1 2013-10-21 Marek Polacek <polacek@redhat.com>
4 * fold-const.c (fold_range_test): Return 0 if the type is not
7 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
9 * system.h: Move hwint.h include further down.
10 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
12 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
14 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
16 Fix volatile issues in optimize_bit_field_compare.
17 * fold-const.c (optimize_bit_field_compare): Bail out if
18 lvolatilep or rvolatilep.
20 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
22 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
23 and get_inner_reference returning different pmode for non-volatile
24 bit-field members dependent on flag_strict_volatile_bitfields.
25 * stor-layout.c (layout_decl): Remove special handling of
26 flag_strict_volatile_bitfields.
27 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
28 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
30 2013-10-21 Paulo Matos <pmatos@broadcom.com>
32 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
35 2013-10-21 Jeff Law <law@redhat.com>
37 * tree-ssa-threadedge.c (thread_through_normal_block): New argument VISITED.
38 Remove VISISTED as a local variable. When we have a threadable jump, verify
39 the destination of the jump has not been visised.
40 (thread_across_edge): Allocate VISITED bitmap once at function scope and
41 use it throughout. Make sure to set appropriate bits in VISITED for E (start
44 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading through
45 a joiner if any edge on the path has a recorded jump thread.
47 2013-10-21 Ian Lance Taylor <iant@google.com>
49 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
50 don't imply that attributes can solve all problems.
51 (Directory Options): Fix typo.
53 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
55 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
56 extend and extend_add.
58 2013-10-21 Richard Biener <rguenther@suse.de>
60 PR tree-optimization/58794
61 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
62 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
64 2013-10-21 Richard Biener <rguenther@suse.de>
67 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
68 to (T) X for sign-changing conversions (or no conversion).
70 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
72 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
74 2013-10-20 Jan Hubicka <jh@suse.cz>
76 * config/i386/i386-tune.def: Add comment; organize into categories
78 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
80 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
81 argument. Update function comment.
82 (expand_set_or_movmem_via_rep): New function combining
83 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
84 (expand_movmem_via_rep_mov): Remove.
85 expand_setmem_via_rep_stos): Remove.
86 (expand_movmem_epilogue): Update calls correspondingly.
87 (expand_setmem_epilogue_via_loop): Likewise.
89 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
90 (expand_set_or_movmem_prologue): New function combining
91 expand_movmem_prologue and expand_setmem_prologue.
92 (expand_movmem_prologue): Remove.
93 (expand_setmem_prologue): Remove.
94 (expand_set_or_movmem_constant_prologue): New function combining
95 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
96 (expand_constant_movmem_prologue): Remove.
97 (expand_constant_setmem_prologue): Remove.
98 (promote_duplicated_reg): Allow vector-const0 value.
99 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
100 and ix86_expand_setmem.
101 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
102 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
104 2013-10-21 Diego Novillo <dnovillo@google.com>
106 * asan.c: Include tree.h
107 * bb-reorder.c: Likewise.
108 * cfgcleanup.c: Likewise.
109 * cfgloopmanip.c: Likewise.
110 * data-streamer-in.c: Likewise.
111 * data-streamer-out.c: Likewise.
112 * data-streamer.c: Likewise.
113 * dwarf2cfi.c: Likewise.
114 * graphite-blocking.c: Likewise.
115 * graphite-clast-to-gimple.c: Likewise.
116 * graphite-dependences.c: Likewise.
117 * graphite-interchange.c: Likewise.
118 * graphite-optimize-isl.c: Likewise.
119 * graphite-poly.c: Likewise.
120 * graphite-scop-detection.c: Likewise.
121 * graphite-sese-to-poly.c: Likewise.
122 * graphite.c: Likewise.
123 * ipa-devirt.c: Likewise.
124 * ipa-profile.c: Likewise.
127 * loop-init.c: Likewise.
128 * loop-unroll.c: Likewise.
129 * lower-subreg.c: Likewise.
130 * lto/lto-object.c: Likewise.
132 * reginfo.c: Likewise.
133 * tree-loop-distribution.c: Likewise.
134 * tree-parloops.c: Likewise.
135 * tree-ssa-strlen.c: Likewise.
136 * tree-streamer.c: Likewise.
137 * value-prof.c: Likewise.
138 * target-globals.c: Likewise.
139 * expr.h: Include tree-core.h instead of tree.h.
140 * gimple.h: Likewise.
141 * ipa-prop.h: Likewise.
142 * ipa-utils.h: Likewise.
143 * lto-streamer.h: Likewise.
144 * streamer-hooks.h: Likewise.
145 * ipa-reference.h: Include cgraph.h instead of tree.h.
146 * cgraph.h: Include basic-block.h instead of tree.h.
147 * tree-streamer.h: Do not include tree.h.
148 * genattrtab.c (write_header): Generate inclusion of
150 * genautomata.c (main): Likewise.
151 * genemit.c: Likewise.
152 * genopinit.c: Likewise.
153 * genoutput.c (output_prologue): Likewise.
154 * genpeep.c: Likewise.
156 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
158 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
160 (vec_unpacku_hi_v8hi): Likewise.
161 (vec_unpacku_lo_v16qi): Likewise.
162 (vec_unpacku_lo_v8hi): Likewise.
164 2013-10-20 Jan Hubicka <jh@suse.cz>
166 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
167 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
168 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
170 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
172 * config/mips/mips.h (ISA_HAS_WSBH): Define.
173 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
175 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
177 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
180 * system.h: Undef m_slot.
182 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
184 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
185 all elements for both endian flavors.
187 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
190 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
191 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
192 and SI_REG for 64bit SYSV ABI targets.
194 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
196 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
197 to multi_reg_return. Clarify that we are skipping USEs of multiple
198 return registers. Use bool type where appropriate.
200 2013-10-18 Jan Hubicka <jh@suse.cz>
202 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
204 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
205 (X86_TUNE_PUSH_MEMORY): Likewise.
206 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
207 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
208 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
210 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
211 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
213 (ix86_option_override_internal): Update to use tune features instead
216 2013-10-18 Cong Hou <congh@google.com>
218 PR tree-optimization/58508
219 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
220 statement that contains data refs with zero-step.
222 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
224 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
225 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
226 * gimple-low.c (gimple_check_call_arg,
227 gimple_check_call_matching_types): Move to cgraph.c.
228 * gimple-low.h: Remove prototype.
229 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
230 Relocate from gimple-low.c.
231 * cgraph.h: Add prototype. Don't include basic-block.h.
232 * gimplify.c: Add gimple-low to include list.
233 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
234 * tree-eh.c: Add gimple-low to include list.
235 * tree-nested.c: Likewise.
236 * cfgexpand.c: Add tree-ssa-address.h to include list.
238 * gimple-fold.c: Likewise.
239 * gimple-ssa-strength-reduction.c: Likewise.
240 * trans-mem.c: Likewise.
241 * tree-mudflap.c: Likewise.
242 * tree-ssa-loop-ivopts.c: Likewise.
243 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
244 (degenerate_phi_result): Move to tree-phinodes.c.
245 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
246 * tree-ssa-threadedge.c: Likewise.
247 * tree-vrp.c: Likewise.
248 * tree-phinodes.c (degenerate_phi_result): Relocate here.
249 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
250 * tree-phinodes.h (degenerate_phi_result): Add prototype.
251 * tree-ssa-copy.c: Include tree-ssa-dom.h.
252 * tree-ssa-forwprop.c: Likewise.
253 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
254 pass_data_cleanup_cfg_post_optimizing,
255 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
256 * tree-optimize.c: Delete File.
257 * graphite.c: Include tree-cfgcleanup.h.
258 * passes.c: Likewise.
259 * tree-cfg.c: Likewise.
260 * tree-profile.c: Likewise.
261 * tree-ssa-dse.c: Likewise.
262 * tree-ssa-loop-ivcanon.c: Likewise.
263 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
264 * tree-outof-ssa.c: Include sbitmap.h.
265 * tree-ssa-live.c: Likewise.
266 * tree-ssa-propagate.c: Likewise.
267 * tree-ssa-structalias.c: Likewise.
268 * tree-stdarg.c: Likewise.
269 * Makefile.in (OBJS): Delete tree-optimize.o.
270 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
271 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
272 * varasm.c: Include basic-block.h.
273 * cfgloop.h: Include function.h instead of basic-block.h
274 (bb_loop_depth): Move to cfgloop.c.
275 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
277 2013-10-18 Teresa Johnson <tejohnson@google.com>
279 * predict.c (probably_never_executed): Compare frequency-based
280 count to number of training runs.
281 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
283 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
285 * config/arm/arm.c (cortexa9_extra_costs): New table.
286 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
288 2013-10-18 Jeff Law <law@redhat.com>
290 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
292 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
293 (register_jump_thread): Add "registered_jump_thread" debug
295 * dbgcnt.def (registered_jump_thread): New debug counter.
297 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
299 * config/rs6000/rs6000.c: Include cgraph.h.
301 2013-10-18 Teresa Johnson <tejohnson@google.com>
303 * tree-ssa-tail-merge.c (replace_block_by): Update edge
304 weights during merging.
306 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
308 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
309 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
310 * tree-ssa-operands.h: Remove prototype.
311 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
312 * gimple.c (virtual_operand_p): Relocate from gimple.c.
313 * gimple.h: Add prototype.
314 * gimple-ssa.h: Include tree-ssa-operands.h.
315 * tree-dump.c: Add tree-cfg.h to include list.
316 * tree-ssa-alias.c: Add ipa-reference.h to include list.
317 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
318 * config/i386/i386.c: Don't include tree-flow.h.
319 * config/rs6000/rs6000.c: Likewise.
321 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
323 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
325 2013-10-18 Richard Biener <rguenther@suse.de>
327 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
328 or TYPE_UNSIGNED of integral types.
329 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
330 NULL_TREE for zero-precision integral types.
332 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
334 * config/aarch64/arm_neon.h
335 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
337 2013-10-17 Sriraman Tallam <tmsriram@google.com>
339 * opth-gen.awk: Define target_flags_explicit.
341 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
343 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
344 fields to the reg_addr array that describes the valid addressing
345 mode for any register, general purpose registers, floating point
346 registers, and Altivec registers.
347 (FIRST_RELOAD_REG_CLASS): Likewise.
348 (LAST_RELOAD_REG_CLASS): Likewise.
349 (struct reload_reg_map_type): Likewise.
350 (reload_reg_map_type): Likewise.
351 (RELOAD_REG_VALID): Likewise.
352 (RELOAD_REG_MULTIPLE): Likewise.
353 (RELOAD_REG_INDEXED): Likewise.
354 (RELOAD_REG_OFFSET): Likewise.
355 (RELOAD_REG_PRE_INCDEC): Likewise.
356 (RELOAD_REG_PRE_MODIFY): Likewise.
357 (reg_addr): Likewise.
358 (mode_supports_pre_incdec_p): New helper functions to say whether
359 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
360 (mode_supports_pre_modify_p): Likewise.
361 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
362 print the valid address mode bits for each mode.
363 (rs6000_debug_print_mode): Likewise.
364 (rs6000_debug_reg_global): Likewise.
365 (rs6000_setup_reg_addr_masks): New function to set up the address
366 mask bits for each type.
367 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
368 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
369 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
370 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
371 PRE_MODIFY are supported.
372 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
373 registers, instead of {src,dest}_av_p.
374 (rs6000_print_options_internal): Tweak the debug output slightly.
376 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
378 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
381 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
383 * tree-flow.h (struct omp_region): Move to omp-low.c.
384 Remove omp_ prototypes and variables.
385 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
386 (copy_var_decl): Relocate prototype from tree-flow.h.
387 * gimple.c (copy_var_decl): Relocate from omp-low.c.
388 * tree.h: Move prototype to omp-low.h.
389 * omp-low.h: New File. Relocate prototypes here.
390 * omp-low.c (struct omp_region): Make local here.
391 (root_omp_region): Make static.
392 (copy_var_decl) Move to gimple.c.
393 (new_omp_region): Make static.
394 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
395 * tree-cfg.c: Include omp-low.h.
396 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
397 * gimplify.c: Include omp-low.h.
398 * tree-parloops.c: Likewise.
400 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
402 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
403 operands are in memory, prefer to force non-matched operand 1 to
406 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
409 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
410 restrict TImode addresses to single indirect registers if both
411 -mquad-memory and -mvsx-timode are used.
412 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
413 we should emit load/store quad. Remove using %y for quad memory
416 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
417 constraints to allow load/store quad on machines where TImode is
418 not allowed in VSX registers. Use 'n' instead of 'F' constraint
419 for TImode to load integer constants.
421 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
423 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
425 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
427 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
428 handling of STACK_REG.
430 2013-10-17 Richard Biener <rguenther@suse.de>
432 PR tree-optimization/58143
433 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
435 (rewrite_to_defined_overflow): Likewise.
436 (move_computations_dom_walker::before_dom): Rewrite stmts
437 with undefined signed overflow that are not always executed
438 into unsigned arithmetic.
440 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
443 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
444 explicit isa flag to be an options variable, instead of using
445 global_options_set. Remove define from rs6000.h.
446 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
448 * config/rs6000/rs6000.c (rs6000_option_override_internal):
449 Initialize rs6000_isa_flags_explicit.
450 (rs6000_function_specific_save): Add gcc_options* parameter, so
451 that the powerpc builds after the 2013-10-15 changes.
452 (rs6000_function_specific_restore): Likewise.
454 2013-10-16 DJ Delorie <dj@redhat.com>
456 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
457 op is a REG before checking REGNO.
458 (rl78_alloc_physical_registers): Verify pattern is a SET before
461 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
463 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
465 (vec_unpacks_lo_v4sf): Likewise.
466 (vec_unpacks_float_hi_v4si): Likewise.
467 (vec_unpacks_float_lo_v4si): Likewise.
468 (vec_unpacku_float_hi_v4si): Likewise.
469 (vec_unpacku_float_lo_v4si): Likewise.
471 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
473 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
474 (vsx_concat_v2sf): Likewise.
476 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
478 * config/aarch64/aarch64.md
479 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
481 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
483 PR tree-optimization/58697
484 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
485 estimated_loop_iterations_int.
486 (max_stmt_executions_int): Call get_max_loop_iterations_int.
487 (get_max_loop_iterations_int): New. HWINT version of
488 get_max_loop_iterations.
489 * cfgloop.h: Add prototypes.
490 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
491 * loop-unroll.c (decide_peel_once_rolling): Call
492 get_estimated_loop_iterations_int.
493 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
494 * tree-ssa-loop-niter.h: Tweak prototypes.
496 2013-10-16 David Malcolm <dmalcolm@redhat.com>
498 * gengtype-parse.c (struct_field_seq): Ignore access-control
499 keywords ("public:" etc).
501 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
503 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
504 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
506 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
508 * config/arm/arm.opt (mlra): New option.
509 * config/arm/arm.c (arm_lra_p): New function.
510 (TARGET_LRA_P): Define.
512 2013-10-16 Paulo Matos <pmatos@broadcom.com>
514 * tree-core.h (tree_code_name): Remove.
515 * tree.h (get_tree_code_name): New prototype.
516 * tree.c (tree_code_name): Make static.
517 (get_tree_code_name): New function.
518 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
519 tree_class_check_failed, tree_range_check_failed,
520 tree_not_class_check_failed, omp_clause_check_failed,
521 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
522 wrapper get_tree_code_name instead of calling tree_code_name directly.
523 * tree-vrp.c (dump_asserts_for): Likewise.
524 * tree-dump.c (dequeue_and_dump): Likewise.
525 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
526 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
527 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
528 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
529 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
530 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
531 dump_gimple_omp_for): Likewise.
532 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
533 * tree-ssa-pre.c (print_pre_expr): Likewise.
534 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
535 * print-tree.c (print_node_brief, print_node): Likewise.
536 * gimple.c (gimple_check_failed): Likewise.
537 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
538 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
539 * config/mep/mep.c (mep_validate_vliw): Likewise.
540 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
541 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
543 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
545 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
548 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
550 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
551 (MULTILIB_MATCHES): Add multilib for -march=v8.
553 2013-10-15 Sriraman Tallam <tmsriram@google.com>
556 * optc-save-gen.awk: Add extra parameter to the save and restore
558 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
559 * tree.c (build_optimization_node): New parameter. Add extra parameter
560 to call to cl_optimization_save.
561 (build_target_option_node): New parameter. Add extra parameter
562 to call to cl_target_option_save.
563 * tree.h (build_optimization_node): New parameter.
564 (build_target_option_node): New parameter.
565 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
566 build_optimization_node and build_target_option_node.
567 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
568 (handle_pragma_push_options): Ditto.
569 * toplev.c (process_options): Ditto.
570 * opts.c (init_options_struct): Check for opts_set non-null.
571 * target.def (target_option.save): New parameter.
572 (target_option.restore): New parameter.
574 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
575 (ix86_pragma_target_parse): Ditto.
576 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
578 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
579 to build_optimization_node and build_target_option_node.
580 (rs6000_valid_attribute_p): Ditto.
581 (rs6000_pragma_target_parse): Ditto.
582 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
584 * config/i386/i386.h:
585 TARGET_64BIT_P: New Macro
586 TARGET_MMX_P: New Macro.
587 TARGET_3DNOW_P: New Macro.
588 TARGET_3DNOW_A_P: New Macro.
589 TARGET_SSE_P: New Macro.
590 TARGET_SSE2_P: New Macro.
591 TARGET_SSE3_P: New Macro.
592 TARGET_SSSE3_P: New Macro.
593 TARGET_SSE4_1_P: New Macro.
594 TARGET_SSE4_2_P: New Macro.
595 TARGET_AVX_P: New Macro.
596 TARGET_AVX2_P: New Macro.
597 TARGET_AVX512F_P: New Macro.
598 TARGET_AVX512PF_P: New Macro.
599 TARGET_AVX512ER_P: New Macro.
600 TARGET_AVX512CD_P: New Macro.
601 TARGET_FMA_P: New Macro.
602 TARGET_SSE4A_P: New Macro.
603 TARGET_FMA4_P: New Macro.
604 TARGET_XOP_P: New Macro.
605 TARGET_LWP_P: New Macro.
606 TARGET_ABM_P: New Macro.
607 TARGET_BMI_P: New Macro.
608 TARGET_BMI2_P: New Macro.
609 TARGET_LZCNT_P: New Macro.
610 TARGET_TBM_P: New Macro.
611 TARGET_POPCNT_P: New Macro.
612 TARGET_SAHF_P: New Macro.
613 TARGET_MOVBE_P: New Macro.
614 TARGET_CRC32_P: New Macro.
615 TARGET_AES_P: New Macro.
616 TARGET_PCLMUL_P: New Macro.
617 TARGET_CMPXCHG16B_P: New Macro.
618 TARGET_FSGSBASE_P: New Macro.
619 TARGET_RDRND_P: New Macro.
620 TARGET_F16C_P: New Macro.
621 TARGET_RTM_P: New Macro.
622 TARGET_HLE_P: New Macro.
623 TARGET_RDSEED_P: New Macro.
624 TARGET_PRFCHW_P: New Macro.
625 TARGET_ADX_P: New Macro.
626 TARGET_FXSR_P: New Macro.
627 TARGET_XSAVE_P: New Macro.
628 TARGET_XSAVEOPT_P: New Macro.
629 TARGET_LP64_P: New Macro.
630 TARGET_X32_P: New Macro.
631 TARGET_FPMATH_DEFAULT_P: New Macro.
632 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
633 * config/i386/i386.c (ix86_option_override_internal): New parameters.
635 Change ix86_tune_string to access opts->x_ix86_tune_string.
636 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
637 Change ix86_arch_string to access opts->x_ix86_arch_string.
638 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
639 Change ix86_pmode to access opts->x_ix86_pmode.
640 Change ix86_abi to access opts->x_ix86_abi.
641 Change ix86_cmodel to access opts->x_ix86_cmodel.
642 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
643 Change ix86_isa_flags_explicit to access
644 opts->x_ix86_isa_flags_explicit.
645 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
646 Change ix86_regparm to access opts->x_ix86_regparm.
647 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
648 Change ix86_preferred_stack_boundary_arg to access
649 opts->x_ix86_preferred_stack_boundary_arg.
650 Change ix86_force_align_arg_pointer to access
651 opts->x_ix86_force_align_arg_pointer.
652 Change ix86_incoming_stack_boundar_arg to access
653 opts->x_ix86_incoming_stack_boundar_arg.
654 Change ix86_fpmath to access opts->x_ix86_fpmath.
655 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
656 Change ix86_recip_name to access opts->x_ix86_recip_name.
657 Change ix86_stack_protector_guard to access
658 opts->x_ix86_stack_protector_guard.
659 Change ix86_tune_memcpy_strategy to access
660 opts->x_ix86_tune_memcpy_strategy.
661 Change ix86_tune_memset_strategy to access
662 opts->x_ix86_tune_memset_strategy.
663 Change global_options to access opts.
664 Change global_options_set to access opts_set.
665 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
666 Change TARGET_MMX to TARGET_MMX_P (opts->...).
667 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
668 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
669 Change TARGET_SSE to TARGET_SSE_P (opts->...).
670 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
671 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
672 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
673 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
674 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
675 Change TARGET_AVX to TARGET_AVX_P (opts->...).
676 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
677 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
678 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
679 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
680 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
681 Change TARGET_FMA to TARGET_FMA_P (opts->...).
682 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
683 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
684 Change TARGET_XOP to TARGET_XOP_P (opts->...).
685 Change TARGET_LWP to TARGET_LWP_P (opts->...).
686 Change TARGET_ABM to TARGET_ABM_P (opts->...).
687 Change TARGET_BMI to TARGET_BMI_P (opts->...).
688 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
689 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
690 Change TARGET_TBM to TARGET_TBM_P (opts->...).
691 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
692 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
693 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
694 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
695 Change TARGET_AES to TARGET_AES_P (opts->...).
696 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
697 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
698 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
699 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
700 Change TARGET_F16C to TARGET_F16C_P (opts->...).
701 Change TARGET_RTM to TARGET_RTM_P (opts->...).
702 Change TARGET_HLE to TARGET_HLE_P (opts->...).
703 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
704 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
705 Change TARGET_ADX to TARGET_ADX_P (opts->...).
706 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
707 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
708 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
709 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
710 Change TARGET_X32 to TARGET_X32_P (opts->...).
711 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
712 Change TARGET_FLOAT_RETURNS_IN_80387 to
713 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
714 (ix86_function_specific_save): New parameter. Use opts-> fields
715 to replace global fields.
716 (ix86_function_specific_restore): Ditto.
717 (ix86_valid_target_attribute_inner_p): New parameters.
719 Fix call to ix86_handle_option and set_option.
720 (ix86_valid_target_attribute_tree): New parameters.
721 Change global_options to access opts.
722 Change global_options_set to access opts_set.
723 Fix call to ix86_valid_target_attribute_inner_p.
724 Change ix86_tune_string to access opts->x_ix86_tune_string.
725 Change ix86_arch_string to access opts->x_ix86_arch_string.
726 Change ix86_fpmath to access opts->x_ix86_fpmath
727 Fix call to ix86_option_override_internal.
728 Fix call to ix86_add_new_builtins.
729 Fix calls to build_optimization_node and build_target_option_node.
730 (ix86_valid_target_attribute_p): Remove access to global_options.
731 Use new gcc_options structure func_options.
732 Fix call to ix86_valid_target_attribute_tree.
733 Fix call to build_optimization_node.
734 (get_builtin_code_for_version): Fix call to
735 ix86_valid_target_attribute_tree.
737 2013-10-15 David Malcolm <dmalcolm@redhat.com>
739 * Makefile.in (PICFLAG): New.
740 (enable_host_shared): New.
741 (INTERNAL_CFLAGS): Use PICFLAG.
742 (LIBIBERTY): Use pic build of libiberty.a if configured with
743 --enable-host-shared.
744 * configure.ac: Add --enable-host-shared, setting up new
746 * configure: Regenerate.
747 * doc/install.texi (--enable-shared): Add note contrasting it with ...
748 (--enable-host-shared): New option.
750 2013-10-15 Richard Biener <rguenther@suse.de>
752 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
753 for built-in functions.
755 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
757 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
758 (optimize_range_tests_1): New function,
759 extracted from optimize_range_tests.
760 (optimize_range_tests_xor): Similarly.
761 (optimize_range_tests_diff): New function.
762 (optimize_range_tests): Use optimize_range_tests_1.
764 2013-10-15 Cong Hou <congh@google.com>
766 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
767 requirement of the reduction pattern so that one operand of the
768 reduction operation can come from outside of the loop.
770 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
772 * config/arm/neon-schedgen.ml: Remove.
773 * config/arm/cortex-a9-neon.md: Remove comment regarding
776 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
778 * config/arm/types: Remove old neon types.
780 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
782 * config/arm/cortex-a7.md
783 (cortex_a7_neon_type): New.
784 (cortex_a7_neon_mul): Update for new types.
785 (cortex_a7_neon_mla): Likewise.
786 (cortex_a7_neon): Likewise.
788 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
790 * config/arm/cortex-a15-neon.md
791 (cortex_a15_neon_type): New,
793 (cortex_a15_neon_int_1): Remove.
794 (cortex_a15_neon_int_2): Likewise.
795 (cortex_a15_neon_int_3): Likewise.
796 (cortex_a15_neon_int_4): Likewise.
797 (cortex_a15_neon_int_5): Likewise.
798 (cortex_a15_neon_vqneg_vqabs): Likewise.
799 (cortex_a15_neon_vmov): Likewise.
800 (cortex_a15_neon_vaba): Likewise.
801 (cortex_a15_neon_vaba_qqq): Likewise.
802 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
803 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
804 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
806 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
807 (cortex_a15_neon_mla_qqq_8_16): Likewise.
808 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
809 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
810 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
811 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
812 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
813 (cortex_a15_neon_shift_1): Likewise.
814 (cortex_a15_neon_shift_2): Likewise.
815 (cortex_a15_neon_shift_3): Likewise.
816 (cortex_a15_neon_vshl_ddd): Likewise.
817 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
818 (cortex_a15_neon_vsra_vrsra): Likewise.
819 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
820 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
821 (cortex_a15_neon_bp_3cycle): Likewise.
822 (cortex_a15_neon_ldm_2): Likewise.
823 (cortex_a15_neon_stm_2): Likewise.
824 (cortex_a15_neon_mcr): Likewise.
825 (cortex_a15_neon_mrc): Likewise.
826 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
827 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
828 (cortex_a15_neon_fp_vmul_ddd): Likewise.
829 (cortex_a15_neon_fp_vmul_qqd): Likewise.
830 (cortex_a15_neon_fp_vmla_ddd): Likewise.
831 (cortex_a15_neon_fp_vmla_qqq): Likewise.
832 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
833 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
834 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
835 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
836 (cortex_a15_neon_bp_simple): Likewise.
837 (cortex_a15_neon_bp_2cycle): Likewise.
838 (cortex_a15_neon_bp_3cycle): Likewise.
839 (cortex_a15_neon_vld1_1_2_regs): Likewise.
840 (cortex_a15_neon_vld1_3_4_regs): Likewise.
841 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
842 (cortex_a15_neon_vld2_4_regs): Likewise.
843 (cortex_a15_neon_vld3_vld4): Likewise.
844 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
845 (cortex_a15_neon_vst1_3_4_regs): Likewise.
846 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
847 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
848 (cortex_a15_neon_vst3_vst4): Rename to...
