguenther@suse.de>
27107 PR tree-optimization/55862
27108 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
27109 translating them through PHI nodes.
27111 2013-01-04 Martin Jambor <mjambor@suse.cz>
27113 PR tree-optimization/55755
27114 * tree-sra.c (sra_modify_assign): Do not check that an access has no
27115 children when trying to avoid producing a VIEW_CONVERT_EXPR.
27117 2013-01-04 Marek Polacek <polacek@redhat.com>
27119 PR middle-end/55859
27120 * opts.c (default_options_optimization): Clarify error message.
27122 2013-01-04 Richard Biener <rguenther@suse.de>
27124 PR middle-end/55863
27125 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
27128 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27131 * config/pa/pa.md (movsi): Revert previous change.
27132 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
27135 2013-01-03 Richard Henderson <rth@redhat.com>
27137 * config/i386/i386.c (ix86_expand_move): Always assign to op1
27138 after eliminating TLS symbols.
27140 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
27143 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
27144 * graphite-poly.c (debug_gmp_value): Likewise.
27146 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
27149 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
27150 selected code model, define __code_mode_small__, __code_model_medium__,
27151 __code_model_large__, __code_model_32__ or __code_model_kernel__.
27152 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
27153 xchg temporary register with %k. Declare temporary register as
27155 [__x86_64__]: For medium and large code models, preserve %rbx register.
27157 2013-01-03 Richard Biener <rguenther@suse.de>
27159 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
27160 (dump_subscript): Adjust.
27161 (finalize_ddr_dependent): Do not dump redundant info.
27162 (analyze_siv_subscript): Adjust.
27163 (subscript_dependence_tester): Likewise.
27164 (compute_affine_dependence): Likewise.
27166 2013-01-03 Richard Biener <rguenther@suse.de>
27169 2013-01-03 Richard Biener <rguenther@suse.de>
27171 PR tree-optimization/55857
27172 * tree-vect-stmts.c (vectorizable_load): Do not setup
27173 re-alignment for invariant loads.
27175 2013-01-02 Richard Biener <rguenther@suse.de>
27177 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
27178 invariant load do not generate a vector load from the scalar location.
27180 2013-01-03 Richard Biener <rguenther@suse.de>
27182 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
27183 for not vectorizing.
27184 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
27185 not build INDIRECT_REFs, call get_name once only.
27186 (vect_create_data_ref_ptr): Likewise. Dump base object kind
27187 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
27189 2013-01-03 Richard Biener <rguenther@suse.de>
27191 PR tree-optimization/55857
27192 * tree-vect-stmts.c (vectorizable_load): Do not setup
27193 re-alignment for invariant loads.
27195 2013-01-03 Richard Biener <rguenther@suse.de>
27198 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
27199 prefer a built-in decl.
27201 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27203 * gcc.c (process_command): Update copyright notice dates.
27204 * gcov.c (print_version): Likewise.
27205 * gcov-dump.c (print_version): Likewise.
27207 PR rtl-optimization/55838
27208 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
27209 iv0.step, iv1.step and step.
27211 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27212 Marc Glisse <marc.glisse@inria.fr>
27214 PR tree-optimization/55832
27215 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
27216 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
27217 integer_{one,zero}_node.
27219 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27222 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
27223 * var-tracking.c (reverse_op): Don't add reverse ops to
27224 VALUEs that have already
27225 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
27227 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
27229 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
27231 2013-01-02 Teresa Johnson <tejohnson@google.com>
27233 * dumpfile.c (dump_loc): Print filename with location.
27234 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
27235 new location_t parameter to emit complete unroll message with
27236 new dump framework.
27237 (canonicalize_loop_induction_variables): Compute loops location
27238 and pass to try_unroll_loop_completely.
27239 * loop-unroll.c (report_unroll_peel): New function.
27240 (peel_loops_completely): Use new dump format with location
27241 for main dumpfile message, and invoke report_unroll_peel on success.
27242 (decide_unrolling_and_peeling): Ditto.
27243 (decide_peel_once_rolling): Remove old dumpfile message subsumed
27244 by report_unroll_peel.
27245 (decide_peel_completely): Ditto.
27246 (decide_unroll_constant_iterations): Ditto.
27247 (decide_unroll_runtime_iterations): Ditto.
27248 (decide_peel_simple): Ditto.
27249 (decide_unroll_stupid): Ditto.
27250 * cfgloop.c (get_loop_location): New function.
27251 * cfgloop.h (get_loop_location): Declare.
27253 2013-01-02 Sriraman Tallam <tmsriram@google.com>
27255 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
27258 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27260 PR middle-end/55198
27261 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
27262 BLKmode objects when EXPAND_MEMORY is specified.
27264 2013-01-02 Sriraman Tallam <tmsriram@google.com>
27266 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
27268 (fold_builtin_cpu): Do not share cpu model decls across statements.
27270 2013-01-02 Jason Merrill <jason@redhat.com>
27273 * tree.c (build_array_type_1): Revert earlier change.
27275 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
27277 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
27279 * config/aarch64/aarch64-tune.md: Re-generate.
27281 2013-01-02 Richard Biener <rguenther@suse.de>
27283 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
27284 invariant load do not generate a vector load from the scalar location.
27286 2013-01-02 Richard Biener <rguenther@suse.de>
27289 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
27290 * configure: Regenerate.
27292 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27294 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
27295 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
27296 (expand_builtin_int_roundingfn_2): Keep the original target around
27297 for the fallback case.
27299 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27301 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
27302 to be clear for sign changes.
27304 2013-01-01 Jan Hubicka <jh@suse.cz>
27306 * ipa-inline-analysis.c: Fix formatting.
27308 2013-01-01 Jakub Jelinek <jakub@redhat.com>
27310 PR tree-optimization/55831
27311 * tree-vect-loop.c (get_initial_def_for_induction): Use
27312 gsi_after_labels instead of gsi_start_bb.
27314 Copyright (C) 2013 Free Software Foundation, Inc.
27316 Copying and distribution of this file, with or without modification,
27317 are permitted in any medium without royalty provided the copyright
27318 notice and this notice are preserved.