849 (cortex_a15_neon_vst4): This, update for new attributes.
850 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
851 (cortex_a15_neon_vld3_vld4_lane): Likewise.
852 (cortex_a15_neon_vst1_vst2_lane): Likewise.
853 (cortex_a15_neon_vst3_vst4_lane): Likewise.
854 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
855 (cortex_a15_neon_ldm_2): Likewise.
856 (cortex_a15_neon_stm_2): Likewise.
857 (cortex_a15_neon_mcr): Likewise.
858 (cortex_a15_neon_mcr_2_mcrr): Likewise.
859 (cortex_a15_neon_mrc): Likewise.
860 (cortex_a15_neon_mrrc): Likewise.
862 (cortex_a15_neon_abd): New.
863 (cortex_a15_neon_abd_q): Likewise.
864 (cortex_a15_neon_aba): Likewise.
865 (cortex_a15_neon_aba_q): Likewise.
866 (cortex_a15_neon_acc): Likewise.
867 (cortex_a15_neon_acc_q): Likewise.
868 (cortex_a15_neon_arith_basic): Likewise.
869 (cortex_a15_neon_arith_complex): Likewise.
870 (cortex_a15_neon_multiply): Likewise.
871 (cortex_a15_neon_multiply_q): Likewise.
872 (cortex_a15_neon_mla): Likewise.
873 (cortex_a15_neon_mla_q): Likewise.
874 (cortex_a15_neon_sat_mla_long): Likewise.
875 (cortex_a15_neon_shift_acc): Likewise.
876 (cortex_a15_neon_shift_imm_basic): Likewise.
877 (cortex_a15_neon_shift_imm_complex): Likewise.
878 (cortex_a15_neon_shift_reg_basic): Likewise.
879 (cortex_a15_neon_shift_reg_basic_q): Likewise.
880 (cortex_a15_neon_shift_reg_complex): Likewise.
881 (cortex_a15_neon_shift_reg_complex_q): Likewise.
882 (cortex_a15_neon_fp_negabs): Likewise
883 (cortex_a15_neon_fp_arith): Likewise
884 (cortex_a15_neon_fp_arith_q): Likewise
885 (cortex_a15_neon_fp_cvt_int): Likewise
886 (cortex_a15_neon_fp_cvt_int_q): Likewise
887 (cortex_a15_neon_fp_cvt_16): Likewise
888 (cortex_a15_neon_fp_mul): Likewise
889 (cortex_a15_neon_fp_mul_q): Likewise
890 (cortex_a15_neon_fp_mla): Likewise
891 (cortex_a15_neon_fp_mla_q): Likewise
892 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
893 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
894 (cortex_a15_neon_bitops): Likewise.
895 (cortex_a15_neon_bitops_q): Likewise.
896 (cortex_a15_neon_from_gp): Likewise.
897 (cortex_a15_neon_from_gp_q): Likewise.
898 (cortex_a15_neon_tbl3_tbl4): Likewise.
899 (cortex_a15_neon_zip_q): Likewise.
900 (cortex_a15_neon_to_gp): Likewise.
901 (cortex_a15_neon_load_a): Likewise.
902 (cortex_a15_neon_load_b): Likewise.
903 (cortex_a15_neon_load_c): Likewise.
904 (cortex_a15_neon_load_d): Likewise.
905 (cortex_a15_neon_load_e): Likewise.
906 (cortex_a15_neon_load_f): Likewise.
907 (cortex_a15_neon_store_a): Likewise.
908 (cortex_a15_neon_store_b): Likewise.
909 (cortex_a15_neon_store_c): Likewise.
910 (cortex_a15_neon_store_d): Likewise.
911 (cortex_a15_neon_store_e): Likewise.
912 (cortex_a15_neon_store_f): Likewise.
913 (cortex_a15_neon_store_g): Likewise.
914 (cortex_a15_neon_store_h): Likewise.
915 (cortex_a15_vfp_to_from_gp): Likewise.
917 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
919 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
921 (cortex_a9_neon_vshl_ddd): Remove.
922 (cortex_a9_neon_vst3_vst4): Likewise.
923 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
925 (cortex_a9_neon_bit_ops_q): New.
927 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
928 (cortex_a9_neon_int_2): Likewise.
929 (cortex_a9_neon_int_3): Likewise.
930 (cortex_a9_neon_int_4): Likewise.
931 (cortex_a9_neon_int_5): Likewise.
932 (cortex_a9_neon_vqneg_vqabs): Likewise.
933 (cortex_a9_neon_vmov): Likewise.
934 (cortex_a9_neon_vaba): Likewise.
935 (cortex_a9_neon_vaba_qqq): Likewise.
936 (cortex_a9_neon_shift_1): Likewise.
937 (cortex_a9_neon_shift_2): Likewise.
938 (cortex_a9_neon_shift_3): Likewise.
939 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
940 (cortex_a9_neon_vsra_vrsra): Likewise.
941 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
942 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
943 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
945 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
946 (cortex_a9_neon_mla_qqq_8_16): Likewise.
947 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
949 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
950 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
951 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
952 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
953 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
954 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
955 (cortex_a9_neon_fp_vsum): Likewise.
956 (cortex_a9_neon_fp_vmul_ddd): Likewise.
957 (cortex_a9_neon_fp_vmul_qqd): Likewise.
958 (cortex_a9_neon_fp_vmla_ddd): Likewise.
959 (cortex_a9_neon_fp_vmla_qqq): Likewise.
960 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
961 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
962 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
963 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
964 (cortex_a9_neon_bp_simple): Likewise.
965 (cortex_a9_neon_bp_2cycle): Likewise.
966 (cortex_a9_neon_bp_3cycle): Likewise.
967 (cortex_a9_neon_ldr): Likewise.
968 (cortex_a9_neon_str): Likewise.
969 (cortex_a9_neon_vld1_1_2_regs): Likewise.
970 (cortex_a9_neon_vld1_3_4_regs): Likewise.
971 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
972 (cortex_a9_neon_vld2_4_regs): Likewise.
973 (cortex_a9_neon_vld3_vld4): Likewise.
974 (cortex_a9_neon_vld1_vld2_lane): Likewise.
975 (cortex_a9_neon_vld3_vld4_lane): Likewise.
976 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
977 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
978 (cortex_a9_neon_vst1_3_4_regs): Likewise.
979 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
980 (cortex_a9_neon_vst1_vst2_lane): Likewise.
981 (cortex_a9_neon_vst3_vst4_lane): Likewise.
982 (cortex_a9_neon_mcr): Likewise.
983 (cortex_a9_neon_mcr_2_mcrr): Likewise.
984 (cortex_a9_neon_mrc): Likewise.
985 (cortex_a9_neon_mrrc): Likewise.
987 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
989 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
991 (cortex_a8_neon_vshl_ddd): Remove.
992 (cortex_a8_neon_vst3_vst4): Likewise.
993 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
995 (cortex_a8_neon_bit_ops_q): New.
997 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
998 (cortex_a8_neon_int_2): Likewise..
999 (cortex_a8_neon_int_3): Likewise.
1000 (cortex_a8_neon_int_5): Likewise.
1001 (cortex_a8_neon_vqneg_vqabs): Likewise.
1002 (cortex_a8_neon_int_4): Likewise.
1003 (cortex_a8_neon_vaba): Likewise.
1004 (cortex_a8_neon_vaba_qqq): Likewise.
1005 (cortex_a8_neon_shift_1): Likewise.
1006 (cortex_a8_neon_shift_2): Likewise.
1007 (cortex_a8_neon_shift_3): Likewise.
1008 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1009 (cortex_a8_neon_vsra_vrsra): Likewise.
1010 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1011 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1012 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1014 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1015 (cortex_a8_neon_mla_qqq_8_16): Likewise.
1016 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
1018 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1019 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1020 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
1021 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1022 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
1023 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
1024 (cortex_a8_neon_fp_vsum): Likewise.
1025 (cortex_a8_neon_fp_vmul_ddd): Likewise.
1026 (cortex_a8_neon_fp_vmul_qqd): Likewise.
1027 (cortex_a8_neon_fp_vmla_ddd): Likewise.
1028 (cortex_a8_neon_fp_vmla_qqq): Likewise.
1029 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
1030 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
1031 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1032 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1033 (cortex_a8_neon_bp_simple): Likewise.
1034 (cortex_a8_neon_bp_2cycle): Likewise.
1035 (cortex_a8_neon_bp_3cycle): Likewise.
1036 (cortex_a8_neon_ldr): Likewise.
1037 (cortex_a8_neon_str): Likewise.
1038 (cortex_a8_neon_vld1_1_2_regs): Likewise.
1039 (cortex_a8_neon_vld1_3_4_regs): Likewise.
1040 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1041 (cortex_a8_neon_vld2_4_regs): Likewise.
1042 (cortex_a8_neon_vld3_vld4): Likewise.
1043 (cortex_a8_neon_vld1_vld2_lane): Likewise.
1044 (cortex_a8_neon_vld3_vld4_lane): Likewise.
1045 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1046 (cortex_a8_neon_vst1_3_4_regs): Likewise.
1047 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
1048 (cortex_a8_neon_vst1_vst2_lane): Likewise.
1049 (cortex_a8_neon_vst3_vst4_lane): Likewise.
1050 (cortex_a8_neon_mcr): Likewise.
1051 (cortex_a8_neon_mcr_2_mcrr): Likewise.
1052 (cortex_a8_neon_mrc): Likewise.
1053 (cortex_a8_neon_mrrc): Likewise.
1055 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1057 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
1059 * config/aarch64/aarch64-simd.md (neon_type): Remove.
1060 (aarch64_simd_dup<mode>): Add "type" attribute.
1061 (aarch64_dup_lane<mode>): Likewise.
1062 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
1063 (*aarch64_simd_mov<mode>): Likewise.
1064 (aarch64_simd_mov_from_<mode>low): Likewise.
1065 (aarch64_simd_mov_from_<mode>high): Likewise.
1066 (orn<mode>3): Likewise.
1067 (bic<mode>3): Likewise.
1068 (add<mode>3): Likewise.
1069 (sub<mode>3): Likewise.
1070 (mul<mode>3): Likewise.
1071 (*aarch64_mul3_elt<mode>): Likewise.
1072 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
1073 (*aarch64_mul3_elt_to_128df): Likewise.
1074 (*aarch64_mul3_elt_to_64v2df): Likewise.
1075 (neg<mode>2): Likewise.
1076 (abs<mode>2): Likewise.
1077 (abd<mode>_3): Likewise.
1078 (aba<mode>_3): Likewise.
1079 (fabd<mode>_3): Likewise.
1080 (*fabd_scalar<mode>3): Likewise.
1081 (and<mode>3): Likewise.
1082 (ior<mode>3): Likewise.
1083 (xor<mode>3): Likewise.
1084 (one_cmpl<mode>2): Likewise.
1085 (aarch64_simd_vec_set<mode>): Likewise.
1086 (aarch64_simd_lshr<mode>): Likewise.
1087 (aarch64_simd_ashr<mode>): Likewise.
1088 (aarch64_simd_imm_shl<mode>): Likewise.
1089 (aarch64_simd_reg_sshl<mode): Likewise.
1090 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
1091 (aarch64_simd_reg_shl<mode>_signed): Likewise.
1092 (aarch64_simd_vec_setv2di): Likewise.
1093 (aarch64_simd_vec_set<mode>): Likewise.
1094 (aarch64_mla<mode>): Likewise.
1095 (*aarch64_mla_elt<mode>): Likewise.
1096 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
1097 (aarch64_mls<mode>): Likewise.
1098 (*aarch64_mls_elt<mode>): Likewise.
1099 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
1100 (<su><maxmin><mode>3): Likewise.
1101 (move_lo_quad_<mode>): Likewise.
1102 (aarch64_simd_move_hi_quad_<mode>): Likewise.
1103 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
1104 (vec_pack_trunc_<mode>): Likewise.
1105 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
1106 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
1107 (*aarch64_<su>mlal_lo<mode>): Likewise.
1108 (*aarch64_<su>mlal_hi<mode>): Likewise.
1109 (*aarch64_<su>mlsl_lo<mode>): Likewise.
1110 (*aarch64_<su>mlsl_hi<mode>): Likewise.
1111 (*aarch64_<su>mlal<mode>): Likewise.
1112 (*aarch64_<su>mlsl<mode>): Likewise.
1113 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
1114 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
1115 (add<mode>3): Likewise.
1116 (sub<mode>3): Likewise.
1117 (mul<mode>3): Likewise.
1118 (div<mode>3): Likewise.
1119 (neg<mode>2): Likewise.
1120 (abs<mode>2): Likewise.
1121 (fma<mode>4): Likewise.
1122 (*aarch64_fma4_elt<mode>): Likewise.
1123 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
1124 (*aarch64_fma4_elt_to_128df): Likewise.
1125 (*aarch64_fma4_elt_to_64v2df): Likewise.
1126 (fnma<mode>4): Likewise.
1127 (*aarch64_fnma4_elt<mode>): Likewise.
1128 (*aarch64_fnma4_elt_<vswap_width_name><mode>
1129 (*aarch64_fnma4_elt_to_128df): Likewise.
1130 (*aarch64_fnma4_elt_to_64v2df): Likewise.
1131 (<frint_pattern><mode>2): Likewise.
1132 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
1133 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
1134 (vec_unpacks_lo_v4sf): Likewise.
1135 (aarch64_float_extend_lo_v2df): Likewise.
1136 (vec_unpacks_hi_v4sf): Likewise.
1137 (aarch64_float_truncate_lo_v2sf): Likewise.
1138 (aarch64_float_truncate_hi_v4sf): Likewise.
1139 (aarch64_vmls<mode>): Likewise.
1140 (<su><maxmin><mode>3): Likewise.
1141 (<maxmin_uns><mode>3): Likewise.
1142 (reduc_<sur>plus_<mode>): Likewise.
1143 (reduc_<sur>plus_v2di): Likewise.
1144 (reduc_<sur>plus_v2si): Likewise.
1145 (reduc_<sur>plus_<mode>): Likewise.
1146 (aarch64_addpv4sf): Likewise.
1147 (clz<mode>2): Likewise.
1148 (reduc_<maxmin_uns>_<mode>): Likewise.
1149 (reduc_<maxmin_uns>_v2di): Likewise.
1150 (reduc_<maxmin_uns>_v2si): Likewise.
1151 (reduc_<maxmin_uns>_<mode>): Likewise.
1152 (reduc_<maxmin_uns>_v4sf): Likewise.
1153 (aarch64_simd_bsl<mode>_internal): Likewise.
1154 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
1155 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
1156 (aarch64_get_lane<mode>): Likewise.
1157 (*aarch64_combinez<mode>): Likewise.
1158 (aarch64_combine<mode>): Likewise.
1159 (aarch64_simd_combine<mode>): Likewise.
1160 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
1161 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
1162 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
1163 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
1164 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
1165 (aarch64_<sur>h<addsub><mode>): Likewise.
1166 (aarch64_<sur><addsub>hn<mode>): Likewise.
1167 (aarch64_<sur><addsub>hn2<mode>): Likewise.
1168 (aarch64_pmul<mode>): Likewise.
1169 (aarch64_<su_optab><optab><mode>): Likewise.
1170 (aarch64_<sur>qadd<mode>): Likewise.
1171 (aarch64_sqmovun<mode>): Likewise.
1172 (aarch64_<sur>qmovn<mode>): Likewise.
1173 (aarch64_s<optab><mode>): Likewise.
1174 (aarch64_sq<r>dmulh<mode>): Likewise.
1175 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1176 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
1177 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1178 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
1179 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1180 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1181 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
1182 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
1183 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
1184 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
1185 (aarch64_sqdmull<mode>): Likewise.
1186 (aarch64_sqdmull_lane<mode>_internal): Likewise.
1187 (aarch64_sqdmull_n<mode>): Likewise.
1188 (aarch64_sqdmull2<mode>_internal): Likewise.
1189 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
1190 (aarch64_sqdmull2_n<mode>_internal): Likewise.
1191 (aarch64_<sur>shl<mode>): Likewise.
1192 (aarch64_<sur>q<r>shl<mode>
1193 (aarch64_<sur>shll_n<mode>): Likewise.
1194 (aarch64_<sur>shll2_n<mode>): Likewise.
1195 (aarch64_<sur>shr_n<mode>): Likewise.
1196 (aarch64_<sur>sra_n<mode>): Likewise.
1197 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
1198 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
1199 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
1200 (aarch64_cm<optab><mode>): Likewise.
1201 (aarch64_cm<optab>di): Likewise.
1202 (aarch64_cm<optab><mode>): Likewise.
1203 (aarch64_cm<optab>di): Likewise.
1204 (aarch64_cmtst<mode>): Likewise.
1205 (aarch64_cmtstdi): Likewise.
1206 (aarch64_cm<optab><mode>): Likewise.
1207 (*aarch64_fac<optab><mode>): Likewise.
1208 (aarch64_addp<mode>): Likewise.
1209 (aarch64_addpdi): Likewise.
1210 (sqrt<mode>2): Likewise.
1211 (vec_load_lanesoi<mode>): Likewise.
1212 (vec_store_lanesoi<mode>): Likewise.
1213 (vec_load_lanesci<mode>): Likewise.
1214 (vec_store_lanesci<mode>): Likewise.
1215 (vec_load_lanesxi<mode>): Likewise.
1216 (vec_store_lanesxi<mode>): Likewise.
1217 (*aarch64_mov<mode>): Likewise.
1218 (aarch64_ld2<mode>_dreg): Likewise.
1219 (aarch64_ld2<mode>_dreg): Likewise.
1220 (aarch64_ld3<mode>_dreg): Likewise.
1221 (aarch64_ld3<mode>_dreg): Likewise.
1222 (aarch64_ld4<mode>_dreg): Likewise.
1223 (aarch64_ld4<mode>_dreg): Likewise.
1224 (aarch64_tbl1<mode>): Likewise.
1225 (aarch64_tbl2v16qi): Likewise.
1226 (aarch64_combinev16qi): Likewise.
1227 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
1228 (aarch64_st2<mode>_dreg): Likewise.
1229 (aarch64_st2<mode>_dreg): Likewise.
1230 (aarch64_st3<mode>_dreg): Likewise.
1231 (aarch64_st3<mode>_dreg): Likewise.
1232 (aarch64_st4<mode>_dreg): Likewise.
1233 (aarch64_st4<mode>_dreg): Likewise.
1234 (*aarch64_simd_ld1r<mode>): Likewise.
1235 (aarch64_frecpe<mode>): Likewise.
1236 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
1237 (aarch64_frecps<mode>): Likewise.
1239 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1241 * config/arm/iterators.md (V_elem_ch): New.
1243 (VQH_type): Likewise.
1244 * config/arm/arm.md (is_neon_type): New.
1245 (conds): Use is_neon_type.
1246 (anddi3_insn): Update type attribute.
1247 (xordi3_insn): Likewise.
1248 (one_cmpldi2): Likewise.
1249 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
1250 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
1251 (*movmisalign<mode>_neon_store): Likewise.
1252 (*movmisalign<mode>_neon_load): Likewise.
1253 (vec_set<mode>_internal): Likewise.
1254 (vec_set<mode>_internal): Likewise.
1255 (vec_setv2di_internal): Likewise.
1256 (vec_extract<mode>): Likewise.
1257 (vec_extract<mode>): Likewise.
1258 (vec_extractv2di): Likewise.
1259 (*add<mode>3_neon): Likewise.
1260 (adddi3_neon): Likewise.
1261 (*sub<mode>3_neon): Likewise.
1262 (subdi3_neon): Likewise.
1263 (fma<VCVTF:mode>4): Likewise.
1264 (fma<VCVTF:mode>4_intrinsic): Likewise.
1265 (*fmsub<VCVTF:mode>4): Likewise.
1266 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
1267 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
1268 (ior<mode>3): Likewise.
1269 (and<mode>3): Likewise.
1270 (orn<mode>3_neon): Likewise.
1271 (orndi3_neon): Likewise.
1272 (bic<mode>3_neon): Likewise.
1273 (bicdi3_neon): Likewise.
1274 (xor<mode>3): Likewise.
1275 (one_cmpl<mode>2): Likewise.
1276 (abs<mode>2): Likewise.
1277 (neg<mode>2): Likewise.
1278 (negdi2_neon): Likewise.
1279 (*umin<mode>3_neon): Likewise.
1280 (*umax<mode>3_neon): Likewise.
1281 (*smin<mode>3_neon): Likewise.
1282 (*smax<mode>3_neon): Likewise.
1283 (vashl<mode>3): Likewise.
1284 (vashr<mode>3_imm): Likewise.
1285 (vlshr<mode>3_imm): Likewise.
1286 (ashl<mode>3_signed): Likewise.
1287 (ashl<mode>3_unsigned): Likewise.
1288 (neon_load_count): Likewise.
1289 (ashldi3_neon_noclobber): Likewise.
1290 (ashldi3_neon): Likewise.
1291 (signed_shift_di3_neon): Likewise.
1292 (unsigned_shift_di3_neon): Likewise.
1293 (ashrdi3_neon_imm_noclobber): Likewise.
1294 (lshrdi3_neon_imm_noclobber): Likewise.
1295 (<shift>di3_neon): Likewise.
1296 (widen_ssum<mode>3): Likewise.
1297 (widen_usum<mode>3): Likewise.
1298 (quad_halves_<code>v4si): Likewise.
1299 (quad_halves_<code>v4sf): Likewise.
1300 (quad_halves_<code>v8hi): Likewise.
1301 (quad_halves_<code>v16qi): Likewise.
1302 (reduc_splus_v2di): Likewise.
1303 (neon_vpadd_internal<mode>): Likewise.
1304 (neon_vpsmin<mode>): Likewise.
1305 (neon_vpsmax<mode>): Likewise.
1306 (neon_vpumin<mode>): Likewise.
1307 (neon_vpumax<mode>): Likewise.
1308 (*ss_add<mode>_neon): Likewise.
1309 (*us_add<mode>_neon): Likewise.
1310 (*ss_sub<mode>_neon): Likewise.
1311 (*us_sub<mode>_neon): Likewise.
1312 (neon_vadd<mode>_unspec): Likewise.
1313 (neon_vaddl<mode>): Likewise.
1314 (neon_vaddw<mode>): Likewise.
1315 (neon_vhadd<mode>): Likewise.
1316 (neon_vqadd<mode>): Likewise.
1317 (neon_vaddhn<mode>): Likewise.
1318 (neon_vmul<mode>): Likewise.
1319 (neon_vfms<VCVTF:mode>): Likewise.
1320 (neon_vmlal<mode>): Likewise.
1321 (neon_vmls<mode>): Likewise.
1322 (neon_vmlsl<mode>): Likewise.
1323 (neon_vqdmulh<mode>): Likewise.
1324 (neon_vqdmlal<mode>): Likewise.
1325 (neon_vqdmlsl<mode>): Likewise.
1326 (neon_vmull<mode>): Likewise.
1327 (neon_vqdmull<mode>): Likewise.
1328 (neon_vsub<mode>_unspec): Likewise.
1329 (neon_vsubl<mode>): Likewise.
1330 (neon_vsubw<mode>): Likewise.
1331 (neon_vqsub<mode>): Likewise.
1332 (neon_vhsub<mode>): Likewise.
1333 (neon_vsubhn<mode>): Likewise.
1334 (neon_vceq<mode>): Likewise.
1335 (neon_vcge<mode>): Likewise.
1336 (neon_vcgeu<mode>): Likewise.
1337 (neon_vcgt<mode>): Likewise.
1338 (neon_vcgtu<mode>): Likewise.
1339 (neon_vcle<mode>): Likewise.
1340 (neon_vclt<mode>): Likewise.
1341 (neon_vcage<mode>): Likewise.
1342 (neon_vcagt<mode>): Likewise.
1343 (neon_vtst<mode>): Likewise.
1344 (neon_vabd<mode>): Likewise.
1345 (neon_vabdl<mode>): Likewise.
1346 (neon_vaba<mode>): Likewise.
1347 (neon_vabal<mode>): Likewise.
1348 (neon_vmax<mode>): Likewise.
1349 (neon_vmin<mode>): Likewise.
1350 (neon_vpaddl<mode>): Likewise.
1351 (neon_vpadal<mode>): Likewise.
1352 (neon_vpmax<mode>): Likewise.
1353 (neon_vpmin<mode>): Likewise.
1354 (neon_vrecps<mode>): Likewise.
1355 (neon_vrsqrts<mode>): Likewise.
1356 (neon_vqabs<mode>): Likewise.
1357 (neon_vqneg<mode>): Likewise.
1358 (neon_vcls<mode>): Likewise.
1359 (clz<mode>2): Likewise.
1360 (popcount<mode>2): Likewise.
1361 (neon_vrecpe<mode>): Likewise.
1362 (neon_vrsqrte<mode>): Likewise.
1363 (neon_vget_lane<mode>_sext_internal): Likewise.
1364 (neon_vget_lane<mode>_zext_internal): Likewise.
1365 (neon_vdup_n<mode>): Likewise.
1366 (neon_vdup_n<mode>): Likewise.
1367 (neon_vdup_nv2di): Likewise.
1368 (neon_vdup_lane<mode>_interal): Likewise.
1369 (*neon_vswp<mode>): Likewise.
1370 (neon_vcombine<mode>): Likewise.
1371 (float<mode><V_cvtto>2): Likewise.
1372 (floatuns<mode><V_cvtto>2): Likewise.
1373 (fix_trunc<mode><V_cvtto>2): Likewise.
1374 (fixuns_trunc<mode><V_cvtto>2
1375 (neon_vcvt<mode>): Likewise.
1376 (neon_vcvt<mode>): Likewise.
1377 (neon_vcvtv4sfv4hf): Likewise.
1378 (neon_vcvtv4hfv4sf): Likewise.
1379 (neon_vcvt_n<mode>): Likewise.
1380 (neon_vcvt_n<mode>): Likewise.
1381 (neon_vmovn<mode>): Likewise.
1382 (neon_vqmovn<mode>): Likewise.
1383 (neon_vqmovun<mode>): Likewise.
1384 (neon_vmovl<mode>): Likewise.
1385 (neon_vmul_lane<mode>): Likewise.
1386 (neon_vmul_lane<mode>): Likewise.
1387 (neon_vmull_lane<mode>): Likewise.
1388 (neon_vqdmull_lane<mode>): Likewise.
1389 (neon_vqdmulh_lane<mode>): Likewise.
1390 (neon_vqdmulh_lane<mode>): Likewise.
1391 (neon_vmla_lane<mode>): Likewise.
1392 (neon_vmla_lane<mode>): Likewise.
1393 (neon_vmlal_lane<mode>): Likewise.
1394 (neon_vqdmlal_lane<mode>): Likewise.
1395 (neon_vmls_lane<mode>): Likewise.
1396 (neon_vmls_lane<mode>): Likewise.
1397 (neon_vmlsl_lane<mode>): Likewise.
1398 (neon_vqdmlsl_lane<mode>): Likewise.
1399 (neon_vext<mode>): Likewise.
1400 (neon_vrev64<mode>): Likewise.
1401 (neon_vrev32<mode>): Likewise.
1402 (neon_vrev16<mode>): Likewise.
1403 (neon_vbsl<mode>_internal): Likewise.
1404 (neon_vshl<mode>): Likewise.
1405 (neon_vqshl<mode>): Likewise.
1406 (neon_vshr_n<mode>): Likewise.
1407 (neon_vshrn_n<mode>): Likewise.
1408 (neon_vqshrn_n<mode>): Likewise.
1409 (neon_vqshrun_n<mode>): Likewise.
1410 (neon_vshl_n<mode>): Likewise.
1411 (neon_vqshl_n<mode>): Likewise.
1412 (neon_vqshlu_n<mode>): Likewise.
1413 (neon_vshll_n<mode>): Likewise.
1414 (neon_vsra_n<mode>): Likewise.
1415 (neon_vsri_n<mode>): Likewise.
1416 (neon_vsli_n<mode>): Likewise.
1417 (neon_vtbl1v8qi): Likewise.
1418 (neon_vtbl2v8qi): Likewise.
1419 (neon_vtbl3v8qi): Likewise.
1420 (neon_vtbl4v8qi): Likewise.
1421 (neon_vtbl1v16qi): Likewise.
1422 (neon_vtbl2v16qi): Likewise.
1423 (neon_vcombinev16qi): Likewise.
1424 (neon_vtbx1v8qi): Likewise.
1425 (neon_vtbx2v8qi): Likewise.
1426 (neon_vtbx3v8qi): Likewise.
1427 (neon_vtbx4v8qi): Likewise.
1428 (*neon_vtrn<mode>_insn): Likewise.
1429 (*neon_vzip<mode>_insn): Likewise.
1430 (*neon_vuzp<mode>_insn): Likewise.
1431 (neon_vld1<mode>): Likewise.
1432 (neon_vld1_lane<mode>): Likewise.
1433 (neon_vld1_lane<mode>): Likewise.
1434 (neon_vld1_dup<mode>): Likewise.
1435 (neon_vld1_dup<mode>): Likewise.
1436 (neon_vld1_dupv2di): Likewise.
1437 (neon_vst1<mode>): Likewise.
1438 (neon_vst1_lane<mode>): Likewise.
1439 (neon_vst1_lane<mode>): Likewise.
1440 (neon_vld2<mode>): Likewise.
1441 (neon_vld2<mode>): Likewise.
1442 (neon_vld2_lane<mode>): Likewise.
1443 (neon_vld2_lane<mode>): Likewise.
1444 (neon_vld2_dup<mode>): Likewise.
1445 (neon_vst2<mode>): Likewise.
1446 (neon_vst2<mode>): Likewise.
1447 (neon_vst2_lane<mode>): Likewise.
1448 (neon_vst2_lane<mode>): Likewise.
1449 (neon_vld3<mode>): Likewise.
1450 (neon_vld3qa<mode>): Likewise.
1451 (neon_vld3qb<mode>): Likewise.
1452 (neon_vld3_lane<mode>): Likewise.
1453 (neon_vld3_lane<mode>): Likewise.
1454 (neon_vld3_dup<mode>): Likewise.
1455 (neon_vst3<mode>): Likewise.
1456 (neon_vst3qa<mode>): Likewise.
1457 (neon_vst3qb<mode>): Likewise.
1458 (neon_vst3_lane<mode>): Likewise.
1459 (neon_vst3_lane<mode>): Likewise.
1460 (neon_vld4<mode>): Likewise.
1461 (neon_vld4qa<mode>): Likewise.
1462 (neon_vld4qb<mode>): Likewise.
1463 (neon_vld4_lane<mode>): Likewise.
1464 (neon_vld4_lane<mode>): Likewise.
1465 (neon_vld4_dup<mode>): Likewise.
1466 (neon_vst4<mode>): Likewise.
1467 (neon_vst4qa<mode>): Likewise.
1468 (neon_vst4qb<mode>): Likewise.
1469 (neon_vst4_lane<mode>): Likewise.
1470 (neon_vst4_lane<mode>): Likewise.
1471 (neon_vec_unpack<US>_lo_<mode>): Likewise.
1472 (neon_vec_unpack<US>_hi_<mode>): Likewise.
1473 (neon_vec_<US>mult_lo_<mode>): Likewise.
1474 (neon_vec_<US>mult_hi_<mode>): Likewise.
1475 (neon_vec_<US>shiftl_<mode>): Likewise.
1476 (neon_unpack<US>_<mode>): Likewise.
1477 (neon_vec_<US>mult_<mode>): Likewise.
1478 (vec_pack_trunc_<mode>): Likewise.
1479 (neon_vec_pack_trunc_<mode>): Likewise.
1480 (neon_vabd<mode>_2): Likewise.
1481 (neon_vabd<mode>_3): Likewise.
1483 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1485 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
1486 (load_pair): Update type attribute.
1487 (store_pair): Update type attribute.
1488 * config/aarch64/iterators.md (q): New.
1490 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1492 * config/arm/types.md: Add new types for Neon insns.
1494 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1495 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1496 Sergey Lega <sergey.s.lega@intel.com>
1497 Anna Tikhonova <anna.tikhonova@intel.com>
1498 Ilya Tocar <ilya.tocar@intel.com>
1499 Andrey Turetskiy <andrey.turetskiy@intel.com>
1500 Ilya Verbin <ilya.verbin@intel.com>
1501 Kirill Yukhin <kirill.yukhin@intel.com>
1502 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1504 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
1505 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
1506 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
1508 (srcp14<mode>): Ditto.
1509 (rsqrt14<mode>): Ditto.
1510 (rsqrt14<mode>): Ditto.
1511 (avx512f_vmscalef<mode>): Ditto.
1512 (avx512f_scalef<mode>): Ditto.
1513 (avx512f_getexp<mode>): Ditto.
1514 (avx512f_sgetexp<mode>): Ditto.
1515 (avx512f_fixupimm<mode>): Ditto.
1516 (avx512f_sfixupimm<mode>): Ditto.
1517 (avx512f_rndscale<mode>): Ditto.
1518 (*avx512er_exp2<mode>): Ditto.
1519 (*avx512er_rcp28<mode>): Ditto.
1520 (avx512er_rsqrt28<mode>): Ditto.
1521 (avx512f_getmant<mode>): Ditto.
1522 (avx512f_getmant<mode>): Ditto.
1523 (avx512f_rndscale<mode>): Fix formatting.
1525 2013-10-15 Martin Jambor <mjambor@suse.cz>
1527 * ipa-utils.h (ipa_edge_within_scc): Declare.
1528 * ipa-cp.c (edge_within_scc): Moved...
1529 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
1531 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1532 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1533 Sergey Lega <sergey.s.lega@intel.com>
1534 Anna Tikhonova <anna.tikhonova@intel.com>
1535 Ilya Tocar <ilya.tocar@intel.com>
1536 Andrey Turetskiy <andrey.turetskiy@intel.com>
1537 Ilya Verbin <ilya.verbin@intel.com>
1538 Kirill Yukhin <kirill.yukhin@intel.com>
1539 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1541 * config/i386/predicates.md (const_8_to_15_operand): New.
1542 (const_16_to_31_operand): Ditto.
1543 * config/i386/sse.md (V8FI): New.
1545 (reduc_splus_v8df): Ditto.
1546 (reduc_splus_v16sf): Ditto.
1547 (avx512f_vextract<shuffletype>32x4_1): Ditto.
1548 (vec_extract_hi_<mode>): Ditto.
1549 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
1550 (vec_set_lo_<mode>): Ditto.
1551 (vec_set_hi_<mode>): Ditto.
1552 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
1553 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
1554 (avx512f_pshufd_1): Ditto.
1555 (avx512f_broadcast<mode>): Ditto.
1556 (avx512f_broadcast<mode>): Ditto.
1557 (define_split): Split vec_extract_lo into move.
1558 (ssequartermode): Ditto.
1559 (ssedoublemode): Extened with wider modes.
1560 (vec_extract_lo_<mode>): Ditto.
1562 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1563 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1564 Sergey Lega <sergey.s.lega@intel.com>
1565 Anna Tikhonova <anna.tikhonova@intel.com>
1566 Ilya Tocar <ilya.tocar@intel.com>
1567 Andrey Turetskiy <andrey.turetskiy@intel.com>
1568 Ilya Verbin <ilya.verbin@intel.com>
1569 Kirill Yukhin <kirill.yukhin@intel.com>
1570 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1572 * config/i386/predicates.md (register_or_constm1_operand): New.
1573 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
1574 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
1575 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
1576 UNSPEC_SCATTER_PREFETCH
1578 (avx512f_ucmp<mode>3): Ditto.
1579 (avx512f_vternlog<mode>): Ditto.
1580 (avx512f_align<mode>): Ditto.
1581 (<shift_insn><mode>3): Ditto.
1582 (avx512f_<rotate>v<mode>): Ditto.
1583 (avx512f_<rotate><mode>): Ditto.
1584 (avx512f_eq<mode>3): Ditto.
1585 (avx512f_eq<mode>3_1): Ditto.
1586 (avx512f_gt<mode>3): Ditto.
1587 (avx512f_testm<mode>3): Ditto.
1588 (avx512f_testnm<mode>3): Ditto.
1589 (avx512pf_gatherpf<mode>): Ditto.
1590 (*avx512pf_gatherpf<mode>_mask): Ditto.
1591 (*avx512pf_gatherpf<mode>): Ditto.
1592 (avx512pf_scatterpf<mode>): Ditto.
1593 (*avx512pf_scatterpf<mode>_mask): Ditto.
1594 (*avx512pf_scatterpf<mode>): Ditto.
1595 (avx512f_vec_dup_gpr<mode>): Ditto.
1596 (clz<mode>2): Ditto.
1597 (conflict<mode>): Ditto.
1598 (REDUC_SMINMAX_MODE): Extened with wider modes.
1599 (reduc_<code>_<mode>): Ditto.
1600 (vlshr<mode>3): Ditto.
1601 (vashl<mode>3): Ditto.
1603 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1604 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1605 Sergey Lega <sergey.s.lega@intel.com>
1606 Anna Tikhonova <anna.tikhonova@intel.com>
1607 Ilya Tocar <ilya.tocar@intel.com>
1608 Andrey Turetskiy <andrey.turetskiy@intel.com>
1609 Ilya Verbin <ilya.verbin@intel.com>
1610 Kirill Yukhin <kirill.yukhin@intel.com>
1611 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1613 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
1616 (avx512fmaskmode): Ditto.
1617 (bcstscalarsuff): Ditto.
1618 (avx512f_blendm<mode>): Ditto.
1619 (cmp_imm_predicate): Ditto.
1620 (avx512f_cmp<mode>3): Ditto.
1621 (avx512f_vec_dup<mode>): Ditto.
1622 (avx512f_vec_dup_mem<mode>): Ditto.
1623 (avx512f_vpermi2var<mode>3): Ditto.
1624 (avx512f_vpermt2var<mode>3): Ditto.
1625 (vec_init<mode>): Ditto.
1626 (avx512f_gathersi<mode>): Ditto.
1627 (*avx512f_gathersi<mode>): Ditto.
1628 (*avx512f_gathersi<mode>_2): Ditto.
1629 (avx512f_gatherdi<mode>): Ditto.
1630 (*avx512f_gatherdi<mode>): Ditto.
1631 (*avx512f_gatherdi<mode>_2): Ditto.
1632 (avx512f_scattersi<mode>): Ditto.
1633 (*avx512f_scattersi<mode>): Ditto.
1634 (avx512f_scatterdi<mode>): Ditto.
1635 (*avx512f_scatterdi<mode>): Ditto.
1636 (sseintprefix): Extened with wider modes.
1637 (VEC_GATHER_IDXSI): Ditto.
1638 (VEC_GATHER_IDXDI): Ditto.
1639 (VEC_GATHER_SRCDI): Ditto.
1641 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
1642 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1644 * config/arm/t-aprofile: New file.
1645 * config.gcc: Handle --with-multilib-list option.
1647 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
1649 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
1652 2013-10-15 Richard Biener <rguenther@suse.de>
1654 * tree-loop-distribution.c (build_empty_rdg): Inline into
1656 (rdg_flag_vertex): Inline into single user.
1657 (rdg_flag_vertex_and_dependent): Likewise.
1658 (build_rdg_partition_for_vertex): Remove processed bitmap.
1659 (rdg_build_partitions): Simplify.
1661 2013-10-15 Richard Biener <rguenther@suse.de>
1663 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
1664 Restructure forwarding through conversions and copies to
1665 avoid performing copy-propagation the wrong way. Adjust
1666 recursion invocations.
1667 (forward_propagate_addr_expr): Add argument stating if we
1668 are recursing from a single-use.
1669 (ssa_forward_propagate_and_combine): Adjust.
1671 2013-10-14 David Malcolm <dmalcolm@redhat.com>
1673 * dumpfile.h (gcc::dump_manager): New class, to hold state
1674 relating to dumpfile management.
1675 (get_dump_file_name): Remove in favor of method of dump_manager.
1676 (dump_initialized_p): Likewise.
1677 (dump_start): Likewise.
1678 (dump_finish): Likewise.
1679 (dump_switch_p): Likewise.
1680 (dump_register): Likewise.
1681 (get_dump_file_info): Likewise.
1682 * context.c (gcc::context::context): Construct the dump_manager
1684 * context.h (gcc::context::get_dumps): New.
1685 (gcc::context::m_dumps): New.
1686 * coverage.c (coverage_init): Port to dump_manager API.
1687 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
1688 (extra_dump_files_in_use): Likewise.
1689 (extra_dump_files_alloced): Likewise.
1690 (gcc::dump_manager::dump_manager): New.
1691 (dump_register): Convert to...
1692 (gcc::dump_manager::dump_register): ...method, replacing
1693 function-static next_dump with m_next_dump field.
1694 (get_dump_file_info): Convert to...
1695 (gcc::dump_manager::get_dump_file_info): ...method.
1696 (get_dump_file_name): Convert to...
1697 (gcc::dump_manager::get_dump_file_name): ...method.
1698 (dump_start): Convert to...
1699 (gcc::dump_manager::dump_start): ...method.
1700 (dump_finish): Convert to...
1701 (gcc::dump_manager::dump_finish): ...method.
1702 (dump_begin): Replace body with...
1703 (gcc::dump_manager::dump_begin): ...new method.
1704 (dump_phase_enabled_p): Convert to...
1705 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1706 (dump_phase_enabled_p): Convert to...
1707 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1708 (dump_initialized_p): Convert to...
1709 (gcc::dump_manager::dump_initialized_p): ...method.
1710 (dump_flag_name): Replace body with...
1711 (gcc::dump_manager::dump_flag_name): ...new method.
1712 (dump_enable_all): Convert to...
1713 (gcc::dump_manager::dump_enable_all): ...new method.
1714 (opt_info_enable_passes): Convert to...
1715 (gcc::dump_manager::opt_info_enable_passes): ...new method.
1716 (dump_switch_p_1): Convert to...
1717 (gcc::dump_manager::dump_switch_p_1): ...new method.
1718 (dump_switch_p): Convert to...
1719 (gcc::dump_manager::dump_switch_p): ...new method.
1720 (opt_info_switch_p): Port to dump_manager API.
1721 (enable_rtl_dump_file): Likewise.
1722 * opts-global.c (handle_common_deferred_options): Port to new
1724 * passes.c (pass_manager::finish_optimization_passes): Likewise.
1725 (pass_manager::register_one_dump_file): Likewise.
1726 (pass_manager::register_pass): Likewise.
1727 (pass_init_dump_file): Likewise.
1728 (pass_fini_dump_file): Likewise.
1729 * statistics.c (statistics_early_init): Likewise.
1731 2013-10-14 Richard Biener <rguenther@suse.de>
1733 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
1734 iterative_hash_canonical_type, gimple_canonical_type_hash,
1735 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
1736 gimple_register_canonical_type, print_gimple_types_stats,
1737 free_gimple_type_tables): Move to lto/lto.c
1738 (gt-gimple.h): Do not include.
1739 * gimple.h (gimple_register_canonical_type,
1740 print_gimple_types_stats, free_gimple_type_tables): Remove.
1741 * Makefile.in (GTFILES): Remove gimple.c.
1743 2013-10-14 Travis Snoozy <quandary@remstate.com>
1746 * config/msp430/msp430.c (msp430_option_override): Correct thinko
1747 scanning for msp430x targets.
1749 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
1752 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
1753 (registers_ok_for_ldd_peep): Move around.
1754 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
1755 * config/sparc/sparc.md (widening peepholes): Use it.
1757 2013-10-14 Richard Biener <rguenther@suse.de>
1761 * gimple.c (iterative_hash_canonical_type): Make sure to
1762 record the hash into the correct hashtable slot.
1764 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
1766 PR rtl-optimization/58662
1767 * combine.c (try_combine): Take into account death nodes on I2 when
1768 splitting a PARALLEL of two independent SETs. Fix dump message.
1770 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
1773 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
1774 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
1775 * config/sh/t-sh (sh_treg_combine.o): New entry.
1776 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
1777 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
1778 (register_sh_passes): New function. Register sh_treg_combine pass.
1779 (sh_option_override): Invoke it.
1780 (sh_canonicalize_comparison): Handle op0_preserve_value.
1781 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
1782 opportunities. Canonicalize branch condition.
1783 (nott): Allow only if pseudos can be created for non-SH2A.
1785 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
1788 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
1789 (ix86_expand_movmem): Replace copy_addr_to_reg with
1790 ix86_copy_addr_to_reg.
1791 (ix86_expand_setmem): Likewise.
1793 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
1795 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
1796 provided by ix86_fp_compare_mode instead of CCFPUmode.
1798 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
1800 * config/aarch64/arm_neon.h
1801 (vtbx<1,3>_<psu>8): Fix register constriants.
1803 2013-10-11 Jeff Law <law@redhat.com>
1805 PR tree-optimization/58640
1806 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
1807 threading paths that cross over two loop entry points.
1809 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1811 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
1812 handle vector float as well.
1813 (*vsx_le_perm_load_v4si): Likewise.
1814 (*vsx_le_perm_store_v2di): Likewise.
1815 (*vsx_le_perm_store_v4si): Likewise.
1817 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1819 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
1820 directly to circumvent subtract from splat{31} workaround.
1821 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
1823 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
1824 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
1825 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
1826 (altivec_vperm_<mode>): Convert to define_insn_and_split to
1827 separate big and little endian logic.
1828 (*altivec_vperm_<mode>_internal): New define_insn.
1829 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
1830 separate big and little endian logic.
1831 (*altivec_vperm_<mode>_uns_internal): New define_insn.
1832 (vec_permv16qi): Add little endian logic.
1834 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
1836 * doc/extend.texi (returns_nonnull): Remove arguments.
1838 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1839 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1840 Sergey Lega <sergey.s.lega@intel.com>
1841 Anna Tikhonova <anna.tikhonova@intel.com>
1842 Ilya Tocar <ilya.tocar@intel.com>
1843 Andrey Turetskiy <andrey.turetskiy@intel.com>
1844 Ilya Verbin <ilya.verbin@intel.com>
1845 Kirill Yukhin <kirill.yukhin@intel.com>
1846 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1848 * config/i386/sse.md (VI48F_256_512): New.
1849 (avx2_permvar<mode>): Change to ...
1850 (<avx2_avx512f>_permvar<mode>): This.
1852 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1853 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1854 Sergey Lega <sergey.s.lega@intel.com>
1855 Anna Tikhonova <anna.tikhonova@intel.com>
1856 Ilya Tocar <ilya.tocar@intel.com>
1857 Andrey Turetskiy <andrey.turetskiy@intel.com>
1858 Ilya Verbin <ilya.verbin@intel.com>
1859 Kirill Yukhin <kirill.yukhin@intel.com>
1860 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1862 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
1863 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
1864 * config/i386/sse.md (VI4_AVX): New.
1865 (sf2simodelower): Ditto.
1866 (sse2_cvtps2dq): Change to ...
1867 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
1869 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1870 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1871 Sergey Lega <sergey.s.lega@intel.com>
1872 Anna Tikhonova <anna.tikhonova@intel.com>
1873 Ilya Tocar <ilya.tocar@intel.com>
1874 Andrey Turetskiy <andrey.turetskiy@intel.com>
1875 Ilya Verbin <ilya.verbin@intel.com>
1876 Kirill Yukhin <kirill.yukhin@intel.com>
1877 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1879 * config/i386/sse.md (V_512): New.
1881 (vcond<V_512:mode><VF_512:mode>): Ditto.
1882 (vcond<V_512:mode><VI_512:mode>): Ditto.
1883 (vcondu<V_512:mode><VI_512:mode>): Ditto.
1885 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1886 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1887 Sergey Lega <sergey.s.lega@intel.com>
1888 Anna Tikhonova <anna.tikhonova@intel.com>
1889 Ilya Tocar <ilya.tocar@intel.com>
1890 Andrey Turetskiy <andrey.turetskiy@intel.com>
1891 Ilya Verbin <ilya.verbin@intel.com>
1892 Kirill Yukhin <kirill.yukhin@intel.com>
1893 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1895 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
1896 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
1898 (fma<mode>4): Removed condition.
1899 (fms<mode>4): Ditto.
1900 (fnma<mode>4): Ditto.
1901 (fnms<mode>4): Ditto.
1902 (fma4i_fmadd_<mode>): Ditto.
1903 (*fma_fmadd_<mode>): Ditto.
1904 (*fma_fmsub_<mode>): Ditto.
1905 (*fma_fnmadd_<mode>): Ditto.
1906 (*fma_fnmsub_<mode>): Ditto.
1907 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
1908 (*fma_fmaddsub_<mode>): Ditto.
1909 (*fma_fmsubadd_<mode>): Ditto.
1910 (*fmai_fmadd_<mode>): Ditto.
1911 (*fmai_fmsub_<mode>): Ditto.
1912 (*fmai_fnmadd_<mode>): Ditto.
1913 (*fmai_fnmsub_<mode>): Ditto.
1915 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1916 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1917 Sergey Lega <sergey.s.lega@intel.com>
1918 Anna Tikhonova <anna.tikhonova@intel.com>
1919 Ilya Tocar <ilya.tocar@intel.com>
1920 Andrey Turetskiy <andrey.turetskiy@intel.com>
1921 Ilya Verbin <ilya.verbin@intel.com>
1922 Kirill Yukhin <kirill.yukhin@intel.com>
1923 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1925 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
1926 (VI124_256): Changed to ...
1927 (VI124_256_48_512): This.
1928 (ssepackmode): Extended with wider modes.
1929 (<code><mode>3): Changed iterator.
1930 (*avx2_<code><mode>3): Ditto.
1931 (vec_pack_trunc_<mode>): Ditto.
1933 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1934 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1935 Sergey Lega <sergey.s.lega@intel.com>
1936 Anna Tikhonova <anna.tikhonova@intel.com>
1937 Ilya Tocar <ilya.tocar@intel.com>
1938 Andrey Turetskiy <andrey.turetskiy@intel.com>
1939 Ilya Verbin <ilya.verbin@intel.com>
1940 Kirill Yukhin <kirill.yukhin@intel.com>
1941 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1943 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
1944 (VI8F_256_512): Ditto.
1945 (abs<mode>2): Changed iterator.
1946 (avx2_perm<mode>): Changed to ...
1947 (<avx2_avx512f>_perm<mode>): This.
1948 (avx2_perm<mode>_1): Changed to ...
1949 (<avx2_avx512f>_perm<mode>_1): This.
1951 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1952 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1953 Sergey Lega <sergey.s.lega@intel.com>
1954 Anna Tikhonova <anna.tikhonova@intel.com>
1955 Ilya Tocar <ilya.tocar@intel.com>
1956 Andrey Turetskiy <andrey.turetskiy@intel.com>
1957 Ilya Verbin <ilya.verbin@intel.com>
1958 Kirill Yukhin <kirill.yukhin@intel.com>
1959 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1961 * config/i386/sse.md (VI48_AVX512F): New.
1962 (VI48_AVX2): Changed to ...
1963 (VI48_AVX2_48_AVX512F): This.
1964 (avx2_ashrv<mode>): Changed to ...
1965 (<avx2_avx512f>_ashrv<mode>): This.
1966 (avx2_<shift_insn>v<mode>): Changed to ...
1967 (<avx2_avx512f>_<shift_insn>v<mode>): This.
1969 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1970 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1971 Sergey Lega <sergey.s.lega@intel.com>
1972 Anna Tikhonova <anna.tikhonova@intel.com>
1973 Ilya Tocar <ilya.tocar@intel.com>
1974 Andrey Turetskiy <andrey.turetskiy@intel.com>
1975 Ilya Verbin <ilya.verbin@intel.com>
1976 Kirill Yukhin <kirill.yukhin@intel.com>
1977 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1979 * config/i386/sse.md (VI4_AVX512F): New.
1980 (VI8_AVX2_AVX512F): Ditto.
1981 (mul<mode>3): Extended with wider modes.
1982 (*<sse4_1_avx2>_mul<mode>3): Ditto.
1983 (mul<mode>3): Ditto.
1984 (vec_widen_<s>mult_odd_<mode>): Ditto.
1986 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1987 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1988 Sergey Lega <sergey.s.lega@intel.com>
1989 Anna Tikhonova <anna.tikhonova@intel.com>
1990 Ilya Tocar <ilya.tocar@intel.com>
1991 Andrey Turetskiy <andrey.turetskiy@intel.com>
1992 Ilya Verbin <ilya.verbin@intel.com>
1993 Kirill Yukhin <kirill.yukhin@intel.com>
1994 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1996 * config/i386/sse.md (VI2_AVX512F): New.
1997 (VI124_AVX512F): Ditto.
1998 (sseunpackmode): Extended with wider modes.
1999 (sseunpackfltmode): Ditto.
2000 (vec_unpacks_float_hi_<mode>): Ditto.
2001 (vec_unpacks_float_lo_<mode>): Ditto.
2002 (vec_unpacku_float_hi_<mode>): Ditto.
2003 (vec_unpacku_float_lo_<mode>): Ditto.
2004 (vec_unpacks_lo_<mode>): Ditto.
2005 (vec_unpacks_hi_<mode>): Ditto.
2006 (vec_unpacku_lo_<mode>): Ditto.
2007 (vec_unpacku_hi_<mode>): Ditto.
2009 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2010 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2011 Sergey Lega <sergey.s.lega@intel.com>
2012 Anna Tikhonova <anna.tikhonova@intel.com>
2013 Ilya Tocar <ilya.tocar@intel.com>
2014 Andrey Turetskiy <andrey.turetskiy@intel.com>
2015 Ilya Verbin <ilya.verbin@intel.com>
2016 Kirill Yukhin <kirill.yukhin@intel.com>
2017 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2019 * config/i386/i386.md (multdiv): New.
2020 (multdiv_mnemonic): Ditto.
2021 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
2022 (<sse>_vm<multdiv_mnemonic><mode>3): This.
2023 (<sse>_vmdiv<mode>3): Removed.
2025 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2026 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2027 Sergey Lega <sergey.s.lega@intel.com>
2028 Anna Tikhonova <anna.tikhonova@intel.com>
2029 Ilya Tocar <ilya.tocar@intel.com>
2030 Andrey Turetskiy <andrey.turetskiy@intel.com>
2031 Ilya Verbin <ilya.verbin@intel.com>
2032 Kirill Yukhin <kirill.yukhin@intel.com>
2033 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2035 * config/i386/sse.md (V): Extended with wider modes.
2037 (ssehalfvecmode): Ditto.
2039 (ssepackfltmode): Ditto.
2040 (avx_vec_concat<mode>): Ditto.
2041 (V_256_512): New iterator.
2042 (VF2_512_256): Ditto.
2043 (si2dfmode): New attribute.
2044 (si2dfmodelower): Ditto.
2046 (concat_tg_mode): Ditto.
2047 (floatv4siv4df2): Changed to ...
2048 (float<si2dfmodelower><mode>2): This.
2049 (avx_cvtps2pd256): Changed to ...
2050 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
2051 (vec_pack_trunc_v4df): Changed to ...
2052 (vec_pack_trunc_<mode>): This.
2053 (avx_vpermil<mode>): Changed to ...
2054 (<sse2_avx_avx512f>_vpermil<mode>): This.
2055 (<fixsuffix>fix_truncv8dfv8si2): New.
2056 (vec_pack_sfix_trunc_v8df): Ditto.
2057 (avx512f_rndscale<mode>): Ditto.
2058 (avx512f_roundpd512): Ditto.
2059 (vec_pack_ufix_trunc_<mode>): Updated iterator.
2061 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2062 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2063 Sergey Lega <sergey.s.lega@intel.com>
2064 Anna Tikhonova <anna.tikhonova@intel.com>
2065 Ilya Tocar <ilya.tocar@intel.com>
2066 Andrey Turetskiy <andrey.turetskiy@intel.com>
2067 Ilya Verbin <ilya.verbin@intel.com>
2068 Kirill Yukhin <kirill.yukhin@intel.com>
2069 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2071 * config/i386/i386.md (any_fix): New iterator.
2072 (fixsuffix): New attribute.
2073 * config/i386/sse.md (VF1): Extened with wider modes.
2077 (sseintvecmodelower): Ditto.
2078 (ssescalarmode): Ditto.
2079 (ssescalarnum): Ditto.
2081 (ssexmmmode): Ditto.
2082 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
2083 (<sse>_rcp<mode>2): Change iterator.
2084 (rsqrt<mode>2): Ditto.
2085 (<sse>_rsqrt<mode>2): Ditto.
2086 (avx2_vec_dup<mode>): Ditto.
2087 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
2088 (round<mode>2_sfix): Ditto.
2089 (avx2_pbroadcast<mode>): Ditto.
2090 (*andnot<mode>3): Handle XI mode.
2091 (*<code><mode>3): Ditto.
2092 (AVXTOSSEMODE): Removed.
2093 (avx_vpermil<mode>): Changed to ...
2094 (<sse2_avx_avx512f>_vpermil<mode>): This.
2096 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2097 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2098 Sergey Lega <sergey.s.lega@intel.com>
2099 Anna Tikhonova <anna.tikhonova@intel.com>
2100 Ilya Tocar <ilya.tocar@intel.com>
2101 Andrey Turetskiy <andrey.turetskiy@intel.com>
2102 Ilya Verbin <ilya.verbin@intel.com>
2103 Kirill Yukhin <kirill.yukhin@intel.com>
2104 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2106 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
2107 (<sse>_comi): Ditto.
2108 (<sse>_ucomi): Ditto.
2109 (sse_cvtss2siq_2): Ditto.
2110 (sse2_cvtsd2si): Ditto.
2111 (sse2_cvtsd2siq): Ditto.
2112 (sse2_cvttsd2si): Ditto.
2113 (sse2_cvttsd2siq): Ditto.
2114 (<shift_insn><mode>3): Ditto.
2115 (sse2_cvtsi2sdq): Update constraint and prefix.
2116 (sse_cvtsi2ss): Update prefix.
2117 (sse_cvtsi2ssq): Ditto.
2119 2013-10-11 Jakub Jelinek <jakub@redhat.com>
2121 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
2122 ignore internal calls.
2124 2013-10-11 Richard Biener <rguenther@suse.de>
2126 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
2127 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
2128 TREE_OVERFLOW is set.
2130 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
2132 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
2134 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
2135 * gimple.h: Likewise.
2137 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
2139 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
2142 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
2144 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
2145 * configure: Regenerate.
2147 2013-10-11 Jakub Jelinek <jakub@redhat.com>
2149 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
2150 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
2151 formatting fixes, use pp_colon instead of pp_character (..., ':'),
2152 similarly pp_right_paren.
2153 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
2154 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
2155 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
2156 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
2158 (omp_declare_simd_clauses_equal,
2159 omp_remove_redundant_declare_simd_attrs): New functions.
2160 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
2161 (walk_tree_1): Handle new OpenMP 4.0 clauses.
2162 * tree.h (OMP_LOOP_CHECK): Define.
2163 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
2164 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
2165 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
2166 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
2167 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
2168 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
2169 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
2170 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
2171 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
2172 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
2173 OMP_CLAUSE_SIMDLEN_EXPR): Define.
2174 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
2175 (omp_remove_redundant_declare_simd_attrs): New prototype.
2176 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
2177 GIMPLE_OMP_TEAMS): New codes.
2178 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
2179 * omp-low.c (struct omp_context): Add cancel_label and cancellable
2181 (target_nesting_level): New variable.
2182 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
2183 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
2184 for collapse > 1 static schedule unless ordered.
2185 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
2186 (determine_parallel_type): Adjust get_ws_args_for caller.
2187 (install_var_field): Handle mask & 4 for double indirection.
2188 (scan_sharing_clauses): Ignore shared clause on teams construct.
2189 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
2190 (create_omp_child_function): If inside target or declare target
2191 constructs, set "omp declare target" attribute on the child function.
2192 (find_combined_for): New function.
2193 (scan_omp_parallel): Handle combined loops.
2194 (scan_omp_target, scan_omp_teams): New functions.
2195 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
2196 restrictions and set ctx->cancellable for cancellable constructs.
2197 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
2198 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
2199 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
2200 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
2201 (omp_clause_aligned_alignment): New function.
2202 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
2203 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
2204 on teams constructs. Handle user defined reductions and new
2206 (lower_reduction_clauses): Don't set placeholder to address of ref
2207 if it has already the right type.
2208 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
2209 (expand_parallel_call): Use the new non-_start suffixed builtins,
2210 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
2211 and GOMP_parallel_end after the call.
2212 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
2213 (expand_omp_for_init_counts): Handle combined loops.
2214 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
2216 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
2217 end of cancellable loops.
2218 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
2219 Likewise. Handle collapse > 1 loops.
2220 (expand_omp_simd): Handle combined loops.
2221 (expand_omp_for): Add inner_stmt argument, adjust callers of
2222 expand_omp_for* functions, use expand_omp_for_static*chunk even
2223 for collapse > 1 unless ordered.
2224 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
2225 of cancellable sections.
2226 (expand_omp_single): Remove need_barrier variable, just rely on
2227 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
2228 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
2229 (expand_omp_atomic_load, expand_omp_atomic_store,
2230 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
2231 (expand_omp_target): New function.
2232 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
2233 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
2234 (build_omp_regions_1): Immediately close region for
2235 GF_OMP_TARGET_KIND_UPDATE.
2236 (maybe_add_implicit_barrier_cancel): New function.
2237 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
2239 (lower_omp_single): Likewise. Add clobber after the barrier.
2240 (lower_omp_taskgroup): New function.
2241 (lower_omp_for): Handle combined loops. Adjust
2242 lower_rec_input_clauses caller. Handle cancellation.
2243 (lower_depend_clauses): New function.
2244 (lower_omp_taskreg): Lower depend clauses. Adjust
2245 lower_rec_input_clauses caller. Add clobber after the call. Handle
2247 (lower_omp_target, lower_omp_teams): New functions.
2248 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
2249 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
2250 and GOMP_cancellation_point calls.
2251 (lower_omp): Fold stmts inside of target region.
2252 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
2253 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2254 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
2255 (BT_FN_VOID_OMPFN_PTR_UINT,
2256 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
2257 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
2258 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
2259 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
2260 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
2261 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
2262 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
2263 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
2264 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
2265 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
2266 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
2267 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
2268 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
2269 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
2270 BUILT_IN_GOMP_PARALLEL_END.
2271 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
2272 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2273 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
2274 GF_OMP_FOR_KIND_DISTRIBUTE.
2275 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
2276 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
2277 (dump_gimple_omp_return): Print lhs if it has any.
2278 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
2279 gimple_omp_atomic_seq_cst_p.
2280 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
2281 and GIMPLE_OMP_TEAMS.
2282 * langhooks.c (lhd_omp_mappable_type): New function.
2283 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
2284 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
2286 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
2287 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
2288 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
2289 (struct gimplify_omp_ctx): Add combined_loop field.
2290 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
2291 on stmts inside of target region.
2292 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
2293 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
2294 ORT_TARGET and ORT_TARGET_DATA.
2295 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
2297 (omp_notice_threadprivate_variable): Complain about threadprivate
2298 variables in target region.
2299 (omp_notice_variable): Complain about vars with non-mappable type
2300 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
2301 (omp_check_private): Ignore ORT_TARGET* regions.
2302 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
2303 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
2304 (find_combined_omp_for): New function.
2305 (gimplify_omp_for): Handle gimplification of combined loops.
2306 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
2308 (gimplify_omp_target_update): New function.
2309 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
2310 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
2311 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
2312 (gimplify_body): If fndecl has "omp declare target" attribute, add
2313 implicit ORT_TARGET context around it.
2314 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
2315 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
2316 * tree-nested.c (convert_nonlocal_reference_stmt,
2317 convert_local_reference_stmt, convert_gimple_call): Handle
2318 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2319 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
2320 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
2321 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
2322 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
2323 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
2324 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
2325 BUILT_IN_GOMP_LOOP_END_CANCEL,
2326 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
2327 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
2328 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
2329 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
2330 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
2331 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
2332 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
2333 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
2334 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
2335 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
2336 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
2337 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
2338 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
2339 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
2340 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
2341 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2342 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
2343 gimple_build_omp_teams): New functions.
2344 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
2345 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
2346 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
2347 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2348 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
2349 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
2350 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
2351 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
2352 GF_OMP_ATOMIC_SEQ_CST): New.
2353 (gimple_build_omp_taskgroup, gimple_build_omp_target,
2354 gimple_build_omp_teams): New prototypes.
2355 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
2356 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2357 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
2358 GIMPLE_OMP_SINGLE as end of range.
2359 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
2360 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
2361 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
2362 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
2363 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
2364 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
2365 gimple_omp_target_kind, gimple_omp_target_set_kind,
2366 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
2367 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
2368 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
2369 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
2370 gimple_omp_teams_set_clauses): New inlines.
2371 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
2372 and GIMPLE_OMP_TASKGROUP.
2373 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
2374 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
2375 enum omp_clause_proc_bind_kind): New.
2376 (union omp_clause_subcode): Add depend_kind, map_kind and
2377 proc_bind_kind fields.
2378 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
2379 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2380 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
2381 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
2382 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
2384 2013-10-10 Teresa Johnson <tejohnson@google.com>
2386 * predict.c (tree_estimate_probability): Add new parameter
2387 for estimate_bb_frequencies.
2388 (estimate_bb_frequencies): Add new parameter to force estimation.
2389 (rebuild_frequencies): When max frequency in function is small,
2390 recompute counts from frequencies.
2391 * predict.h (estimate_bb_frequencies): New parameter.
2393 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2395 * ipa-inline.c (ipa_inline): Fix leak of "order" when
2396 optimizations are disabled.
2398 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2400 * coverage.c (coverage_finish): Fix leak of da_file_name.
2402 2013-10-10 Jan Hubicka <jh@suse.cz>
2404 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
2405 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
2408 2013-10-10 Jakub Jelinek <jakub@redhat.com>
2411 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
2412 if any labels are in FALLTHRU_BB, use a special label emitted
2413 immediately after the asm goto insn rather than label_rtx
2415 (expand_asm_stmt): Adjust caller.
2416 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
2417 edge if the last insn in predecessor is a jump with single successor,
2418 but it isn't simplejump_p.
2420 2013-10-10 Richard Biener <rguenther@suse.de>
2422 PR tree-optimization/58656
2423 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
2425 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2427 * gimplify.c: Include expr.h and tm_p.h for targets with special
2428 va-arg padding requirements.
2430 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2432 * tree-flow.h: Move some prototypes to gimple.h.
2433 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
2434 * gimple.h: Relocate some prototypes from tree-flow.h
2435 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
2437 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
2438 (build_va_arg_indirect_ref): Relocate and make static.
2439 (std_gimplify_va_arg_expr): Relocate here.
2440 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
2441 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
2443 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2445 * doc/md.texi: Document the mnemonic attribute.
2447 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2450 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
2451 (set_attr_alternative x ...) when searching for user defined
2454 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2456 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
2458 2013-10-09 Easwaran Raman <eraman@google.com>
2460 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
2461 * cfgexpand.c (defer_stack_allocation): ...use here
2462 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
2464 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
2466 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
2467 (operand_equal_for_value_replacement): New function, extracted from
2468 value_replacement and enhanced to catch more cases.
2469 (value_replacement): Use operand_equal_for_value_replacement.
2471 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2473 * loop-doloop.c (doloop_modify, doloop_optimize): Use
2474 get_max_loop_iterations.
2476 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2478 * config/arm/aarch-common.c (arm_early_load_addr_dep):
2479 Place comment above function.
2481 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2483 * tree-flow.h: Remove all remaining prototypes, enums and structs that
2484 are not related to tree-cfg.c.
2485 * tree-ssa-address.h: New file. Relocate prototypes.
2486 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
2487 (addr_for_mem_ref): New. Combine call to get_address_description and
2488 return addr_for_mem_ref.
2489 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
2490 * tree-ssa-live.h: Adjust prototypes.
2491 * passes.c: Include tree-ssa-live.h.
2492 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
2493 * graphite.c (graphite_transform_loops): Make static.
2494 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2495 make_pass_graphite, pass_data_graphite_transforms,
2496 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
2497 * ipa-pure-const.c (warn_function_noreturn): Make static.
2498 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2499 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2500 Relocate from tree-cfg.c
2501 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
2503 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2504 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2505 Move to ipa-pure-const.c.
2506 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
2507 Relocate from tree-optimize.c.
2508 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
2509 make_pass_fixup_cfg): Move to tree-cfg.c.
2510 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
2511 Relocate some prototypes.
2512 * tree-data-ref.h (tree_check_data_deps) Add prototype.
2513 * tree-dump.c (dump_function_to_file): Remove prototype.
2514 Add tree-flow.h to the include file.
2515 * tree-dump.h: Remove prototype.
2516 * tree-parloops.h: New File. Add prototypes.
2517 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
2518 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
2519 from tree-ssa-loop.c.
2520 * tree-predcom.c (run_tree_predictive_commoning,
2521 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
2522 Relocate here from tree-ssa-loop.c.
2523 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
2524 ssa_name_values.release ().
2525 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
2526 (ssa_name_values): Relocate from tree-flow.h.
2527 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
2528 * tree-ssa-loop.c (run_tree_predictive_commoning,
2529 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
2530 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2531 make_pass_graphite, pass_data_graphite_transforms,
2532 make_pass_graphite_transforms, gate_tree_parallelize_loops,
2533 tree_parallelize_loops, pass_data_parallelize_loops,
2534 make_pass_parallelize_loops): Move to other files.
2535 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
2537 * tree.h: Remove prototypes from tree-address.c.
2539 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2541 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
2542 (struct int_tree_map): Move to tree-hasher.h
2543 (SCALE, LABEL, PERCENT): Move to gimple.h
2544 * tree-flow-inline.h: Delete. Move functions to other files.
2545 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
2546 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
2547 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
2548 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
2549 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
2550 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
2551 * tree-hasher.h: Don't include tree-flow.h.
2552 (struct int_tree_map): Relocate from tree-flow.h.
2553 * tree-sra.c (contains_view_convert_expr_p): Relocate from
2554 tree-flow-inline.h and make static.
2555 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
2557 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
2558 tree-flow-inline.h and make static.
2559 * tree.h (is_global_var, may_be_aliased): Relocate from
2561 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
2562 * value-prof.c: No longer include tree-flow-inline.h.
2563 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
2565 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2567 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
2568 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
2569 (enum move_pos): Move to tree-ssa-loop-im.h
2570 * cfgloop.h: Move some prototypes.
2571 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
2572 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
2573 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
2574 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
2575 (loop_containing_stmt): Relocate from tree-flow-inline.h.
2576 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
2577 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
2578 (enum move_pos): Relocate here.
2579 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
2581 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
2582 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
2583 make_pass_lim): Relocate here from tree-ssa-loop.c.
2584 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
2586 (loop_edge_to_cancel, unloop_loops): Make static.
2587 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2588 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
2589 (tree_complete_unroll, gate_tree_complete_unroll,
2590 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
2591 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2592 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
2593 * tree-ssa-loop-ivopts.c: Remove local prototypes.
2594 (stmt_invariant_in_loop_p): Remove unused function.
2595 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
2596 * tree-ssa-loop-manip.h: New file. Add prototypes.
2597 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
2598 (gcov_type_to_double_int): Move to cfgloop.h.
2599 (double_int_cmp, bound_index,
2600 estimate_numbers_of_iterations_loop): Make static.
2601 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
2602 (max_loop_iterations): Factor out get_max_loop_iterations.
2603 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
2605 * tree-ssa-loop-niter.h: New file. Add prototypes.
2606 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
2607 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
2608 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
2609 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
2610 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
2611 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
2612 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
2613 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
2614 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
2615 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
2616 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2617 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
2618 pass_data_complete_unroll, make_pass_complete_unroll,
2619 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2620 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
2621 tree-ssa-loop-ivcanon.c.
2622 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
2623 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
2624 tree-ssa-loop-prefetch.c.
2625 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
2627 (get_lsm_tmp_name): Relocate and add suffix parameter.
2628 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
2629 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
2630 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
2631 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
2632 (get_estimated_loop_iterations): Factor out accessor from
2633 estimated_loop_iterations in tree-ssa-loop-niter.c.
2634 (get_max_loop_iterations): Factor out accessor from
2635 _max_loop_iterations in tree-ssa-niter.c.
2636 * loop-unroll.c (decide_unroll_constant_iterations,
2637 decide_unroll_runtime_iterations, decide_peel_simple,
2638 decide_unroll_stupid): Use new get_* accessors.
2640 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
2642 PR tree-optimization/20318
2643 * doc/extend.texi (returns_nonnull): New function attribute.
2644 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
2646 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
2647 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
2649 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
2652 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
2653 false if both components are bitfields.
2655 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2657 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
2658 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
2659 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
2660 * config/aarch64/aarch64.h
2661 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
2662 * config/aarch64/aarch64-simd-builtins.def
2663 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
2665 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2667 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
2669 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2671 * config/aarch64/arm_neon.h (vdiv_f64): Added.
2673 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2675 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
2676 (vneg_f64): New intrinsic.
2677 (vneg_s8): Asm replaced with C.
2678 (vneg_s16): Likewise.
2679 (vneg_s32): Likewise.
2680 (vneg_s64): New intrinsic.
2681 (vnegq_f32): Asm replaced with C.
2682 (vnegq_f64): Likewise.
2683 (vnegq_s8): Likewise.
2684 (vnegq_s16): Likewise.
2685 (vnegq_s32): Likewise.
2686 (vnegq_s64): Likewise.
2688 2013-10-09 Renlin Li <Renlin.Li@arm.com>
2690 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
2692 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2694 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
2695 packed stack special handling.
2696 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
2697 back to fixed stack slots for FPRs saved due to stdarg.
2699 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2701 * config/s390/s390.c (s390_frame_info): Restructure function.
2703 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2705 * config/s390/s390.c (struct s390_frame_layout): New field
2707 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
2708 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
2709 regs_ever_clobbered to char*.
2710 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
2711 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
2712 (s390_register_info): Enable FPR save slots. Move/Copy some
2713 functionality into ...
2714 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
2715 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
2717 (s390_frame_info): Do gpr slot allocation here now. stdarg does
2718 not imply a stack frame.
2719 (s390_init_frame_layout): Remove variable clobbered_regs.
2720 (s390_update_register_info): Remove function.
2721 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
2722 cannot be used for register renaming.
2723 (s390_hard_regno_scratch_ok): New function.
2724 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
2725 (s390_initial_elimination_offset): Change offset calculation of
2726 the return address pointer.
2727 (save_gprs): Deal with only r6 being saved from the call-saved regs.
2728 (restore_gprs): Set frame related flag.
2729 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
2730 (s390_emit_prologue): Call s390_register_info instead of
2731 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
2732 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
2733 (s390_optimize_prologue): Call s390_optimize_register_info.
2734 Try to remove also FPR slot save/restore INSNs. Remove frame
2735 related flags from restore INSNs.
2737 2013-10-08 DJ Delorie <dj@redhat.com>
2739 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
2742 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
2743 avoid conflict with the MI use of %c.
2744 * config/rl78/rl78-real.md: change %c to %C throughout.
2745 * config/rl78/rl78-virt.md: Likewise.
2747 2013-10-08 Jan Hubicka <jh@suse.cz>
2749 * config/i386/i386.c (ix86_option_override_internal): Switch
2750 to SSE math for -ffast-math when target ISA supports SSE2.
2752 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
2754 * tree-flow.h: Remove some prototypes.
2755 * tree.h: Remove some protypes, add a couple.
2756 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
2757 using_eh_for_cleanups_p): Add interface routines for front ends.
2758 * tree-eh.h: New file. Add protoptyes.
2759 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
2760 (add_stmt_to_eh_lp_fn): Make static.
2761 (lower_try_finally): Use new using_eh_for_cleanups_p.
2762 * emit-rtl.c: Include tree-eh.h.
2763 * gimple.h: Include tree-eh.h.
2765 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
2767 PR tree-optimization/58480
2768 * tree-vrp.c (infer_nonnull_range): New function.
2769 (infer_value_range): Call infer_nonnull_range.
2771 2013-10-08 Dehao Chen <dehao@google.com>
2773 PR tree-optimization/58619
2774 * tree-inline.c (copy_phis_for_bb): Combine location data
2777 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2780 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
2781 RTX_FRAME_RELATED_P on INSN.
2783 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2785 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
2786 (altivec_expand_vec_perm_const): Call it.
2788 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2790 * config/rs6000/vector.md (mov<mode>): Emit permuted move
2791 sequences for LE VSX loads and stores at expand time.
2792 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
2794 * config/rs6000/rs6000.c (rs6000_const_vec): New.
2795 (rs6000_gen_le_vsx_permute): New.
2796 (rs6000_gen_le_vsx_load): New.
2797 (rs6000_gen_le_vsx_store): New.
2798 (rs6000_gen_le_vsx_move): New.
2799 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
2800 (*vsx_le_perm_load_v4si): New.
2801 (*vsx_le_perm_load_v8hi): New.
2802 (*vsx_le_perm_load_v16qi): New.
2803 (*vsx_le_perm_store_v2di): New.
2804 (*vsx_le_perm_store_v4si): New.
2805 (*vsx_le_perm_store_v8hi): New.
2806 (*vsx_le_perm_store_v16qi): New.
2807 (*vsx_xxpermdi2_le_<mode>): New.
2808 (*vsx_xxpermdi4_le_<mode>): New.
2809 (*vsx_xxpermdi8_le_V8HI): New.
2810 (*vsx_xxpermdi16_le_V16QI): New.
2811 (*vsx_lxvd2x2_le_<mode>): New.
2812 (*vsx_lxvd2x4_le_<mode>): New.
2813 (*vsx_lxvd2x8_le_V8HI): New.
2814 (*vsx_lxvd2x16_le_V16QI): New.
2815 (*vsx_stxvd2x2_le_<mode>): New.
2816 (*vsx_stxvd2x4_le_<mode>): New.
2817 (*vsx_stxvd2x8_le_V8HI): New.
2818 (*vsx_stxvd2x16_le_V16QI): New.
2820 2013-10-07 Renlin Li <Renlin.Li@arm.com>
2822 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
2824 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2826 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
2827 loop to work also for 31bit ABI.
2828 Save the stack pointer for frame_size > 0.
2830 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2832 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
2833 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
2834 constraint letters from expanders.
2835 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
2836 retry count to general_operand.
2837 ("tabort"): Give operand 0 a mode.
2838 ("tabort_1"): Add mode and constraint letter for operand 0.
2839 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
2841 2013-10-04 Jeff Law <law@redhat.com>
2843 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
2845 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
2847 (thread_across_edge): Here. Call it.
2849 2013-10-04 Cary Coutant <ccoutant@google.com>
2851 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
2852 discarding a location list expression (or a piece of one).
2854 2013-10-03 Jan Hubicka <jh@suse.cz>
2856 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
2857 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
2858 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
2860 2013-10-03 Jan Hubicka <jh@suse.cz>
2862 * config/i386/i386.c (ix86_option_override_internal): Do not enable
2863 accumulate-outgoing-args when producing unwind info.
2865 2013-10-03 Wei Mi <wmi@google.com>
2867 * lra-constraints.c (insert_move_for_subreg): New function
2868 extracted from simplify_operand_subreg.
2869 (simplify_operand_subreg): Add reload for paradoxical subreg.
2871 2013-10-03 Rong Xu <xur@google.com>
2873 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
2874 the candidate of builtin_expect such that we should fix the
2875 size/time estimation.
2876 (estimate_function_body_sizes): Do the acutally size/time fix-up
2879 2013-10-03 Rong Xu <xur@google.com>
2881 * predict.c (tree_predict_by_opcode): Get the probability
2882 for builtin_expect from param builtin_expect_probability.
2883 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
2884 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
2885 * doc/invoke.texi: Add documentation for builtin-expect-probability.
2887 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
2890 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
2891 * calls.c (alloca_call_p): Use get_callee_fndecl.
2892 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
2893 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
2895 (vrp_visit_stmt): Remove duplicated code.
2897 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
2899 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
2900 ceildf2, btruncdf2, instead of vsx_* name.
2902 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
2903 iterators to only do V2DF and V4SF here. Move the DF code to
2904 rs6000.md where it is combined with SF mode. Replace <VSv> with
2905 just 'v' since only vector operations are handled with these insns
2906 after moving the DF support to rs6000.md.
2907 (vsx_sub<mode>3): Likewise.
2908 (vsx_mul<mode>3): Likewise.
2909 (vsx_div<mode>3): Likewise.
2910 (vsx_fre<mode>2): Likewise.
2911 (vsx_neg<mode>2): Likewise.
2912 (vsx_abs<mode>2): Likewise.
2913 (vsx_nabs<mode>2): Likewise.
2914 (vsx_smax<mode>3): Likewise.
2915 (vsx_smin<mode>3): Likewise.
2916 (vsx_sqrt<mode>2): Likewise.
2917 (vsx_rsqrte<mode>2): Likewise.
2918 (vsx_fms<mode>4): Likewise.
2919 (vsx_nfma<mode>4): Likewise.
2920 (vsx_copysign<mode>3): Likewise.
2921 (vsx_btrunc<mode>2): Likewise.
2922 (vsx_floor<mode>2): Likewise.
2923 (vsx_ceil<mode>2): Likewise.
2924 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
2925 (vsx_sminsf3): Likewise.
2926 (vsx_fmadf4): Likewise.
2927 (vsx_fmsdf4): Likewise.
2928 (vsx_nfmadf4): Likewise.
2929 (vsx_nfmsdf4): Likewise.
2930 (vsx_cmpdf_internal1): Likewise.
2932 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
2933 simpler to select whether a target has SPE or traditional floating
2934 point support in iterators.
2935 (TARGET_DF_SPE): Likewise.
2936 (TARGET_SF_FPR): Likewise.
2937 (TARGET_DF_FPR): Likewise.
2938 (TARGET_SF_INSN): Macros to say whether floating point support
2939 exists for a given operation for expanders.
2940 (TARGET_DF_INSN): Likewise.
2942 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
2943 combining of SF/DF mode operations, using both traditional and VSX
2951 (abs<mode>2): Combine SF/DF modes using traditional floating point
2952 instructions. Add support for using the upper DF registers with
2953 VSX support, and SF registers with power8-vector support. Update
2954 expanders for operations supported by both the SPE and traditional
2955 floating point units.
2956 (abs<mode>2_fpr): Likewise.
2957 (nabs<mode>2): Likewise.
2958 (nabs<mode>2_fpr): Likewise.
2959 (neg<mode>2): Likewise.
2960 (neg<mode>2_fpr): Likewise.
2961 (add<mode>3): Likewise.
2962 (add<mode>3_fpr): Likewise.
2963 (sub<mode>3): Likewise.
2964 (sub<mode>3_fpr): Likewise.
2965 (mul<mode>3): Likewise.
2966 (mul<mode>3_fpr): Likewise.
2967 (div<mode>3): Likewise.
2968 (div<mode>3_fpr): Likewise.
2969 (sqrt<mode>3): Likewise.
2970 (sqrt<mode>3_fpr): Likewise.
2971 (fre<Fs>): Likewise.
2972 (rsqrt<mode>2): Likewise.
2973 (cmp<mode>_fpr): Likewise.
2974 (smax<mode>3): Likewise.
2975 (smin<mode>3): Likewise.
2976 (smax<mode>3_vsx): Likewise.
2977 (smin<mode>3_vsx): Likewise.
2978 (negsf2): Delete SF operations that are merged with DF.
2985 (fmasf4_fpr): Likewise.
2986 (fmssf4_fpr): Likewise.
2987 (nfmasf4_fpr): Likewise.
2988 (nfmssf4_fpr): Likewise.
2989 (sqrtsf2): Likewise.
2990 (rsqrtsf_internal1): Likewise.
2991 (smaxsf3): Likewise.
2992 (sminsf3): Likewise.
2993 (cmpsf_internal1): Likewise.
2994 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
2995 (negdf2): Delete DF operations that are merged with SF.
2997 (nabsdf2): Likewise.
3003 (rsqrtdf_internal1): Likewise.
3004 (fmadf4_fpr): Likewise.
3005 (fmsdf4_fpr): Likewise.
3006 (nfmadf4_fpr): Likewise.
3007 (nfmsdf4_fpr): Likewise.
3008 (sqrtdf2): Likewise.
3009 (smaxdf3): Likewise.
3010 (smindf3): Likewise.
3011 (cmpdf_internal1): Likewise.
3012 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
3013 (btrunc<mode>2): Delete separate expander, and combine with the
3014 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
3015 (btrunc<mode>2_fpr): Likewise.
3016 (ceil<mode>2): Likewise.
3017 (ceil<mode>2_fpr): Likewise.
3018 (floor<mode>2): Likewise.
3019 (floor<mode>2_fpr): Likewise.
3020 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
3021 Add support for using the upper registers with VSX and
3022 power8-vector. Move insns to be closer to the define_expands. On
3023 VSX systems, prefer the traditional form of FMA over the VSX
3024 version, since the traditional form allows the target not to
3025 overlap with the inputs.
3026 (fms<mode>4_fpr): Likewise.
3027 (nfma<mode>4_fpr): Likewise.
3028 (nfms<mode>4_fpr): Likewise.
3030 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3031 Richard Earnshaw <richard.earnshaw@arm.com>
3033 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
3034 (struct mult_cost_table): Likewise.
3035 (struct mem_cost_table): Likewise.
3036 (struct fp_cost_table): Likewise.
3037 (struct vector_cost_table): Likewise.
3038 (cpu_cost_table): Likewise.
3039 * config/arm/arm.opt (mold-rts-costs): New option.
3040 (mnew-generic-costs): Likewise.
3041 * config/arm/arm.c (generic_extra_costs): New table.
3042 (cortexa15_extra_costs): Likewise.
3043 (arm_slowmul_tune): Use NULL as new costs.
3044 (arm_fastmul_tune): Likewise.
3045 (arm_strongarm_tune): Likewise.
3046 (arm_xscale_tune): Likewise.
3047 (arm_9e_tune): Likewise.
3048 (arm_v6t2_tune): Likewise.
3049 (arm_cortex_a5_tune): Likewise.
3050 (arm_cortex_a9_tune): Likewise.
3051 (arm_v6m_tune): Likewise.
3052 (arm_fa726te_tune): Likewise.
3053 (arm_cortex_a15_tune): Use cortex15_extra_costs.
3054 (arm_cortex_tune): Use generict_extra_costs.
3055 (shifter_op_p): New function.
3056 (arm_unspec_cost): Likewise.
3057 (LIBCALL_COST): Define.
3058 (arm_new_rtx_costs): New function.
3059 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
3060 table is available. Use old costs otherwise unless mnew-generic-costs
3062 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
3063 (cpu_cost_table): Declare.
3065 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
3068 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
3069 (*subs_mul_imm_<mode>)
3070 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
3071 (*sub_<shift>_<mode>)
3072 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
3073 Remove k constraint.
3075 2013-10-03 Ian Bolton <ian.bolton@arm.com>
3077 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
3079 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
3081 2013-10-02 Teresa Johnson <tejohnson@google.com>
3083 * predict.c (probably_never_executed): New function.
3084 (probably_never_executed_bb_p): Invoke probably_never_executed.
3085 (probably_never_executed_edge_p): Ditto.
3086 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
3087 Treat profile insanities conservatively.
3089 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
3091 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
3093 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
3095 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
3096 use it. Use smaller increase for scratch. Don't increase reject
3097 for early clobber scratch.
3098 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
3099 setting eliminated regs except setting fp from hfp.
3100 (lra_eliminate): Check lra_insn_recog_data on NULL.
3102 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3105 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
3106 setting -mvsx-timode by default until the underlying problem is fixed.
3107 (RS6000_CPU, power7 defaults): Likewise.
3109 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3111 * config/x-linux (host-linux.o): Remove header dependencies.
3112 Use $(COMPILE) and $(POSTCOMPILE).
3113 * config/t-linux-android (linux-android.o): Ditto.
3115 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3117 * Makefile.in (expmed.o-warn): Remove.
3119 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3121 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
3122 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
3124 2013-10-02 Teresa Johnson <tejohnson@google.com>
3126 * dojump.c (do_jump_1): Divide probability between
3127 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
3129 2013-10-02 Tom Tromey <tromey@redhat.com>
3131 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
3133 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3135 * tree-flow.h: Remove some prototypes.
3136 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
3137 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
3138 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
3139 mark_virtual_phi_result_for_renaming): Relocate here.
3140 * tree-into-ssa.h: Add prototypes.
3141 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
3142 single_pred_before_succ_order.
3143 (blocks_in_phiopt_order): Rename and move to cfganal.c.
3144 (nonfreeing_call_p) Move to gimple.c.
3145 * cfganal.c (single_pred_before_succ_order): Move and renamed from
3147 * basic-block.h (single_pred_before_succ_order): Add prototype.
3148 * gimple.c (nonfreeing_call_p): Relocate here.
3149 * gimple.h: Add prototype.
3150 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
3151 * tree-ssa-dom.h: New file. Relocate prototypes here.
3152 * tree-ssa.h: Include tree-ssa-dom.h.
3154 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3156 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
3157 Use $(COMPILE) and $(POSTCOMPILE).
3159 * config/alpha/x-alpha (driver-alpha.o): Ditto.
3161 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3163 * tree-flow.h: Remove some prototypes.
3164 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
3165 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3166 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
3167 propagate_tree_value*): Move from here to...
3168 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
3169 propagate_tree_value*): Relocate here.
3170 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3171 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
3172 * gimple-fold.c: Remove gimple-fold.h from include list.
3173 * tree-vrp.c: Remove gimple-fold.h from include list.
3174 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
3175 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
3176 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
3177 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
3178 * sese.c: Add tree-ssa-propagate.h to include list.
3180 2013-10-02 Richard Biener <rguenther@suse.de>
3182 * tree-loop-distribution.c: Include tree-vectorizer.h for
3184 (enum partition_kind): Remove PKIND_REDUCTION.
3185 (struct partition_s): Remove has_writes member, add reduction_p member.
3186 (partition_alloc): Adjust.
3187 (partition_builtin_p): Likewise.
3188 (partition_has_writes): Remove.
3189 (partition_reduction_p): New function.
3190 (partition_merge_into): Likewise.
3191 (generate_code_for_partition): Commonize builtin partition
3193 (rdg_cannot_recompute_vertex_p): Remove.
3194 (already_processed_vertex_p): Likewise.
3195 (rdg_flag_vertex): Do not set has_writes.
3196 (classify_partition): Adjust.
3197 (rdg_build_partitions): Do not set has_writes, treat all
3198 partitions as useful.
3199 (distribute_loop): Record number of library calls generated. Adjust.
3200 (tree_loop_distribution): Report number of loops and library
3201 calls generated as opt-info.
3203 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3205 * tree-flow.h: Include new .h files. Move prototypes.
3206 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
3207 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
3208 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
3209 * tree-pretty-print.h: Add prototypes from tree-flow.h.
3210 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
3211 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
3212 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
3213 * tree.h (get_ref_base_and_extent): Move prototype out.
3214 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
3216 * gimple-low.h: New File. Add prototypes from tree-flow.h.
3217 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
3218 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
3219 * tree-scalar-evolution.c: Include tree.h.
3220 * sese.c: Include tree.h.
3221 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
3222 * dwarf2out.c: Include tree-dfa.h.
3223 * tree-chrec.c: Include tree.h.
3224 * tree-data-ref.c: Include tree.h.
3226 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
3228 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
3231 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3233 * config/t-sol2 (sol2-c.o): Remove header dependencies.
3234 Use $(COMPILE) and $(POSTCOMPILE).
3235 (sol2-cxx.o): Likewise.
3236 (sol2-stubs.o): Likewise.
3238 * config/x-solaris (host-solaris.o): Likewise.
3240 * config/sparc/t-sparc (sparc.o): Remove.
3241 (sparc-c.o): Remove header dependencies.
3242 Use $(COMPILE) and $(POSTCOMPILE).
3243 * config/sparc/x-sparc: Likewise.
3245 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
3247 * config/arc/arc-opts.h: Add 2013 to Copyright years.
3248 * config/arc/arc700.md: Likewise.
3249 * config/arc/arc-modes.def: Likewise.
3250 * config/arc/arc-simd.h: Likewise.
3251 * config/arc/t-arc-uClibc: Likewise.
3252 * config/arc/t-arc-newlib: Likewise.
3254 2013-10-02 Renlin Li <renlin.li@arm.com>
3256 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
3258 (aarch64_expand_epilogue): Likewise.
3260 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3261 Yufeng Zhang <yufeng.zhang@arm.com>
3263 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
3265 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
3266 'base_in' represent a conversion and legal_cast_p_1 holds; set
3267 'base_in' with the returned value from get_unwidened.
3269 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3271 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
3272 plus_constant is not used.
3274 2013-10-01 Wei Mi <wmi@google.com>
3276 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
3277 * config/i386/i386.md: Add define_peephole2 to
3278 break partial reg stall for cvtss2sd/cvtsd2ss.
3280 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3282 * config/arc/arc.c (pass_arc_ifcvt::clone):
3283 Update for ctxt_ -> m_ctxt change.
3285 2013-10-01 Jeff Law <law@redhat.com>
3287 * tree-ssa-threadupdate.c (struct redirection_data): Delete
3288 outgoing_edge and intermediate_edge fields. Instead store the path.
3289 (redirection_data::hash): Hash on the last edge's destination index.
3290 (redirection_data::equal): Check the entire thread path.
3291 (lookup_redirectio_data): Corresponding changes.
3292 (create_edge_and_update_destination_phis): Likewise.
3293 (thread_single_edge): Likewise.
3295 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3296 Diego Novillo <dnovillo@google.com>
3298 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
3299 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
3300 (vld32wh_insn, vld32wl_insn): Delete commented-out old
3301 versions of these patterns.
3303 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
3304 (__builtin_arc_aligned): Likewise.
3306 * config/arc/arc.md: Expand adc_0 comment stating the intended
3307 purpose and why it isn't ready.
3308 Replace commented out call_value_via_label_mixed with a
3309 plain comment about bl_s.
3311 * config/arc/arc.c (stdio.h): Don't include directly.
3312 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
3313 Remove if (1) condition.
3314 (arc_encode_section_info): Fix comment.
3316 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3318 * config/arc/arc.c (arc_conditional_register_usage):
3319 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
3320 Also set reg_alloc_order for DMA config regs.
3322 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3323 Jeremy Bennett <jeremy.bennett@embecosm.com>
3325 * doc/install.texi (--with-cpu): Mention ARC.
3326 (arc-*-elf32): New paragraph.
3327 (arc-linux-uclibc): Likewise.
3328 * doc/md.texi (Machine Constraints): Add ARC part.
3329 * doc/invoke.texi: (menu): Add ARC Options.
3330 (Machine Dependent Options) <ARC Options>: Add synopsis.
3331 (node ARC Options): Add.
3332 * doc/extend.texi (long_call / short_call attribute): Add ARC.
3333 (ARC Built-in Functions): New section defining
3334 generic ARC built-in functions.
3335 (ARC SIMD Built-in Functions): New section defining SIMD specific
3337 (Declaring Attributes of Functions): Extended
3338 description of short_call and long_call attributes for ARC and
3339 added index entries.
3341 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
3342 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
3343 Joern Rennecke <joern.rennecke@embecosm.com>
3344 Muhammad Khurram Riaz <khurram.riaz@arc.com>
3345 Brendan Kehoe <brendan@zen.org>
3346 Michael Eager <eager@eagercon.com>
3347 Simon Cook <simon.cook@embecosm.com>
3348 Jeremy Bennett <jeremy.bennett@embecosm.com>
3350 * config/arc, common/config/arc: New directories.
3352 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3353 Brendan Kehoe <brendan@zen.org>
3354 Simon Cook <simon.cook@embecosm.com>
3356 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
3358 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3360 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
3361 * tree-ssa-coalesce.h: New. Move prototype to here.
3362 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
3363 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
3364 (gimple_can_coalesce_p): Move to...
3365 * gimple.c (gimple_can_coalesce_p): Here.
3367 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3369 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
3370 (dump_decl_set): Move to gimple.c.
3371 * gimple.h: Don't include tree-ssa-operands.h.
3372 (dump_decl_set): Add prototype.
3373 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3374 Move to gimple-ssa.h.
3375 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
3376 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
3377 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
3378 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
3379 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
3381 (dump_decl_set): Relocate here.
3382 * gimple-ssa.h: New file.
3383 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3384 Relocate from gimple.h.
3385 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
3386 * tree-ssa-operands.c (swap_ssa_operands): Rename from
3387 swap_tree_operands and remove non-ssa path.
3388 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
3389 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
3391 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
3392 vect_is_simple_reduction_1): Use swap_ssa_operands.
3393 * tree-flow.h: Move various prototypes to tree-phinodes.h.
3394 (enum need_phi_state): Move to tree-into-ssa.c.
3395 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3396 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
3397 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
3398 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
3399 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3400 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3401 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3402 num_imm_uses): Move to ssa-iterators.h.
3403 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
3404 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
3406 (op_iter_done, op_iter_next_def, op_iter_next_tree,
3407 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
3408 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
3409 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
3410 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
3411 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
3412 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
3413 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
3414 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
3415 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
3416 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
3417 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
3418 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
3419 (set_phi_nodes): Move to tree-phinodes.h.
3420 * tree-ssa-operands.h (enum ssa_op_iter_type,
3421 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
3422 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
3423 (dump_decl_set): Remove prototype.
3424 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
3425 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
3426 (set_phi_nodes): Relocate from tree-flow-inline.h.
3427 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
3429 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
3430 include list. Temporarily add gimple.h to include list.
3431 * ssa-iterators.h: New file.
3432 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3433 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
3434 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
3435 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
3436 Relocate from tree-ssa-operands.h.
3437 (delink_imm_use, link_imm_use_to_list, link_imm_use,
3438 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3439 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3440 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3441 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
3442 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
3443 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
3444 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
3445 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
3446 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
3447 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
3448 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
3449 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
3450 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
3451 Relocate from tree-flow-inline.h.
3452 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
3454 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
3457 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
3458 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
3459 Insert '\t' to output template.
3460 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
3461 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
3462 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
3463 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
3465 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
3467 * doc/install.texi (Host/target specific installation notes for GCC):
3468 Put @anchor before @heading.
3469 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
3470 Use @email for email addresses.
3472 2013-10-01 Jeff Law <law@redhat.com>
3474 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
3475 a vec. Only delete the path if we create one without successfully
3476 registering a jump thread.
3477 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
3479 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
3480 (paths): New vector of jump threading paths.
3481 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
3482 (THREAD_PATH): New accessor macro for the entire thread path.
3483 (lookup_redirection_data): Get intermediate and final outgoing edge
3484 from the thread path.
3485 (create_edge_and_update_destination_phis): Copy the threading path.
3486 (ssa_fix_duplicate_block_edges): Get edges and block types from the
3487 jump threading path.
3488 (ssa_redirect_edges): Get edges and block types from the jump threading
3489 path. Free the path vector.
3490 (thread_block): Get edges from the jump threading path. Look at the
3491 entire path to see if we thread to a loop exit. If we cancel a jump
3492 thread request, then free the path vector.
3493 (thread_single_edge): Get edges and block types from the jump threading
3494 path. Free the path vector.
3495 (thread_through_loop_header): Get edges and block types from the jump
3496 threading path. Free the path vector.
3497 (mark_threaded_blocks): Iterate over the vector of paths and store
3498 the path on the appropriate edge. Get edges and block types from the
3499 jump threading path.
3500 (mark_threaded_blocks): Get edges and block types from the jump
3501 threading path. Free the path vector.
3502 (thread_through_all_blocks): Use the vector of paths rather than
3503 a vector of 3-edge sets.
3504 (register_jump_thread): Accept pointer to a path vector rather
3505 than the path vector itself. Store the path vector for later use.
3508 2013-10-01 Jakub Jelinek <jakub@redhat.com>
3509 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3512 * config/s390/s390.c (s390_split_branches): Modify check for table
3514 (s390_chunkify_start): Rearrange table jump insn check in order to
3515 deal with compare and branch insns correctly.
3517 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
3521 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
3522 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
3523 define_insn_and_split.
3524 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
3525 (shiftsi3_compare): New pattern.
3527 * config/arm/unspecs.md (UNSPEC_RRX): New.
3529 2013-10-01 Alan Modra <amodra@gmail.com>
3531 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
3532 casing inout operands.
3534 2013-10-01 Richard Biener <rguenther@suse.de>
3536 PR tree-optimization/58553
3537 * tree-loop-distribution.c (struct partition_s): Add niter member.
3538 (classify_partition): Populate niter member for the partition
3539 and properly identify whether the relevant store happens before
3540 or after the loop exit.
3541 (generate_memset_builtin): Use niter member from the partition.
3542 (generate_memcpy_builtin): Likewise.
3544 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3546 * vec.h (vec_prefix, vec): Prefix member names with "m_".
3547 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
3549 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3551 * basic-block.h (edge_list): Prefix member names with "m_".
3552 * context.h (context): Likewise.
3553 * domwalk.h (dom_walker): Likewise.
3554 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
3555 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
3556 * hash-table.h (hash_table): Likewise.
3557 * machmode.h (bit_field_mode_iterator): Likewise.
3558 * pass_manager.h (pass_list): Likewise.
3559 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
3560 * tree-pass.h (pass_data): Likewise.
3561 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
3562 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
3563 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
3564 * asan.c (pass_data_asan): Update accordingly.
3565 * cfganal.c (control_dependences::find_control_dependence): Likewise.
3566 (control_dependences::control_dependences): Likewise.
3567 (control_dependences::~control_dependences): Likewise.
3568 (control_dependences::~control_dependences): Likewise.
3569 (control_dependences::get_edges_dependent_on): Likewise.
3570 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
3571 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
3572 * context.c (gcc::context::context): Likewise.
3573 * cprop.c (pass_rtl_cprop::clone): Likewise.
3574 * domwalk.c (dom_walker::walk): Likewise.
3575 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
3576 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
3577 * mode-switching.c (pass_mode_switching::clone): Likewise.
3578 * passes.c (opt_pass::opt_pass): Likewise.
3579 (pass_manager::pass_manager): Likewise.
3580 * predict.c (pass_strip_predict_hints::clone): Likewise.
3581 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
3582 (pass_split_all_insns::clone): Likewise.
3583 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
3585 (bit_field_mode_iterator::next_mode): Likewise.
3586 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
3587 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
3588 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
3589 * tree-complex.c (pass_lower_complex::clone): Likewise.
3590 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
3591 * tree-object-size.c (pass_object_sizes::clone): Likewise.
3592 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
3593 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
3594 (pass_fold_builtins::clone): Likewise.
3595 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
3596 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
3597 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
3598 (pass_cd_dce::clone): Likewise.
3599 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
3600 (pass_phi_only_cprop::clone): Likewise.
3601 * tree-ssa-dse.c (pass_dse::clone): Likewise.
3602 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
3603 * tree-ssa-loop.c (pass_lim::clone): Likewise.
3604 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
3605 * tree-ssa-pre.c (pass_fre::clone): Likewise.
3606 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
3607 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
3608 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
3609 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
3610 * tree-vrp.c (pass_vrp::clone): Likewise.
3611 * tsan.c (pass_tsan::clone): Likewise.
3613 2013-09-30 Jakub Jelinek <jakub@redhat.com>
3616 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
3617 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
3620 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
3621 optimization, punt if sign_bit_p looked through any zero extension.
3623 2013-09-30 Teresa Johnson <tejohnson@google.com>
3625 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
3626 Update redirected out edge count in joiner case.
3627 (ssa_redirect_edges): Common the joiner and non-joiner cases
3628 so that joiner case gets profile updates.
3630 2013-09-30 Richard Biener <rguenther@suse.de>
3632 PR tree-optimization/58554
3633 * tree-loop-distribution.c (classify_partition): Require
3634 unconditionally executed stores for memcpy and memset recognition.
3635 (tree_loop_distribution): Calculate dominance info.
3637 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3639 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
3640 (NO_PROFILE_COUNTERS): Likewise.
3641 (PROFILE_HOOK): Likewise.
3642 (FUNCTION_PROFILER): Likewise.
3643 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
3645 2013-09-30 Iain Sandoe <iain@codesourcery.com>
3647 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
3648 calls and defines in TARGET_MACHO conditional.
3649 (load_macho_picbase_di): Likewise.
3650 (reload_macho_picbase): Likewise.
3651 (reload_macho_picbase_si): Likewise.
3652 (reload_macho_picbase_di): Likewise.
3653 (nonlocal_goto_receiver): Likewise.
3655 2013-09-30 Nick Clifton <nickc@redhat.com>
3657 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
3658 that use the MSP430X ISA.
3659 (msp430_option_override): Scan -mmcu command line option for any
3660 MCU name that supports the MSP430X ISA.
3661 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
3662 -mmcu options which enable the MSP430X ISA.
3664 2013-09-30 Richard Biener <rguenther@suse.de>
3667 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
3668 before looking for setjmp-like calls.
3670 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3673 * config/darwin-protos.h (machopic_get_function_picbase): New.
3674 * config/darwin.c (machopic_get_function_picbase): New.
3675 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
3676 label for a new func. (load_macho_picbase_di): Likewise.
3677 (reload_macho_picbase): New expand.
3678 (reload_macho_picbase_si): New insn.
3679 (reload_macho_picbase_di): New insn.
3680 (nonlocal_goto_receiver): New define and split.
3681 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
3682 (unspecv enum): Add UNSPECV_NLGR.
3684 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3686 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
3687 that altivec registers are correctly sized on Darwin.
3689 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3691 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
3692 darwin-driver.o): Use COMPILE and POSTCOMPILE.
3693 * config/x-darwin (host-darwin.o): Likewise.
3694 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
3695 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
3696 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
3698 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
3700 * doc/invoke.texi: Fix usage of @tie{} command.
3702 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
3704 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
3706 2013-09-28 Jan Hubicka <jh@suse.cz>
3708 * config/i386/x86-tune.def: Add documentation for each of the options;
3711 2013-09-28 Jan Hubicka <jh@suse.cz>
3713 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
3715 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
3716 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
3717 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
3718 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
3719 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
3721 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
3723 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
3724 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
3725 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
3726 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
3727 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
3728 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
3729 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
3730 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
3731 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
3732 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
3733 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
3734 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
3735 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
3736 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
3737 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
3738 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
3739 gimple.h, godump.c, graphite-clast-to-gimple.c,
3740 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
3741 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
3742 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
3743 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3744 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
3745 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
3746 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
3747 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
3748 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
3749 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
3750 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
3751 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
3752 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
3753 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
3754 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
3755 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
3756 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
3757 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
3758 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
3759 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
3760 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
3761 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
3762 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
3763 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
3764 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3765 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3766 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
3767 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
3768 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
3769 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3770 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3771 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
3772 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
3773 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3774 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
3775 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
3776 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
3777 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
3778 whitespace before "(".
3780 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
3782 * expr.h (extract_bit_field): Remove packedp parameter.
3783 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
3784 from forward declaration.
3785 (store_split_bit_field): Remove packedp arg from calls to
3786 extract_fixed_bit_field.
3787 (extract_bit_field_1): Remove packedp parameter and packedp
3788 argument from recursive calls and calls to extract_fixed_bit_field.
3789 (extract_bit_field): Remove packedp parameter and corresponding
3790 arg to extract_bit_field_1.
3791 (extract_fixed_bit_field): Remove packedp parameter. Remove code
3793 (extract_split_bit_field): Remove packedp arg from call to
3794 extract_fixed_bit_field.
3795 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
3796 (copy_blkmode_from_reg): Likewise.
3797 (copy_blkmode_to_reg): Likewise.
3798 (read_complex_part): Likewise.
3799 (store_field): Likewise.
3800 (expand_expr_real_1): Likewise.
3801 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
3802 to extract_bit_field.
3803 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
3804 call to extract_bit_field.
3805 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
3806 call to extract_bit_field.
3807 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
3808 and special packedp behavior from -fstrict-volatile-bitfields
3811 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3813 * lra-eliminations.c (init_elim_table): Guard value_p.
3815 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3817 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
3818 DFmode, DImode, and SFmode in the upper VSX registers based on the
3819 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
3820 if -mpower8-vector. Combine -mvsx-timode handling with the rest
3821 of the VSX register handling.
3823 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
3825 (zero_extendsidi2_lfiwzx): Add support for loading into the
3826 Altivec registers with -mpower8-vector. Use wu/wv constraints to
3827 only do VSX memory options on Altivec registers.
3828 (extendsidi2_lfiwax): Likewise.
3829 (extendsfdf2_fpr): Likewise.
3830 (mov<mode>_hardfloat, SF/SD modes): Likewise.
3831 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
3832 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
3833 (movdi_internal64): Likewise.
3835 2013-09-27 Xinliang David Li <davidxl@google.com>
3837 * opts.c (finish_options): Adjust parameters
3838 according to vect cost model.
3839 (common_handle_option): Set dynamic vect cost
3841 targhooks.c (default_add_stmt_cost): Compute stmt cost
3843 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
3844 Use helper function.
3845 * tree-vectorizer.h (unlimited_cost_model): New function.
3846 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
3847 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
3849 (vect_enhance_data_refs_alignment): Ditto.
3850 * flag-types.h: New enum.
3851 * common/config/i386/i386-common.c (ix86_option_init_struct):
3852 No need to initialize vect_cost_model flag.
3853 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
3856 2013-09-27 Diego Novillo <dnovillo@google.com>
3858 * gimple.h (enum ssa_mode): Remove.
3860 2013-09-27 Paulo Matos <pmatos@broadcom.com>
3862 * cfgloop.h (number_of_loops): Fix typo in check for null.
3864 2013-09-27 Jakub Jelinek <jakub@redhat.com>
3867 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
3868 are children of outermost saved_cfun's loop, and set it up to
3869 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
3870 if loop != loop0 and SESE region contains bbs that belong to loop0.
3872 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3874 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
3875 (binary_scale_code_p, get_base_term, get_index_term): New functions.
3876 (set_address_segment, set_address_base, set_address_index)
3877 (set_address_disp): Accept the argument unconditionally.
3878 (baseness): Remove must_be_base_p and must_be_index_p checks.
3879 (decompose_normal_address): Classify as much as possible in the
3882 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3884 * cse.c (count_reg_usage): Handle INT_LIST.
3885 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
3886 * reginfo.c (reg_scan_mark_refs): Likewise.
3887 * reload1.c (eliminate_regs_1): Likewise.
3889 2013-09-27 Iain Sandoe <iain@codesourcery.com>
3892 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
3895 2013-09-27 Richard Biener <rguenther@suse.de>
3897 PR tree-optimization/58459
3898 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
3899 restriction not propagating into loops.
3901 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
3903 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
3904 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
3905 * doc/tree-ssa.texi (Walking use-def chains): Delete.
3907 2013-09-26 Richard Biener <rguenther@suse.de>
3909 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
3911 2013-09-26 Richard Biener <rguenther@suse.de>
3913 * alias.h (component_uses_parent_alias_set): Rename to ...
3914 (component_uses_parent_alias_set_from): ... this.
3915 * alias.c (component_uses_parent_alias_set): Rename to ...
3916 (component_uses_parent_alias_set_from): ... this and return
3918 (reference_alias_ptr_type_1): Use the result from
3919 component_uses_parent_alias_set_from instead of stripping
3920 components one at a time.
3921 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
3923 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3925 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
3926 Move prototypes to...
3927 * tree-ssa-ter.h: New File. Move prototypes here.
3928 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
3929 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
3930 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
3931 from is_replaceable_p.
3932 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
3933 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
3934 (process_replaceable): Use ter_is_replaceable_p.
3935 (find_replaceable_in_bb): Use ter_is_replaceable_p.
3936 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
3937 newly refactored ssa_is_replaceable_p.
3938 * cfgexpand.c: Include tree-outof-ssa.h.
3939 * ssaexpand.h: Delete.
3941 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3943 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
3944 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
3946 (create_gimple_tmp): Delete.
3947 (get_expr_type, build_assign, build_type_cast): Move to...
3948 * gimple-builder.c: New File.
3949 (get_expr_type): Relocate from gimple.c.
3950 (build_assign, build_type_cast): Change to only create ssanames.
3951 * gimple.h: Move prototypes to...
3952 * gimple-builder.h: New File. Here.
3953 * tree-ssa.h: And here.
3954 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
3955 count_uses_and_derefs): Relocate from gimple.c.
3956 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
3957 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
3958 * tree-ssa-math-opts (execute_cse_reciprocals): Use
3959 gimple_replace_ssa_lhs.
3960 * asan.c: Include gimple-builder.h.
3961 * Makefile.in: Add gimple-builder.o.
3963 2013-09-26 Richard Biener <rguenther@suse.de>
3965 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
3966 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
3967 (loe_visit_block): Use gcc_checking_assert.
3968 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
3969 gimple_assign_ssa_name_copy_p.
3970 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
3973 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
3975 * config/rs6000/t-rs6000 (rs6000.o): Remove.
3976 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
3978 2013-09-26 Richard Biener <rguenther@suse.de>
3980 PR tree-optimization/58539
3981 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
3982 the fact that debug statements are not taking part in loop-closed
3985 2013-09-26 Nick Clifton <nickc@redhat.com>
3987 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
3988 time warning message.
3989 (msp430_print_operand_raw): Delete unused letter parameter.
3990 (TARGET_PRINT_OPERAND_ADDRESS): Define.
3991 (msp430_print_operand_address): New function.
3992 (msp430_print_operand): Move address printing code from here to
3994 * config/msp430/msp430.md (movsipsi2): Add comment in generated
3996 (zero_extendpsisi2): Likewise.
3997 (extendpsisi2): New pattern.
3998 (andneghi3): New pattern.
4000 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
4002 * config/aarch64/aarch64.opt (mlra): New option.
4003 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
4004 (TARGET_LRA_P): Define.
4006 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
4008 * expr.c (expand_assignment): Remove obsolete comment.
4010 2013-09-25 Jeff Law <law@redhat.com>
4012 * tree-flow.h (thread_through_all_blocks): Prototype moved into
4013 tree-ssa-threadupdate.h.
4014 (register_jump_thread): Similarly.
4015 * tree-ssa-threadupdate.h: New header file.
4016 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
4017 * tree-vrp.c: Likewise.
4018 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
4019 (thread_around_empty_blocks): Change type of path vector argument to
4020 an edge,type pair from just an edge. Initialize both elements when
4021 appending to a jump threading path. Tweak references to elements
4023 (thread_across_edge): Similarly. Release memory for the elements
4025 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
4026 (dump_jump_thread_path): New function broken out from
4027 register_jump_thread.
4028 (register_jump_thread): Use dump_jump_thread_path. Change type of
4029 path vector entries. Search the path for NULL edges and dump
4030 the path if one is found. Tweak the conversion of path to 3-edge
4031 form to use the block copy type information embedded in the path.
4033 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
4035 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
4037 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
4038 Vladimir Makarov <vmakarov@redhat.com>
4040 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
4041 from the least significant bit.
4042 (strip_address_mutations): Add bitfield operations handling.
4043 (must_be_index_p): Add shifting and rotate operations handling.
4044 (set_address_base): Use must_be_base_p predicate.
4045 (set_address_index): Use must_be_index_p predicate.
4047 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
4048 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4049 Sergey Lega <sergey.s.lega@intel.com>
4050 Anna Tikhonova <anna.tikhonova@intel.com>
4051 Ilya Tocar <ilya.tocar@intel.com>
4052 Andrey Turetskiy <andrey.turetskiy@intel.com>
4053 Ilya Verbin <ilya.verbin@intel.com>
4054 Kirill Yukhin <kirill.yukhin@intel.com>
4055 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4057 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
4059 (ix86_expand_vector_move_misalign): Support new unaligned load and
4060 stores and use new names.
4061 (CODE_FOR_sse2_storedqu): Rename to ...
4062 (CODE_FOR_sse2_storedquv16qi): ... this.
4063 (CODE_FOR_sse2_loaddqu): Rename to ...
4064 (CODE_FOR_sse2_loaddquv16qi): ... this.
4065 (CODE_FOR_avx_loaddqu256): Rename to ...
4066 (CODE_FOR_avx_loaddquv32qi): ... this.
4067 (CODE_FOR_avx_storedqu256): Rename to ...
4068 (CODE_FOR_avx_storedquv32qi): ... this.
4069 * config/i386/i386.md (fpint_logic): New.
4070 * config/i386/sse.md (VMOVE): Extend for AVX512.
4074 (VI_UNALIGNED_LOADSTORE): Ditto.
4075 (sse2_avx_avx512f): Ditto.
4076 (sse2_avx2): Extend for AVX512.
4077 (sse4_1_avx2): Ditto.
4078 (avx2_avx512f): New.
4079 (sse): Extend for AVX512.
4082 (avxsizesuffix): Ditto.
4083 (sseintvecmode): Ditto.
4085 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
4086 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
4087 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
4088 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
4089 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
4090 (<sse2_avx_avx512f>_storedqu<mode): ... this.
4091 (<sse>_movnt<mode>): Replace constraint "x" with "v".
4092 (STORENT_MODE): Extend for AVX512.
4093 (*absneg<mode>2): Replace constraint "x" with "v".
4094 (*mul<mode>3): Ditto.
4095 (*ieee_smin<mode>3): Ditto.
4096 (*ieee_smax<mode>3): Ditto.
4097 (avx_cmp<mode>3): Replace VF with VF_128_256.
4098 (*<sse>_maskcmp<mode>3_comm): Ditto.
4099 (<sse>_maskcmp<mode>3): Ditto.
4100 (<sse>_andnot<mode>3): Extend for AVX512.
4101 (<code><mode>3, anylogic): Replace VF with VF_128_256.
4102 (<code><mode>3, fpint_logic): New.
4103 (*<code><mode>3): Extend for AVX512.
4104 (avx512flogicsuff): New.
4105 (avx512f_<logic><mode>): Ditto.
4106 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
4108 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
4109 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
4110 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
4111 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
4112 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
4113 (xop_vpermil2<mode>3): Ditto.
4114 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
4115 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
4116 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
4117 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
4119 2013-09-25 Tom Tromey <tromey@redhat.com>
4121 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
4122 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
4123 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
4124 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
4125 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
4126 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
4127 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
4128 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
4129 (GRAPHITE_HTAB_H): Remove.
4131 2013-09-25 Tom Tromey <tromey@redhat.com>
4133 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
4135 2013-09-25 Tom Tromey <tromey@redhat.com>
4137 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
4139 2013-09-25 Tom Tromey <tromey@redhat.com>
4141 * config/i386/t-i386 (i386.o): Remove.
4142 (i386-c.o): Use COMPILE and POSTCOMPILE.
4144 2013-09-25 Tom Tromey <tromey@redhat.com>
4146 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
4148 2013-09-25 Tom Tromey <tromey@redhat.com>
4150 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
4151 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
4152 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
4153 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
4154 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
4155 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
4156 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
4157 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
4158 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
4159 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
4160 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
4161 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
4162 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
4163 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
4164 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
4165 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
4166 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
4167 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
4168 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
4169 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
4170 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
4171 (tree-ssa-pre.o, tree-ssa-sccvn.o)
4172 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
4173 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
4174 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
4175 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
4176 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
4177 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
4178 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
4179 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
4180 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
4181 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
4182 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
4183 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
4184 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
4185 (graphite-clast-to-gimple.o, graphite-dependences.o)
4186 (graphite-interchange.o, graphite-poly.o)
4187 (graphite-scop-detection.o, graphite-sese-to-poly.o)
4188 (graphite-optimize-isl.o, tree-vect-loop.o)
4189 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
4190 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
4191 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
4192 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
4193 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
4194 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
4195 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
4196 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
4197 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
4198 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
4199 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
4200 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
4201 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
4202 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
4203 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
4204 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
4205 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
4206 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
4207 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
4208 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
4209 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
4210 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
4211 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
4212 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
4213 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
4214 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
4215 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
4216 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
4217 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
4218 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
4219 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
4220 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
4221 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
4222 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
4223 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
4224 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
4225 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
4226 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
4227 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
4228 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
4229 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
4230 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
4231 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
4232 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
4233 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
4234 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
4235 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
4236 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
4237 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
4238 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
4239 (gcov-dump.o): Remove.
4240 (default-c.o): Use COMPILE and POSTCOMPILE.
4241 (CFLAGS-gcc.o): New variable.
4242 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
4244 2013-09-25 Tom Tromey <tromey@redhat.com>
4246 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
4247 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
4248 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
4249 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
4250 (c-family/c-pragma.o, c-family/c-pretty-print.o)
4251 (c-family/c-semantics.o, c-family/c-ada-spec.o)
4252 (c-family/array-notation-common.o, c-family/stub-objc.o)
4253 (c-family/c-ubsan.o): Remove.
4255 2013-09-25 Tom Tromey <tromey@redhat.com>
4257 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
4259 2013-09-25 Tom Tromey <tromey@redhat.com>
4261 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
4262 to add -DENABLE_SHARED_LIBGCC.
4263 (gcc.o): Don't use subshell.
4265 2013-09-25 Tom Tromey <tromey@redhat.com>
4267 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
4268 * configure.ac: Don't invoke AM_PROG_CC_C_O.
4269 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
4270 * configure, config.in: Rebuild.
4272 2013-09-25 Tom Tromey <tromey@redhat.com>
4274 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
4275 (COMPILE, POSTCOMPILE): New variables.
4276 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
4277 (DEPFILES): New variable.
4278 Include ".Po" files.
4279 * configure.ac: Add checks for dependency checking.
4280 * configure, aclocal.m4: Regenerate.
4282 2013-09-25 Tom Tromey <tromey@redhat.com>
4284 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
4285 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
4287 2013-09-25 Tom Tromey <tromey@redhat.com>
4289 * Makefile.in (generated_files): Add options.h,
4290 target-hooks-def.h, insn-opinit.h,
4291 common/common-target-hooks-def.h, pass-instances.def,
4292 c-family/c-target-hooks-def.h.
4294 2013-09-25 Jeff Law <law@redhat.com>
4296 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
4297 than foo[foo.length () - 1] to access last member in a vec.
4298 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
4300 2013-09-25 Richard Biener <rguenther@suse.de>
4303 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
4305 2013-09-25 Jan Hubicka <jh@suse.cz>
4307 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
4310 2013-09-25 Marek Polacek <polacek@redhat.com>
4313 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
4314 when determining the type name.
4316 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
4318 * config/sh/sh.md: Fix formatting.
4320 2013-09-24 Xinliang David Li <davidxl@google.com>
4322 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
4323 max peel iterations parameter.
4324 * param.def: New parameter.
4325 * doc/invoke.texi: Document New parameter.
4327 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
4329 * gimple-pretty-print.c: Various whitespace tweaks.
4330 * tree-core.h: Likewise.
4331 * tree-pretty-print.c: Likewise.
4332 * tree-ssa-alias.c: Likewise.
4333 * tree-ssa-copy.c: Likewise.
4334 * tree-ssanames.c: Likewise.
4335 * tree-ssanames.h: Likewise.
4336 * tree-vrp.c: Likewise.
4338 2013-09-24 Alan Modra <amodra@gmail.com>
4342 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
4343 for output operands that disallow regs. Don't use EXPAND_WRITE on
4346 2013-09-24 Richard Biener <rguenther@suse.de>
4349 * tree.c (reference_alias_ptr_type): Move ...
4350 * alias.c (reference_alias_ptr_type): ... here and implement
4351 in terms of the new reference_alias_ptr_type_1.
4352 (ref_all_alias_ptr_type_p): New helper.
4353 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
4354 use ref_all_alias_ptr_type_p.
4355 (get_deref_alias_set): Add flag_strict_aliasing check here.
4356 (reference_alias_ptr_type_1): New function, split out from ...
4357 (get_alias_set): ... here.
4358 (alias_ptr_types_compatible_p): New function.
4359 * alias.h (reference_alias_ptr_type): Declare.
4360 (alias_ptr_types_compatible_p): Likewise.
4361 * tree.h (reference_alias_ptr_type): Remove.
4362 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
4363 to compare MEM_REF alias types.
4365 2013-09-24 Richard Biener <rguenther@suse.de>
4367 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
4369 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
4371 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
4372 reload helper function arrays into a single array reg_addr.
4373 (reload_fpr_gpr): Likewise.
4374 (reload_gpr_vsx): Likewise.
4375 (reload_vsx_gpr): Likewise.
4376 (struct rs6000_reg_addr): Likewise.
4377 (reg_addr): Likewise.
4378 (rs6000_debug_reg_global): Change rs6000_vector_reload,
4379 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
4380 (rs6000_init_hard_regno_mode_ok): Likewise.
4381 (rs6000_secondary_reload_direct_move): Likewise.
4382 (rs6000_secondary_reload): Likewise.
4384 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
4385 constraints: wu, ww, and wy. Repurpose wv constraint added during
4386 power8 changes. Put wg constraint in alphabetical order.
4388 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
4389 for future work to add ISA 2.07 VSX single precision support.
4390 (-mvsx-scalar-double): Change default from -1 to 1, update
4391 documentation comment.
4392 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
4393 (-mupper-regs-df): New debug switch to control whether DF values
4394 can go in the traditional Altivec registers.
4395 (-mupper-regs-sf): New debug switch to control whether SF values
4396 can go in the traditional Altivec registers.
4398 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
4400 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
4401 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
4402 Add new constraints, wu/ww/wy. Repurpose wv constraint.
4403 (rs6000_debug_legitimate_address_p): Print if we are running
4404 before, during, or after reload.
4405 (rs6000_secondary_reload): Add a comment.
4406 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
4408 * config/rs6000/constraints.md (wa constraint): Sort w<x>
4409 constraints. Update documentation string.
4410 (wd constraint): Likewise.
4411 (wf constraint): Likewise.
4412 (wg constraint): Likewise.
4413 (wn constraint): Likewise.
4414 (ws constraint): Likewise.
4415 (wt constraint): Likewise.
4416 (wx constraint): Likewise.
4417 (wz constraint): Likewise.
4418 (wu constraint): New constraint for ISA 2.07 SFmode scalar
4420 (ww constraint): Likewise.
4421 (wy constraint): Likewise.
4422 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
4423 in the previous submissions.
4424 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
4426 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
4428 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
4429 (REG_BR_PROB): Say that the probability is stored in an int_list.
4430 * reg-notes.def: Update commentary to mention INT_LIST.
4431 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
4432 (INT_LIST): New rtx.
4433 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
4434 * rtlanal.c (int_reg_note_p): New function.
4435 (alloc_reg_note): Assert that the note does not have an int argument.
4436 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
4437 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
4438 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
4439 rather than an INSN_LIST. Handle INT_LIST.
4440 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
4441 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
4442 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
4443 Manipulate them as ints rather than rtxes.
4444 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
4445 * regmove.c (copy_src_to_dest): Likewise.
4446 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
4448 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
4449 into the cases that need it.
4450 * config/arm/arm.c (arm_unwind_emit): Likewise.
4452 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
4453 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
4454 * loop-doloop.c (add_test, doloop_modify): Likewise.
4455 * loop-unswitch.c (compare_and_jump_seq): Likewise.
4456 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
4457 * predict.c (combine_predictions_for_insn): Likewise.
4458 * print-rtl.c (print_rtx): Handle INT_LIST.
4459 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
4460 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
4461 * config/arm/arm.c (emit_unlikely_jump): Likewise.
4462 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
4463 (ix86_split_fp_branch, predict_jump): Likewise.
4464 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
4465 * config/sh/sh.c (expand_cbranchsi4): Likewise.
4466 * config/spu/spu.c (ea_load_store_inline): Likewise.
4468 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
4469 value of a REG_BR_PROB note.
4470 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
4471 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
4472 * emit-rtl.c (try_split): Likewise.
4473 * predict.c (br_prob_note_reliable_p): Likewise.
4474 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
4475 * reorg.c (mostly_true_jump): Likewise.
4476 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
4477 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
4478 * config/i386/i386.c (ix86_print_operand): Likewise.
4479 * config/ia64/ia64.c (ia64_print_operand): Likewise.
4480 * config/mmix/mmix.c (mmix_print_operand): Likewise.
4481 * config/rs6000/rs6000.c (output_cbranch): Likewise.
4482 * config/s390/s390.c (s390_expand_tbegin): Likewise.
4483 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
4484 * config/sparc/sparc.c (output_cbranch): Likewise.
4485 * config/spu/spu.c (get_branch_target): Likewise.
4486 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
4487 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
4489 2013-09-23 Jan Hubicka <jh@suse.cz>
4491 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
4493 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
4494 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
4496 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
4497 (gimple_fold_call): Dump inconsistent devirtualizations; add
4498 sanity check for type based devirtualizations.
4499 * ipa-prop.c: Include ipa-utils.h
4500 (ipa_intraprocedural_devirtualization): Add sanity check.
4501 (try_make_edge_direct_virtual_call): Likewise.
4503 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
4505 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
4506 assignment statements.
4508 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
4510 * gimple-pretty-print.c (dump_ssaname_info): New function.
4511 (dump_gimple_phi): Call it.
4512 (pp_gimple_stmt_1): Likewise.
4513 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
4514 (range_info_def): Declare.
4515 * tree-pretty-print.c (pp_double_int): New function.
4516 (dump_generic_node): Call it.
4517 * tree-pretty-print.h (pp_double_int): Declare.
4518 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
4519 * tree-ssanames.h (range_info_def): New structure.
4520 (value_range_type): Move definition here.
4521 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
4523 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
4525 (set_range_info): New function.
4526 (get_range_info): Likewise.
4527 (duplicate_ssa_name_range_info): Likewise.
4528 (duplicate_ssa_name_fn): Check pointer type and call
4529 duplicate_ssa_name_range_info.
4530 * tree-ssa-copy.c (fini_copy_prop): Likewise.
4531 * tree-vrp.c (value_range_type): Remove definition, now in
4533 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
4534 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
4535 (SSA_NAME_RANGE_INFO): New macro.
4537 2013-09-23 Richard Biener <rguenther@suse.de>
4539 PR tree-optimization/58464
4540 * tree-ssa-pre.c (phi_trans_lookup): Remove.
4541 (phi_trans_add): Change to add conditionally on being not present.
4542 (phi_translate_1): Remove recursion detection here.
4543 (phi_translate): Pre-seed the cache with NULL to catch
4544 recursion here in a more generic way.
4545 (bitmap_find_leader): Adjust comment.
4546 (get_representative_for): Dump value-numbers.
4547 (create_expression_by_pieces): Likewise.
4548 (insert_into_preds_of_block): Likewise.
4550 2013-09-23 Christian Bruel <christian.bruel@st.com>
4553 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
4554 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
4556 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
4559 2013-09-20 Renlin Li <renlin.li@arm.com>
4561 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
4563 (aarch64_expand_epilogue): Likewise.
4564 (aarch64_legitimize_reload_address): Likewise.
4566 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
4568 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
4569 NULL_TREE before pushing them onto the vector. Likewise for labels.
4571 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
4573 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
4574 comparison operators when -fno-trapping-math is in effect.
4575 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
4576 comparison operators in TFmode and assert that unsupported operators
4578 (ia64_print_operand): Likewise.
4580 2013-09-21 Jan Hubicka <jh@suse.cz>
4582 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
4583 (sse_typeless_stores): Enable for core
4584 (sse_load0_by_pxor): Likewise.
4585 (four_jump_limit): Disable for core.
4586 (pad_returns): Likewise.
4587 (avoid_vector_decode): Likewise.
4588 (fuse_cmp_and_branch): Enable for cores.
4589 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
4591 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4594 * config/pa/pa.c (pa_option_override): Disable auto increment and
4595 decrement instructions until reload is completed.
4597 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
4598 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
4599 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
4601 2013-09-20 DJ Delorie <dj@redhat.com>
4602 Nick Clifton <nickc@redhat.com>
4604 * config/rl78/rl78.c: Various whitespace and comment tweaks.
4605 (need_to_save): Save bank 0 on interrupts.
4606 (characterize_address): Strip far address wrappers.
4607 (rl78_as_legitimate_address): Likewise.
4608 (transcode_memory_rtx): Likewise.
4609 (rl78_peep_movhi_p): Disable this peephole after devirt.
4610 (rl78_propogate_register_origins): Forget all origins when a
4612 * config/rl78/rl78-virt.md: Various whitespace tweaks.
4613 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
4615 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
4616 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
4617 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
4619 * config/rl78/constraints.md: Various whitespace and paren tweaks.
4621 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4623 * config/pa/pa.md: In "scc" insn patterns, change output template to
4624 handle const0_rtx in reg_or_0_operand operands.
4626 2013-09-20 Martin Husemann <martin@NetBSD.org>
4629 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
4630 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
4632 2013-09-20 Richard Biener <rguenther@suse.de>
4635 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
4636 remembering SSA name version and block index.
4637 (new_scev_info_str): Adjust.
4638 (hash_scev_info): Likewise. Also hash the block index.
4639 (eq_scev_info): Adjust.
4640 (find_var_scev_info): Likewise.
4641 (struct instantiate_cache_entry): Remove.
4642 (struct instantiate_cache_type): Use a htab to map name, block
4644 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4645 (get_instantiated_value_entry): Likewise.
4646 (hash_idx_scev_info, eq_idx_scev_info): New functions.
4647 (instantiate_scev_name): Adjust.
4649 2013-09-20 Jeff Law <law@redhat.com>
4651 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
4653 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
4655 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
4656 Call aarch64_simd_expand_args to update op[argc].
4658 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4660 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
4663 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4665 * gengtype.c (file_rules): Added rule for *.cc files.
4666 (get_output_file_with_visibility): Give fatal message when no
4669 2013-09-20 Renlin Li <renlin.li@arm.com>
4671 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
4672 (aarch64_expand_epilogue): Likewise.
4673 (aarch64_legitimize_reload_address): Likewise.
4675 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
4678 * expr.c (expand_assignment): Remove misalignp code path.
4680 2013-09-20 Marek Polacek <polacek@redhat.com>
4683 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
4684 TYPE_PRECISION. Add asserts.
4686 2013-09-20 Richard Biener <rguenther@suse.de>
4688 PR tree-optimization/58453
4689 * tree-loop-distribution.c (distribute_loop): Apply the cost
4690 model for -ftree-loop-distribute-patterns, too.
4692 2013-09-20 Richard Biener <rguenther@suse.de>
4695 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
4696 make type comparison less strict.
4698 2013-09-20 Alan Modra <amodra@gmail.com>
4700 * configure: Regenerate.
4701 * aclocal.m4: Regenerate.
4703 2013-09-20 Marek Polacek <polacek@redhat.com>
4706 * doc/extend.texi: Document that attribute used is meant to be used
4707 on variables with static storage duration.
4709 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4711 PR tree-optimization/58472
4712 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
4713 simd_lane_access set inv_p = false.
4714 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
4715 the simduid magic VAR_DECL.
4717 2013-09-19 Jan Hubicka <jh@suse.cz>
4719 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
4721 2013-09-17 Jeff Law <law@redhat.com>
4723 * tree-ssa-dom.c (record_temporary_equivalences): New function
4724 split out of dom_opt_dom_walker::after_dom_children.
4725 (dom_opt_dom_walker::thread_across_edge): Move common code
4726 in here from dom_opt_dom_walker::after_dom_children.
4727 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
4729 2013-09-19 Jan Hubicka <jh@suse.cz>
4731 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
4732 (TARGET_GENERIC): Use PROCESOR_GENERIC
4733 (enum processor_type): Unify generic32 and 64.
4734 * i386.md (cpu): Likewise.
4735 * x86-tune.def (use_leave): Enable for generic32.
4736 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
4737 * athlon.md: Change generic64 to generic in all occurences.
4738 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
4739 (ix86_target_macros_internal): Likewise.
4740 * driver-i386.c (host_detect_local_cpu): Likewise.
4741 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
4743 (generic_memcpy, generic_memset, generic_cost): This one.
4744 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
4745 (m_GENERIC32, m_GENERIC64): Remove.
4746 (m_GENERIC): Turn into one flag.
4747 (processor_target): Unify generic tunnings.
4748 (ix86_option_override_internal): Replace generic32/64 by generic.
4749 (ix86_issue_rate): Likewise.
4750 (ix86_adjust_cost): Likewise.
4752 2013-09-19 Jan Hubicka <jh@suse.cz>
4754 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
4755 of speculative flag.
4757 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4759 * omp-low.c (expand_omp_sections): Always pass len - 1 to
4760 GOMP_sections_start, even if !exit_reachable.
4762 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4764 * lra-constraints.c (need_for_all_save_p): Use macro
4765 HARD_REGNO_CALL_PART_CLOBBERED.
4766 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
4767 set up pseudo conflict hard regs.
4769 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4772 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
4775 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4777 PR rtl-optimization/58438
4778 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
4779 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
4780 from previous subpasses.
4782 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
4784 * arm.c (arm_get_frame_offsets): Validate architecture supports
4785 LDRD/STRD before accepting the tuning preference.
4786 (arm_expand_prologue): Likewise.
4787 (arm_expand_epilogue): Likewise.
4789 2013-09-18 Richard Biener <rguenther@suse.de>
4791 PR tree-optimization/58417
4792 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
4793 have chrecs with symbols defined in the loop as operands.
4794 (chrec_fold_multiply): Likewise.
4795 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
4796 parameters before folding binary operations.
4797 (struct instantiate_cache_entry_hasher): Remove.
4798 (struct instantiate_cache_type): Use a pointer-map.
4799 (instantiate_cache_type::instantiate_cache_type): New function.
4800 (instantiate_cache_type::get): Likewise.
4801 (instantiate_cache_type::set): Likewise.
4802 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4803 (get_instantiated_value_entry): Likewise.
4804 (global_cache): New global.
4805 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
4806 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
4807 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
4808 (instantiate_scev_name): Adjust.
4809 (instantiate_scev): Construct global instead of local cache.
4810 (resolve_mixers): Likewise.
4812 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
4813 Paolo Carlini <paolo.carlini@oracle.com>
4816 * doc/implement-cxx.texi: Fix references to the C++ standards.
4818 2013-09-18 Jakub Jelinek <jakub@redhat.com>
4820 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
4821 * tree-vect-data-refs.c (vect_analyze_data_refs): For
4822 simd_lane_access drs, update also DR_ALIGNED_TO.
4824 2013-09-18 Marek Polacek <polacek@redhat.com>
4827 * doc/extend.texi: Document no_sanitize_undefined attribute.
4828 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
4829 no_sanitize_undefined attribute.
4831 2013-09-18 Nick Clifton <nickc@redhat.com>
4833 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
4834 (ASM_DECLARE_FUNCTION_NAME): Define.
4836 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
4838 * compare-elim.c (find_comparison_dom_walker): New class
4839 (find_comparisons_in_bb): Rename to
4840 find_comparison_dom_walker::before_dom_children
4841 (find_comparisons): Adjust
4842 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
4844 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
4845 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
4846 (init_walk_dominator_tree): Remove declaration.
4847 (fini_walk_dominator_tree): Remove declaration.
4848 * fwprop.c (single_def_use_dom_walker): New class
4849 (single_def_use_enter_block): Convert to
4850 single_def_use_dom_walker::before_dom_children.
4851 (single_def_use_leave_block): Convert to
4852 single_def_use_dom_walker::after_dom_children.
4853 (build_single_def_use_links): Adjust.
4854 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
4856 (find_candidates_in_block): Convert to
4857 find_candidates_dom_walker::before_dom_children.
4858 (execute_strength_reduction): Adjust.
4859 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
4860 (sese_dom_walker): New class.
4861 (sese_dom_walker::sese_dom_walker): New constructor.
4862 (sese_dom_walker::~sese_dom_walker): New destructor.
4863 (build_sese_conditions_before): Convert to
4864 sese_dom_walker::before_dom_children.
4865 (build_sese_conditions_after): Convert to
4866 sese_dom_walker::after_dom_children.
4867 (build_poly_scop): Adjust
4868 * tree-into-ssa.c (rewrite_dom_walker): New class
4869 (rewrite_enter_block): Convert to
4870 rewrite_dom_walker::before_dom_children.
4871 (rewrite_leave_block): Convert to
4872 rewrite_dom_walker::after_dom_children.
4873 (rewrite_update_dom_walker): New class.
4874 (rewrite_update_enter_block): Convert to
4875 rewrite_update_dom_walker::before_dom_children.
4876 (rewrite_update_leave_block): Convert to
4877 rewrite_update_dom_walker::after_dom_children.
4878 (rewrite_blocks, rewrite_into_ssa): Adjust.
4879 (mark_def_dom_walker): New class.
4880 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
4881 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
4882 (mark_def_sites_blocks): Convert to
4883 mark_def_dom_walker::before_dom_children.
4884 (mark_def_site_blocks): Remove.
4885 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
4886 (tree_ssa_dominator_optimize): Adjust.
4887 (dom_thread_across_edge): Convert to method
4888 dom_opt_dom_walker::thread_across_edge.
4889 (dom_opt_enter_block): Convert to member function
4890 dom_opt_dom_walker::before_dom_children.
4891 (dom_opt_leave_block): Convert to member function
4892 dom_opt_dom_walker::after_dom_children.
4893 * tree-ssa-dse.c (dse_dom_walker): New class.
4894 (dse_enter_block): Convert to member function
4895 dse_dom_walker::before_dom_children.
4896 (tree_ssa_dse): Adjust.
4897 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
4898 (determine_invariantness_stmt): Convert to method
4899 invariantness_dom_walker::before_dom_children.
4900 (determine_invariantness): Remove
4901 (move_computations_dom_walker): New class.
4902 (move_computations_stmt): Convert to method
4903 move_computations_dom_walker::before_dom_children.
4904 (move_computations, tree_ssa_lim): Adjust.
4905 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
4906 (nt_init_block): Make method
4907 notrappping_dom_walker::before_dom_children.
4908 (nt_fini_block): Make
4909 method nontrapping_dom_walker::after_dom_children.
4910 (get_non_trapping): Adjust.
4911 * tree-ssa-pre.c (eliminate_dom_walker): New class.
4912 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
4913 (eliminate_leave_block): Make method.
4914 eliminate_dom_walker::after_dom_children.
4916 * tree-ssa-strlen.c (strlen_dom_walker): New class.
4917 (strlen_enter_block): Make method
4918 strlen_dom_walker::before_dom_children.
4919 (strlen_leave_block): Make
4920 method strlen_dom_walker::after_dom_children.
4921 (tree_ssa_strlen): Adjust.
4922 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
4923 (tree_ssa_uncprop): Adjust.
4924 (uncprop_leave_block): Make method
4925 uncprop_dom_walker::after_dom_children.
4926 (uncprop_leave_block): Make method
4927 uncprop_dom_walker::before_dom_children.
4929 2013-09-18 Bin Cheng <bin.cheng@arm.com>
4931 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
4932 branch in same basic block. Check both src and dest of the move insn.
4934 2013-09-17 Nick Clifton <nickc@redhat.com>
4936 * config/rl78/rl78-real.md (bf): New pattern.
4938 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
4939 (rl78_print_operand): Do not put a # before a %B.
4940 * config/rl78/rl78.opt: Tweak doc strings.
4942 2013-09-17 DJ Delorie <dj@redhat.com>
4944 * config/rl78/constraints.md (Wcv): Allow up to $r31.
4945 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
4946 (rl78_option_override): Likewise, if -mallregs.
4947 (is_virtual_register): Likewise.
4948 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
4949 (REGNO_OK_FOR_BASE_P): Likewise.
4950 * config/rl78/rl78.opt (-mallregs): New.
4952 2013-09-17 Nick Clifton <nickc@redhat.com>
4954 * config/rl78/rl78.c (need_to_save): Change return type to bool.
4955 For interrupt functions: save all call clobbered registers if the
4956 interrupt handler is not a leaf function.
4957 (rl78_expand_prologue): Always recompute the frame information.
4958 For interrupt functions: only select bank 0 if one of the bank 0
4959 registers is going to be psuhed.
4961 2013-09-17 DJ Delorie <dj@redhat.com>
4963 * config/rl78/constraints.md: For each W* constraint, rename to C*
4964 and create a W* constraint that checks for an optional ES: prefix
4966 * config/rl78/rl78.md (UNS_ES_ADDR): New.
4967 (es_addr): New. Used to wrap far addresses.
4968 * config/rl78/rl78-protos.h (rl78_es_addr): New.
4969 (rl78_es_base): New.
4970 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
4971 wrapped far addresses.
4972 (rl78_print_operand_1): Unwrap far addresses before processing.
4973 (rl78_lo16): Wrap far addresses in unspecs.
4974 (rl78_es_addr): New.
4975 (rl78_es_base): New.
4976 (insn_ok_now): Check for not-yet-wrapped far addresses.
4977 (transcode_memory_rtx): Properly re-wrap far addresses.
4979 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
4981 * config/sparc/t-rtems: Add leon3 multilibs.
4983 2013-09-17 Cong Hou <congh@google.com>
4985 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
4986 when checking the dot production pattern. The type of rhs operand
4987 of multiply is now checked correctly.
4989 2013-09-17 Jeff Law <law@redhat.com>
4991 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
4992 edge implied equivalences into successor phis.
4993 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
4994 here from tree-ssa-threadedge.c.
4995 (mark_threaded_blocks): When threading through a joiner, if both
4996 successors of the joiner's clone reach the same block, verify the
4997 PHI arguments are equal. If not, cancel the jump threading request.
4998 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
4999 tree-ssa-threadupdate.c
5000 (thread_across_edge): Don't check PHI argument equality when
5001 threading through joiner block here.
5003 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
5005 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
5006 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
5007 (warn_uninit, warn_uninitialized_vars,
5008 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
5009 Move to tree-ssa-uninit.c.
5010 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
5011 (has_undefined_value_p): New. Pass dependant parts of
5012 ssa_undefined_value_p.
5013 (uninit_undefined_value_p): Use has_undefined_value_p.
5014 (warn_uninit, warn_uninitialized_vars,
5015 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
5016 Move from tree-ssa.c.
5017 * tree-ssa.h: Adjust prototypes.
5019 2013-09-17 Jan Hubicka <jh@suse.cz>
5022 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
5023 * ipa-inline.c (can_inline_edge_p): Do not downgrade
5024 FUNCTION_NOT_OPTIMIZED.
5025 * ipa-inline-analysis.c (compute_inline_parameters): Function
5026 not optimized is not inlinable unless it is alwaysinline.
5027 (inline_analyze_function): Force calls in not optimized
5028 function not inlinable.
5030 2013-09-17 Jan Hubicka <jh@suse.cz>
5033 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
5035 * ipa.c (function_and_variable_visibility): Likewise.
5036 * ipa-profile.c (ipa_profile): Likewise.
5038 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
5041 * cgraph.c (cgraph_function_body_availability): Check for ifunc
5042 attribute, and don't inline the resolver in this case.
5044 2013-09-17 Teresa Johnson <tejohnson@google.com>
5046 * coverage.c (get_coverage_counts): Add missing newline.
5048 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5050 PR tree-optimization/58088
5051 * fold-const.c (mask_with_trailing_zeros): New function.
5052 (fold_binary_loc): Make sure we don't recurse infinitely
5053 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
5054 Use mask_with_trailing_zeros where appropriate.
5056 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
5058 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
5059 of 'prev' var to get better distance estimation.
5061 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
5063 * tree-inline.h (struct copy_body_data): Add transform_parameter.
5064 * tree-inline.c (is_parameter_of): New predicate.
5065 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
5066 a parameter has been remapped.
5067 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
5068 (optimize_inline_calls): Initialize transform_parameter.
5069 (copy_gimple_seq_and_replace_locals): Likewise.
5070 (tree_function_versioning): Likewise.
5071 (maybe_inline_call_in_expr): Likewise.
5073 2013-09-17 Nick Clifton <nickc@redhat.com>
5075 * config/msp430/msp430-protos.h: Add prototypes for new functions.
5076 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
5078 (is_attr_func): New function.
5079 (msp430_is_interrupt_func): New function.
5080 (is_naked_func): New function.
5081 (is_reentrant_func): New function.
5082 (is_critical_func): New function.
5083 (msp430_start_function): Add annotations for function attributes.
5084 (msp430_attr): New function.
5085 (msp430_attribute_table): New.
5086 (msp430_function_section): New function.
5087 (TARGET_ASM_FUNCTION_SECTION): Define.
5088 (msp430_builtin): New enum.
5089 (msp430_init_builtins): New function.
5090 (msp430_builtin_devl): New function.
5091 (msp430_expand_builtin): New function.
5092 (TARGET_INIT_BUILTINS): Define.
5093 (TARGET_EXPAND_BUILTINS): Define.
5094 (TARGET_BUILTIN_DECL): Define.
5095 (msp430_expand_prologue): Add support for naked, interrupt,
5096 critical and reentrant functions.
5097 (msp430_expand_epilogue): Likewise.
5098 (msp430_print_operand): Handle 'O' character.
5099 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
5101 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
5102 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
5103 (pushm): Use a 'n' rather than an 'i' constraint.
5104 (msp_return): Add generation of the interrupt return instruction.
5105 (disable_interrupts): New pattern.
5106 (enable_interrupts): New pattern.
5107 (push_intr_state): New pattern.
5108 (pop_intr_state): New pattern.
5109 (bic_SR): New pattern.
5110 (bis_SR): New pattern.
5111 * doc/extend.texi: Document MSP430 function attributes and builtin
5114 2013-09-17 Richard Biener <rguenther@suse.de>
5116 PR tree-optimization/58432
5117 * tree-loop-distribution.c (tree_loop_distribution): Also
5118 scan PHIs for outside loop uses and seed a partition from them.
5120 2013-09-17 Bin Cheng <bin.cheng@arm.com>
5122 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
5123 (restructure_reference): Call backtrace_base_for_ref.
5125 2013-09-17 Alan Modra <amodra@gmail.com>
5128 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
5131 2013-09-16 DJ Delorie <dj@redhat.com>
5133 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
5134 vregs location for RL78/G10.
5135 (rl78_expand_prologue): Avoid SEL on G10.
5136 (rl78_expand_epilogue): Likewise.
5137 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
5138 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
5139 __RL78_G10__ when appropriate.
5140 (ASM_SPEC): Pass -mg10 along to the assembler.
5141 * config/rl78/rl78.md (sel_rb): Disable for G10.
5142 * config/rl78/rl78.opt: Add -mg10 option.
5143 * config/rl78/t-rl78: Add -mg10 multilib.
5145 2013-09-16 Xinliang David Li <davidxl@google.com>
5147 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
5148 * omp-low.c (omp_max_vf): Ditto.
5149 (expand_omp_simd): Ditto.
5150 * tree-vectorizer.c (vectorize_loops): Ditto.
5151 (gate_vect_slp): Ditto.
5152 (gate_increase_alignment): Ditto.
5153 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
5154 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
5155 (gate_tree_vectorize): Name change.
5156 (tree_vectorize): Ditto.
5157 (pass_vectorize::gate): Call new function.
5158 (pass_vectorize::execute): Ditto.
5159 * opts.c: O3 default setting change.
5160 (finish_options): Check new flag.
5161 * doc/invoke.texi: Document new flags.
5162 * common.opt: New flags.
5164 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
5166 * doc/tm.texi.in (Cond Exec Macros): Remove node.
5167 (Condition Code): Don't reference it.
5168 * doc/tm.texi: Regenerate.
5170 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
5173 * lra-constraints.c (undo_optional_reloads): Consider all optional
5174 reload even if it did not get a hard reg.
5176 2013-09-16 Teresa Johnson <tejohnson@google.com>
5178 * dumpfile.c (dump_loc): Remove newline emission.
5179 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
5180 emission to dump_printf_loc calls where missing.
5181 (vect_mark_for_runtime_alias_test): Ditto.
5182 (vect_analyze_data_ref_dependence): Ditto.
5183 (vect_analyze_data_ref_dependences): Ditto.
5184 (vect_slp_analyze_data_ref_dependence): Ditto.
5185 (vect_slp_analyze_data_ref_dependences): Ditto.
5186 (vect_compute_data_ref_alignment): Ditto.
5187 (vect_update_misalignment_for_peel): Ditto.
5188 (vect_verify_datarefs_alignment): Ditto.
5189 (vector_alignment_reachable_p): Ditto.
5190 (vect_get_data_access_cost): Ditto.
5191 (vect_enhance_data_refs_alignment): Ditto.
5192 (vect_find_same_alignment_drs): Ditto.
5193 (vect_analyze_data_refs_alignment): Ditto.
5194 (vect_analyze_group_access): Ditto.
5195 (vect_analyze_data_ref_access): Ditto.
5196 (vect_analyze_data_ref_accesses): Ditto.
5197 (vect_prune_runtime_alias_test_list): Ditto.
5198 (vect_analyze_data_refs): Ditto.
5199 (vect_create_addr_base_for_vector_ref): Ditto.
5200 (vect_create_data_ref_ptr): Ditto.
5201 (vect_grouped_store_supported): Ditto.
5202 (vect_grouped_load_supported): Ditto.
5203 * value-prof.c (check_counter): Ditto.
5204 (check_ic_target): Ditto.
5205 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
5206 (vect_recog_widen_mult_pattern): Ditto.
5207 (vect_recog_widen_sum_pattern): Ditto.
5208 (vect_recog_over_widening_pattern): Ditto.
5209 (vect_recog_widen_shift_pattern): Ditto.
5210 (vect_recog_rotate_pattern): Ditto.
5211 (vect_recog_vector_vector_shift_pattern): Ditto.
5212 (vect_recog_divmod_pattern): Ditto.
5213 (vect_recog_mixed_size_cond_pattern): Ditto.
5214 (vect_recog_bool_pattern): Ditto.
5215 (vect_pattern_recog_1): Ditto.
5216 (vect_pattern_recog): Ditto.
5217 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
5218 (vect_is_simple_iv_evolution): Ditto.
5219 (vect_analyze_scalar_cycles_1): Ditto.
5220 (vect_get_loop_niters): Ditto.
5221 (vect_analyze_loop_1): Ditto.
5222 (vect_analyze_loop_form): Ditto.
5223 (vect_analyze_loop_operations): Ditto.
5224 (vect_analyze_loop_2): Ditto.
5225 (vect_analyze_loop): Ditto.
5226 (report_vect_op): Ditto.
5227 (vect_is_slp_reduction): Ditto.
5228 (vect_is_simple_reduction_1): Ditto.
5229 (vect_get_known_peeling_cost): Ditto.
5230 (vect_estimate_min_profitable_iters): Ditto.
5231 (vect_model_reduction_cost): Ditto.
5232 (vect_model_induction_cost): Ditto.
5233 (get_initial_def_for_induction): Ditto.
5234 (vect_create_epilog_for_reduction): Ditto.
5235 (vectorizable_reduction): Ditto.
5236 (vectorizable_induction): Ditto.
5237 (vectorizable_live_operation): Ditto.
5238 (vect_loop_kill_debug_uses): Ditto.
5239 (vect_transform_loop): Ditto.
5240 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
5241 (vect_stmt_relevant_p): Ditto.
5242 (process_use): Ditto.
5243 (vect_mark_stmts_to_be_vectorized): Ditto.
5244 (vect_model_simple_cost): Ditto.
5245 (vect_model_promotion_demotion_cost): Ditto.
5246 (vect_model_store_cost): Ditto.
5247 (vect_get_store_cost): Ditto.
5248 (vect_model_load_cost): Ditto.
5249 (vect_get_load_cost): Ditto.
5250 (vect_init_vector_1): Ditto.
5251 (vect_get_vec_def_for_operand): Ditto.
5252 (vect_finish_stmt_generation): Ditto.
5253 (vectorizable_call): Ditto.
5254 (vectorizable_conversion): Ditto.
5255 (vectorizable_assignment): Ditto.
5256 (vectorizable_shift): Ditto.
5257 (vectorizable_operation): Ditto.
5258 (vectorizable_store): Ditto.
5259 (vectorizable_load): Ditto.
5260 (vectorizable_condition): Ditto.
5261 (vect_analyze_stmt): Ditto.
5262 (vect_transform_stmt): Ditto.
5263 (vect_is_simple_use): Ditto.
5264 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
5265 (vect_can_advance_ivs_p): Ditto.
5266 (vect_update_ivs_after_vectorizer): Ditto.
5267 (vect_do_peeling_for_loop_bound): Ditto.
5268 (vect_gen_niters_for_prolog_loop): Ditto.
5269 (vect_update_inits_of_drs): Ditto.
5270 (vect_create_cond_for_alias_checks): Ditto.
5271 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
5272 (vect_build_slp_tree_1): Ditto.
5273 (vect_supported_load_permutation_p): Ditto.
5274 (vect_analyze_slp_instance): Ditto.
5275 (vect_analyze_slp): Ditto.
5276 (vect_make_slp_decision): Ditto.
5277 (vect_detect_hybrid_slp): Ditto.
5278 (vect_bb_vectorization_profitable_p): Ditto.
5279 (vect_slp_analyze_bb_1): Ditto.
5280 (vect_update_slp_costs_according_to_vf): Ditto.
5281 (vect_get_mask_element): Ditto.
5282 (vect_transform_slp_perm_load): Ditto.
5283 (vect_schedule_slp_instance): Ditto.
5284 (vect_schedule_slp): Ditto.
5285 (vect_slp_transform_bb): Ditto.
5286 * profile.c (read_profile_edge_counts): Ditto.
5287 (compute_branch_probabilities): Ditto.
5288 * coverage.c (get_coverage_counts): Ditto.
5290 2013-09-16 Diego Novillo <dnovillo@google.com>
5292 * tree-core.h: Add missing comment lines from refactoring of tree.h.
5294 2013-09-16 Jan Hubicka <jh@suse.cz>
5296 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
5297 abstract functions; for static functions check the presence of body.
5299 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5301 * config/aarch64/aarch64-simd-builtins.def (fma): New.
5302 * config/aarch64/aarch64-simd.md
5303 (aarch64_mla_elt<mode>): New.
5304 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
5305 (aarch64_mls_elt<mode>): Likewise.
5306 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
5307 (aarch64_fma4_elt<mode>): Likewise.
5308 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
5309 (aarch64_fma4_elt_to_128v2df): Likewise.
5310 (aarch64_fma4_elt_to_64df): Likewise.
5311 (fnma<mode>4): Likewise.
5312 (aarch64_fnma4_elt<mode>): Likewise.
5313 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
5314 (aarch64_fnma4_elt_to_128v2df): Likewise.
5315 (aarch64_fnma4_elt_to_64df): Likewise.
5316 * config/aarch64/iterators.md (VDQSF): New.
5317 * config/aarch64/arm_neon.h
5318 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
5319 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
5321 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5323 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
5324 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
5325 (aarch64_mul3_elt_to_128df): Likewise.
5326 (aarch64_mul3_elt_to_64v2df): Likewise.
5327 * config/aarch64/iterators.md (VEL): Also handle DFmode.
5329 (VMUL_CHANGE_NLANES) Likewise.
5332 * config/aarch64/arm_neon.h
5333 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
5335 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5337 * config/aarch64/arm_neon.h
5338 (vcvtx_high_f32_f64): Fix parameters.
5340 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5341 Uros Bizjak <ubizjak@gmail.com>
5343 * config/alpha.c: Include tree-ssanames.h.
5345 2013-09-16 Richard Biener <rguenther@suse.de>
5347 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
5348 (dot_rdg_1): Handle control_dd.
5349 (create_edge_for_control_dependence): New function.
5350 (create_rdg_edges): Add control dependences if asked for.
5351 (build_rdg): Likewise.
5352 (generate_loops_for_partition): If there are not necessary
5353 control stmts remove all their dependencies.
5354 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
5355 (distribute_loop): Pass on control dependences.
5356 (tree_loop_distribution): Compute control dependences and remove
5357 restriction on number of loop nodes.
5359 2013-09-16 Jakub Jelinek <jakub@redhat.com>
5361 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
5364 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
5366 * cse.c (try_const_anchors): Punt on CC modes.
5368 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5370 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
5372 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
5375 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
5378 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5381 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
5383 (darwin_objc1_section): Likewise.
5384 (darwin_file_end): Emit Image Info section when required.
5386 2013-09-14 Jan Hubicka <jh@suse.cz>
5388 * tree-into-ssa.c (gate_into_ssa): New.
5389 (pass_data_build_ssa): Use it.
5390 * cgraph.h (expand_thunk): Update prototype.
5391 * cgraphunit.c (analyze_function): Expand thunks early.
5392 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
5393 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
5394 set lowered flag; do not add new function.
5395 (assemble_thunks_and_aliases): Update.
5396 * tree-ssa.c (gate_init_datastructures): New gate.
5397 (pass_data_init_datastructures): Use it.
5399 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5402 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
5403 xmm register set described in the psABI.
5405 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
5407 * dwarf2out.c (should_emit_struct_debug): Add check
5408 for type_decl variable is not NULL.
5410 2013-09-13 Jacek Caban <jacek@codeweavers.com>
5412 * config.gcc: Use new winnt-c.c target hooks
5413 * config/t-winnt: New file
5414 * config/winnt-c.c: New file
5415 * doc/tm.texi.in: Document new hook
5416 * doc/tm.texi: Regenerated
5418 2013-09-13 Jan Hubicka <jh@suse.cz>
5421 * ipa-inline.c (check_callers): New function.
5422 (check_caller_edge): Remove.
5423 (want_inline_function_to_all_callers_p): Also permit alises that are
5425 (inline_to_all_callers): Terminate the walk when devirtualization
5426 introduce new calls.
5428 2013-09-13 Jan Hubicka <jh@suse.cz>
5430 * ipa-inline-analysis.c (struct growth_data): Add node.
5431 (do_estimate_growth_1): Fix detection of recursion.
5433 2013-09-13 Jakub Jelinek <jakub@redhat.com>
5435 PR tree-optimization/58392
5436 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
5437 to avoid shadowing of outer loop variable. If
5438 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
5439 replace_by_duplicate_decl simduid of loops that have it set and
5440 set dest_cfun->has_simduid_loops and/or
5441 dest_cfun->has_force_vect_loops.
5442 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
5443 instead of maybe_lookup_decl.
5444 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
5445 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
5446 call. Copy over force_vect and copy and remap simduid. Set
5447 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
5448 (copy_cfg_body): Remove blocks_to_copy argument. Use
5449 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
5450 caller. Don't set cfun->has_simduid_loops and/or
5451 cfun->has_force_vect_loops here.
5452 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
5454 (expand_call_inline, tree_function_versioning): Adjust copy_body
5457 2013-09-13 Martin Jambor <mjambor@suse.cz>
5460 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
5461 the assert if the edge was a speculative one.
5463 2013-09-13 Richard Biener <rguenther@suse.de>
5465 * tree-data-ref.h (known_dependences_p): Move here ...
5466 * tree-loop-distribution.c (known_dependences_p): ... from here.
5467 (dump_rdg_component, debug_rdg_component): Remove.
5469 (generate_loops_for_partition): Use gimple_uid instead of
5470 relying on matching stmt visit order.
5471 (rdg_build_partitions): Take starting stmt vector.
5472 (ldist_gen): Merge into ...
5473 (distribute_loop): ... this function. Do not compute starting
5475 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
5477 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5479 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
5480 Set type attribute correctly. Set predicable_short_it attribute.
5481 (cmpsi_shiftsi): Remove %? from output template.
5483 2013-09-13 Richard Biener <rguenther@suse.de>
5485 * tree-loop-distribution.c (struct rdg_component,
5486 rdg_defs_used_in_other_loops_p, free_rdg_components,
5487 rdg_build_components): Remove.
5488 (stmts_from_loop): Do not record virtual PHIs.
5489 (generate_loops_for_partition): Skip virtual PHIs.
5490 (build_rdg_partition_for_component): Rename to ...
5491 (build_rdg_partition_for_vertex): ... this and adjust.
5492 (rdg_build_partitions): Take a vector of starting vertices
5493 instead of components. Remove unnecessary leftover handling.
5494 (ldist_gen): Do not build components or record other stores.
5495 (distribute_loop): Do not distribute loops containing stmts
5498 2013-09-13 Christian Bruel <christian.bruel@st.com>
5501 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
5503 2013-09-13 Kai Tietz <ktietz@redhat.com>
5505 * config.gcc: Separate cases for mingw and cygwin targets,
5506 and add 64-bit cygwin target case.
5508 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
5509 dll-export inline-functions.
5510 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
5513 2013-09-13 Jeff Law <law@redhat.com>
5517 2013-09-06 Jeff Law <law@redhat.com>
5519 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5520 edge implied equivalences into successor phis.
5522 2013-09-12 DJ Delorie <dj@redhat.com>
5524 * config/rl78/rl78-virt.md: Change from | to \; for asm line
5527 2013-09-12 Brooks Moses <bmoses@google.com>
5530 * Makefile.in: Do not install driver binaries in $(target)/bin.
5532 2013-09-12 DJ Delorie <dj@redhat.com>
5534 * config/rl78/rl78.opt (mrelax): New.
5535 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
5536 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
5538 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
5540 (rl78_expand_epilogue): Likewise.
5542 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
5545 * lra-eliminations.c (remove_reg_equal_offset_note): New.
5546 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
5547 pointer elimination with using remove_reg_equal_offset_note.
5549 2013-09-12 DJ Delorie <dj@redhat.com>
5551 * config/msp430/: New port.
5552 * config.gcc (msp430): Added.
5553 * doc/invoke.texi: Document MSP430 options.
5554 * doc/install.texi: Document msp430-elf
5555 * doc/md.texi: Document msp430-elf
5556 * doc/contrib.texi: Document msp430-elf
5558 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
5561 2013-09-12 Martin Jambor <mjambor@suse.cz>
5564 * ipa-prop.c (remove_described_reference): Give up if the edge in the
5565 reference descriptor is NULL.
5566 (ipa_edge_removal_hook): If owning a reference descriptor, set its
5569 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5571 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
5572 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
5573 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
5574 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
5575 make_temp_ssa_name): move to tree-ssanames.h
5576 * tree-ssa-alias.h: Move prototype.
5577 * tree-ssa.h: Include tree-ssanames.h.
5578 * tree-ssanames.c (FREE_SSANAMES): Move to here.
5579 * tree-ssanames.h: New. Move items from tree-flow*.h
5580 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
5582 2013-09-12 Richard Biener <rguenther@suse.de>
5584 PR tree-optimization/58404
5585 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
5586 propagate non-invariant addresses into dereferences wrapped
5587 in component references.
5589 2013-09-12 Richard Biener <rguenther@suse.de>
5591 PR tree-optimization/58402
5592 * passes.def: Move pass_late_warn_uninitialized later.
5594 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5596 * tree-ssa.h: New. Move content from tree-flow.h and
5598 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
5599 Move prototypes belonging to tree-ssa.c.
5600 * tree-flow-inline.h (redirect_edge_var_map_def,
5601 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
5603 * gimple.h: Adjust prototypes.
5604 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
5606 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
5607 * tree.h: Move prototype to tree-ssa.h.
5608 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
5609 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
5610 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
5611 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
5612 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
5613 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
5614 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
5615 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
5616 graphite-blocking.c, graphite-clast-to-gimple.c,
5617 graphite-dependences.c, graphite-interchange.c,
5618 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
5619 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
5620 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
5621 ipa-reference.c, ipa-split.c, ipa-utils.c,
5622 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
5623 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
5624 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
5625 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5626 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
5627 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
5628 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
5629 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
5630 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
5631 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
5632 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
5633 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
5634 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
5635 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
5636 value-prof.c, var-tracking.c,
5637 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
5639 2013-09-12 Richard Biener <rguenther@suse.de>
5641 PR tree-optimization/58396
5642 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
5643 (build_rdg): Take a loop-nest parameter, fix memleaks.
5644 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
5646 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
5648 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
5651 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
5653 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
5654 The FMA instruction names should have a 'v' prefix.
5656 2013-09-12 Richard Biener <rguenther@suse.de>
5658 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
5659 (dot_rdg): Use popen instead of system in optional code.
5660 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
5661 (already_processed_vertex_p): Adjust.
5662 (has_anti_or_output_dependence, predecessor_has_mem_write,
5663 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
5664 rdg_flag_uses): Remove.
5665 (rdg_flag_vertex): Simplify.
5666 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
5668 (build_rdg_partition_for_component): Process the first vertex
5669 of a component only.
5670 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
5672 2013-09-12 Alan Modra <amodra@gmail.com>
5674 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
5676 2013-09-11 DJ Delorie <dj@redhat.com>
5677 Nick Clifton <nickc@redhat.com>
5679 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
5680 (rl78_stack_based_mem): New.
5681 * config/rl78/constraints.md (Iv08): New.
5690 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
5692 (movsi): Change from expand to insn-and-split.
5693 (ashrsi3): Clobber AX.
5697 * config/rl78/rl78.md (CC_REG): Fix.
5698 (addsi3): Allow memory and immediate operands.
5699 (addsi3_internal): Split into...
5700 (addsi3_internal_virt): ...new, and ...
5701 (addsi3_internal_real): ...new.
5703 (subsi3_internal_virt): New.
5704 (subsi3_internal_real): New.
5705 (mulsi3): Add memory operand.
5706 (mulsi3_rl78): Likewise.
5707 (mulsi3_g13): Likewise.
5708 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
5709 (cbranchqi4_real): Add more constraint options.
5710 (cbranchhi4_real): Expand pattern.
5711 (cbranchhi4_real_signed): New.
5712 (cbranchhi4_real_inverted): New.
5713 (cbranchsi4_real_lt): New.
5714 (cbranchsi4_real_ge): New.
5715 (cbranchsi4_real_signed): New.
5716 (cbranchsi4_real): New.
5718 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
5720 (lshrsi3_virt): Likewise.
5721 (ashlsi3_virt): Likewise.
5722 (cbranchqi4_virt_signed): New.
5723 (cbranchhi4_virt_signed): New.
5724 (cbranchsi4_virt): New.
5725 * config/rl78/rl78.c: Whitespace fixes throughout.
5726 (move_elim_pass): New.
5727 (pass_data_rl78_move_elim): New.
5728 (pass_rl78_move_elim): New.
5729 (make_pass_rl78_move_elim): New.
5730 (rl78_devirt_info): Run devirt earlier.
5731 (rl78_move_elim_info): New.
5732 (rl78_asm_file_start): Register it.
5733 (rl78_split_movsi): New.
5734 (rl78_as_legitimate_address): Allow virtual base registers when
5736 (rl78_addr_space_convert): Remove spurious debug stuff.
5737 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
5738 (rl78_print_operand): More cases for not printing '#'.
5739 (rl78_expand_compare): Remove most of the logic.
5740 (content_memory): New.
5741 (clear_content_memory): New.
5742 (get_content_index): New.
5743 (get_content_name): New.
5744 (display_content_memory): New.
5745 (update_content): New.
5746 (record_content): New.
5747 (already_contains): New.
5748 (insn_ok_now): Re-recog insns with virtual registers.
5749 (add_postponed_content_update): New.
5750 (process_postponed_content_update): New.
5751 (gen_and_emit_move): New.
5752 (transcode_memory_rtx): Record new location content.
5753 Use gen_and_emit_move.
5754 (force_into_acc): New.
5755 (move_to_acc): Use gen_and_emit_move.
5756 (move_from_acc): Likewise.
5757 (move_acc_to_reg): Likewise.
5758 (move_to_x): Likewise.
5759 (move_to_hl): Likewise.
5760 (move_to_de): Likewise.
5761 (rl78_alloc_physical_registers_op1): Record location content.
5762 (has_constraint): New.
5763 (rl78_alloc_physical_registers_op2): Record location content.
5765 (rl78_alloc_physical_registers_ro1): Likewise.
5766 (rl78_alloc_physical_registers_cmp): Likewise.
5767 (rl78_alloc_physical_registers_umul): Likewise.
5768 (rl78_alloc_address_registers_macax): New.
5769 (rl78_alloc_physical_registers): Initialize and set location
5770 content memory as needed.
5771 (rl78_reorg): Make sure split2 is called.
5772 (rl78_rtx_costs): New.
5774 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
5776 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
5777 for (not (neg ...)) and (neg (not ...)) cases.
5779 2013-09-11 Richard Biener <rguenther@suse.de>
5782 * passes.def: Split critical edges before late uninit warning passes.
5783 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
5785 2013-09-11 Jakub Jelinek <jakub@redhat.com>
5787 PR tree-optimization/58385
5788 * fold-const.c (build_range_check): If both low and high are NULL,
5789 use omit_one_operand_loc to preserve exp side-effects.
5791 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5793 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
5795 2013-09-11 Richard Biener <rguenther@suse.de>
5797 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
5798 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
5799 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
5800 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
5801 stmts_from_loop, known_dependences_p, build_empty_rdg,
5802 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
5803 * tree-loop-distribution.c: ... here.
5804 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
5805 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
5806 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
5807 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
5808 * tree-loop-distribution.c: ... here.
5809 * tree-loop-distribution.c: Include gimple-pretty-print.h.
5810 (struct partition_s): Add loops member.
5811 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
5812 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
5813 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
5815 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5816 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5817 Sergey Lega <sergey.s.lega@intel.com>
5818 Anna Tikhonova <anna.tikhonova@intel.com>
5819 Ilya Tocar <ilya.tocar@intel.com>
5820 Andrey Turetskiy <andrey.turetskiy@intel.com>
5821 Ilya Verbin <ilya.verbin@intel.com>
5822 Kirill Yukhin <kirill.yukhin@intel.com>
5823 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5825 * config/i386/constraints.md (k): New.
5827 * config/i386/i386.c (const regclass_map): Add new mask registers.
5828 (dbx_register_map): Ditto.
5829 (dbx64_register_map): Ditto.
5830 (svr4_dbx_register_map): Ditto.
5831 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
5833 (ix86_preferred_reload_class): Disable constants for mask registers.
5834 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
5835 (ix86_hard_regno_mode_ok): Support new mask registers.
5836 (x86_order_regs_for_local_alloc): Ditto.
5837 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
5838 (FIXED_REGISTERS): Add new mask registers.
5839 (CALL_USED_REGISTERS): Ditto.
5840 (REG_ALLOC_ORDER): Ditto.
5841 (VALID_MASK_REG_MODE): New.
5842 (FIRST_MASK_REG): Ditto.
5843 (LAST_MASK_REG): Ditto.
5844 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
5845 (MAYBE_MASK_CLASS_P): New.
5846 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
5847 (REG_CLASS_CONTENTS): Ditto.
5848 (MASK_REGNO_P): New.
5849 (ANY_MASK_REG_P): Ditto.
5850 (HI_REGISTER_NAMES): Add new mask registers.
5851 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
5852 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
5854 (attribute "type"): Add mskmov, msklog.
5855 (attribute "length_immediate"): Support them.
5856 (attribute "memory"): Ditto.
5857 (attribute "prefix_0f"): Ditto.
5858 (*movhi_internal): Support new mask registers.
5859 (*movqi_internal): Ditto.
5860 (define_split): Split out clobber pattern is a logic
5861 insn on mask registers.
5862 (*k<logic><mode>): New.
5863 (*andhi_1): Extend to support mask regs.
5864 (*andqi_1): Extend to support mask regs.
5866 (define_split): Split and-not to and and not if operands
5868 (*<code><mode>_1): Separate HI mode to new pattern...
5869 (*<code>hi_1): This.
5870 (*<code>qi_1): Extend to support mask regs.
5872 (kortestzhi): Ditto.
5873 (kortestchi): Ditto.
5875 (*one_cmpl<mode>2_1): Remove HImode and handle it...
5876 (*one_cmplhi2_1): ...Here, now with mask registers support.
5877 (*one_cmplqi2_1): Support new mask registers.
5878 (HI/QImode arithmetics splitter): Don't split if mask registers
5880 (HI/QImode not splitter): Ditto.
5881 * config/i386/predicated.md (mask_reg_operand): New.
5882 (general_reg_operand): Ditto.
5884 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5886 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
5887 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
5889 2013-09-10 Jeff Law <law@redhat.com>
5891 PR tree-optimization/58380
5892 * tree-ssa-threadupdate.c (thread_block): Recognize another case
5893 of threading through a buried loop header.
5895 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
5896 return value for single successor case.
5898 2013-09-10 Jan Hubicka <jh@suse.cz>
5900 * ipa-devirt.c (ipa_devirt): Enable with LTO.
5902 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
5905 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
5906 support conditional execution.
5907 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
5909 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
5911 * lra.c (lra): Clear lra_optional_reload_pseudos before every
5913 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
5914 Check destination too to check move insn.
5915 (undo_optional_reloads): Add check that the original peudo did not
5916 changed its allocation and the optional reload was inherited on last
5917 inheritance pass. Break loop after deciding to keep optional reload.
5918 (lra_undo_inheritance): Add check that inherited pseudo still in
5921 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
5923 * config/aarch64/aarch64.md (generic_sched): New.
5924 * config/aarch64/aarch64-generic.md (load): Make conditional
5925 on generic_sched attribute.
5926 (nonload): Likewise.
5928 2013-09-10 Jan Hubicka <jh@suse.cz>
5930 * lto-cgraph.c: Include ipa-utils.h.
5931 (compute_ltrans_boundary): Also add possible targets into the boundary.
5933 2013-09-10 Jan Hubicka <jh@suse.cz>
5935 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
5936 VAR_DECL of vtable rather than full expression.
5938 2013-09-10 Jan Hubicka <jh@suse.cz>
5939 Paolo Carlini <paolo.carlini@oracle.com>
5941 * cgraphunit.c (analyze_functions): Save input_location, set it
5942 to UNKNOWN_LOCATION and restore it at the end.
5944 2013-09-10 Martin Jambor <mjambor@suse.cz>
5946 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
5947 represented by a thunk.
5949 2013-09-10 Jeff Law <law@redhat.com>
5951 PR tree-optimization/58343
5952 * tree-ssa-threadupdate.c (thread_block): Identify and disable
5953 jump threading requests through loop headers buried in the middle
5954 of a jump threading path.
5956 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
5957 in return value/type.
5959 2013-09-10 Jakub Jelinek <jakub@redhat.com>
5961 PR rtl-optimization/58365
5962 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
5963 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
5966 2013-09-10 Richard Biener <rguenther@suse.de>
5968 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
5969 * tree-data-ref.c (create_rdg_vertices): Collect all data
5970 references, signal failure to the caller, use data-ref API.
5971 (build_rdg): Compute data references only once. Maintain lifetime
5972 of data references and data dependences from within RDG.
5973 (free_rdg): Free dependence relations.
5974 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
5975 inventing extra dependences.
5976 (distribute_loop): Update for RDG API changes.
5978 2013-09-10 Kai Tietz <ktietz@redhat.com>
5980 * doc/invoke.texi (fms-extensions): Document changed
5981 behavior for ms-abi targets.
5982 * config/i386/i386.c (ix86_option_override_internal):
5983 Set default value of option -fms-extension for ms-abi targets.
5985 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
5987 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
5989 2013-09-10 Alan Modra <amodra@gmail.com>
5992 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
5994 2013-09-10 Alan Modra <amodra@gmail.com>
5996 * config/rs6000/predicates.md (add_cint_operand): New.
5997 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
5998 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
5999 using add_cint_operand.
6000 (largetoc_high_plus_aix): Likewise.
6002 2013-09-09 Jakub Jelinek <jakub@redhat.com>
6004 PR tree-optimization/58364
6005 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
6006 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
6007 the current range can't be an unconditional true or false.
6009 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
6011 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
6013 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
6015 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
6017 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
6020 * doc/invoke.texi (-Wdelete-incomplete): Document it.
6022 2013-09-09 Ian Bolton <ian.bolton@arm.com>
6024 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
6025 NO_REGS for immediate that can't be moved directly into FP_REGS.
6027 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6029 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
6030 comparison with negated operand.
6031 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
6034 2013-09-09 Richard Biener <rguenther@suse.de>
6037 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
6038 of a subloop record all its block as affecting loop-closed SSA form.
6040 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6042 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
6043 of an rtx/bitpos pair.
6044 (store_fixed_bit_field): Update accordingly.
6046 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6048 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
6050 * builtins.c (expand_errno_check): Likewise.
6051 * dwarf2cfi.c (init_return_column_size): Likewise.
6052 * except.c (sjlj_mark_call_sites): Likewise.
6053 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
6054 * lra-constraints.c (emit_inc): Likewise.
6055 * ree.c (combine_set_extension): Likewise.
6056 * regmove.c (fixup_match_2): Likewise.
6057 * reload1.c (inc_for_reload): Likewise.
6059 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6061 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
6062 (force_to_mode, simplify_shift_const_1, simplify_comparison):
6063 Use gen_int_mode with the mode of the associated simplify_* call.
6064 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
6065 * expmed.c (expand_shift_1): Likewise.
6066 * function.c (instantiate_virtual_regs_in_insn): Likewise.
6067 * loop-iv.c (iv_number_of_iterations): Likewise.
6068 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
6069 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
6071 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6073 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
6074 of the associated expand_* call.
6075 (asan_emit_stack_protection): Likewise.
6076 * builtins.c (round_trampoline_addr): Likewise.
6077 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
6078 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
6079 (emit_store_flag): Likewise.
6080 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
6081 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
6083 * function.c (instantiate_virtual_regs_in_insn): Likewise.
6084 * ifcvt.c (noce_try_store_flag_constants): Likewise.
6085 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
6086 * modulo-sched.c (generate_prolog_epilog): Likewise.
6087 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
6088 (expand_ctz, expand_ffs, expand_unop): Likewise.
6090 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6092 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
6093 of the associated gen_rtx_* call.
6094 * caller-save.c (init_caller_save): Likewise.
6095 * combine.c (find_split_point, make_extraction): Likewise.
6096 (make_compound_operation): Likewise.
6097 * dwarf2out.c (mem_loc_descriptor): Likewise.
6098 * explow.c (plus_constant, probe_stack_range): Likewise.
6099 * expmed.c (expand_mult_const): Likewise.
6100 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
6101 * reload1.c (init_reload): Likewise.
6102 * valtrack.c (cleanup_auto_inc_dec): Likewise.
6103 * var-tracking.c (adjust_mems): Likewise.
6104 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
6105 rather than gen_rtx_fmt_ee.
6107 2013-09-09 Jan Hubicka <jh@suse.cz>
6110 * value-prof.c (gimple_ic): Copy also abnormal edges.
6112 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6114 * asan.c (asan_shadow_cst): Use gen_int_mode.
6116 2013-09-08 Jan Hubicka <jh@suse.cz>
6118 * ipa-profile.c: Add toplevel comment.
6119 (ipa_propagate_frequency_1): Be more conservative when profile is read.
6120 (contains_hot_call_p): New function.
6121 (ipa_propagate_frequency): Set frequencies based on counts when
6123 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
6124 profile; do not tamper with profile after inlining if it is read.
6126 2013-09-08 Jan Hubicka <jh@suse.cz>
6128 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
6131 2013-09-08 Jan Hubicka <jh@suse.cz>
6133 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
6136 2013-09-08 Jeff Law <law@redhat.com>
6139 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
6142 2013-09-08 Andi Kleen <ak@linux.intel.com>
6144 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
6146 2013-09-08 Jan Hubicka <jh@suse.cz>
6148 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
6150 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
6152 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
6153 for non-debug insns.
6154 * lra.c (new_insn_reg): Take the containing insn as a parameter.
6155 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
6156 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
6159 2013-09-08 Jan Hubicka <jh@suse.cz>
6161 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
6162 targets and devirtualize to BUILT_IN_UNREACHABLE.
6163 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
6164 * ipa.c (walk_polymorphic_call_targets): New function.
6165 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
6166 functions; use the new timevar.
6167 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
6168 was removed from the program.
6169 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
6170 not consider it in the walk when its vtable is dead.
6171 (possible_polymorphic_call_targets_1): Pass anonymous flag to
6173 (devirt_variable_node_removal_hook): New function.
6174 (possible_polymorphic_call_targets): Also register
6175 devirt_variable_node_removal_hook.
6176 (ipa_devirt): Do not do non-speculative devirtualization.
6177 (gate_ipa_devirt): One execute if devirtualizing speculatively.
6179 2013-09-08 Jan Hubicka <jh@suse.cz>
6181 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
6182 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
6183 varpool_remove_variable_insertion_hook): Declare.
6184 * varpool.c (varpool_node_hook_list): New structure.
6185 (first_varpool_node_removal_hook,
6186 first_varpool_variable_insertion_hook): New variables.
6187 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
6188 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
6189 varpool_remove_variable_insertion_hook,
6190 varpool_call_variable_insertion_hooks): New functions.
6191 (varpool_remove_node): Use it.
6193 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
6196 * diagnostic.c (diagnostic_build_prefix): When s.file is
6197 "<built-in>" don't output line and column numbers.
6199 2013-09-06 Jan Hubicka <jh@suse.cz>
6201 * cgraphunit.c (expand_thunk): Get body before touching arguments.
6202 * lto-streamer-out.c: Stream thunks, too.
6203 * lto-streamer-in.c (input_function): Pop cfun here
6204 (lto_read_body): Instead of here.
6206 2013-09-06 Caroline Tice <cmtice@google.com>
6208 * doc/install.texi: Add documentation for the --enable-vtable-verify
6209 and the --disable-libvtv configure options.
6211 2013-09-06 Jeff Law <law@redhat.com>
6213 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
6214 edge implied equivalences into successor phis.
6216 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
6218 * resource.c (mark_referenced_resources): Handle COND_EXEC.
6220 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
6222 * resource.c (mark_target_live_regs): Compute resources taking
6223 into account if a call is predicated or not.
6225 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6227 * toplev.c (output_stack_usage): Be prepared for suffixes created by
6228 the compiler in the function names.
6230 2013-09-06 Jan Hubicka <jh@suse.cz>
6233 * ipa-inline.c (has_caller_p): New function.
6234 (want_inline_function_to_all_callers_p): Use it.
6235 (sum_callers, inline_to_all_callers): Break out from ...
6236 (ipa_inline): ... here.
6238 2013-09-06 Jan Hubicka <jh@suse.cz>
6240 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
6241 only when AVX is enabled.
6243 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6245 * config/aarch64/aarch64.md
6246 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
6247 is fpsimd_<load/store>2.
6248 (load_pair<mode>): Likewise.
6249 (store_pair<mode>): Likewise.
6251 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6253 * config/arm/types.md (type): Add "mrs" type.
6254 * config/aarch64/aarch64.md
6255 (aarch64_load_tp_hard): Make type "mrs".
6257 (load_tp_hard): Make type "mrs".
6258 * config/arm/cortex-a15.md: Update with new attributes.
6259 * config/arm/cortex-a5.md: Update with new attributes.
6260 * config/arm/cortex-a53.md: Update with new attributes.
6261 * config/arm/cortex-a7.md: Update with new attributes.
6262 * config/arm/cortex-a8.md: Update with new attributes.
6263 * config/arm/cortex-a9.md: Update with new attributes.
6264 * config/arm/cortex-m4.md: Update with new attributes.
6265 * config/arm/cortex-r4.md: Update with new attributes.
6266 * config/arm/fa526.md: Update with new attributes.
6267 * config/arm/fa606te.md: Update with new attributes.
6268 * config/arm/fa626te.md: Update with new attributes.
6269 * config/arm/fa726te.md: Update with new attributes.
6271 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6273 * config/aarch64/aarch64.md
6274 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
6275 (*movtf_aarch64): Likewise.
6277 (thumb1_movdi_insn): Use "multiple" for type where more than one
6278 instruction is used for a move.
6279 (*arm32_movhf): Likewise.
6280 (*thumb_movdf_insn): Likewise.
6282 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6284 * config/arm/types.md (type): Rename fcpys to fmov.
6286 (*arm_movsi_vfp): Rename type fcpys as fmov.
6287 (*thumb2_movsi_vfp): Likewise
6288 (*movhf_vfp_neon): Likewise
6289 (*movhf_vfp): Likewise
6290 (*movsf_vfp): Likewise
6291 (*thumb2_movsf_vfp): Likewise
6292 (*movsfcc_vfp): Likewise
6293 (*thumb2_movsfcc_vfp): Likewise
6294 * config/aarch64/aarch64-simd.md
6295 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
6296 * config/aarch64/aarch64.md
6297 (*movsi_aarch64): Replace type mov_reg with fmovs.
6298 (*movdi_aarch64): Likewise
6299 (*movsf_aarch64): Likewise
6300 (*movdf_aarch64): Likewise
6302 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
6303 * config/arm/iwmmxt.md
6304 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
6305 * config/arm/arm1020e.md: Update with new attributes.
6306 * config/arm/cortex-a15-neon.md: Update with new attributes.
6307 * config/arm/cortex-a5.md: Update with new attributes.
6308 * config/arm/cortex-a53.md: Update with new attributes.
6309 * config/arm/cortex-a7.md: Update with new attributes.
6310 * config/arm/cortex-a8-neon.md: Update with new attributes.
6311 * config/arm/cortex-a9.md: Update with new attributes.
6312 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6313 * config/arm/cortex-r4f.md: Update with new attributes.
6314 * config/arm/marvell-pj4.md: Update with new attributes.
6315 * config/arm/vfp11.md: Update with new attributes.
6317 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6319 * config/aarch64/aarch64.md
6320 (*madd<mode>): Fix type attribute.
6321 (*maddsi_uxtw): Likewise.
6322 (*msub<mode>): Likewise.
6323 (*msubsi_uxtw): Likewise.
6324 (<su_optab>maddsidi4): Likewise.
6325 (<su_optab>msubsidi4): Likewise.
6327 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6329 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
6330 * config/arm/arm.md (core_cycles): Remove fdiv.
6331 * config/arm/vfp.md:
6332 (*sqrtsf2_vfp): Update for attribute changes.
6333 (*sqrtdf2_vfp): Likewise.
6334 * config/aarch64/aarch64.md:
6335 (sqrt<mode>2): Update for attribute changes.
6336 * config/arm/arm1020e.md: Update with new attributes.
6337 * config/arm/cortex-a15-neon.md: Update with new attributes.
6338 * config/arm/cortex-a5.md: Update with new attributes.
6339 * config/arm/cortex-a53.md: Update with new attributes.
6340 * config/arm/cortex-a7.md: Update with new attributes.
6341 * config/arm/cortex-a8-neon.md: Update with new attributes.
6342 * config/arm/cortex-a9.md: Update with new attributes.
6343 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6344 * config/arm/cortex-r4f.md: Update with new attributes.
6345 * config/arm/marvell-pj4.md: Update with new attributes.
6346 * config/arm/vfp11.md: Update with new attributes.
6348 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6350 * config/arm/types.md
6351 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
6352 * config/aarch64/aarch64.md
6353 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
6355 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
6356 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
6357 (float<GPI:mode><GPF:mode>2): Likewise.
6359 (*truncsisf2_vfp): Update with new attributes.
6360 (*truncsidf2_vfp): Likewise.
6361 (fixuns_truncsfsi2): Likewise.
6362 (fixuns_truncdfsi2): Likewise.
6363 (*floatsisf2_vfp): Likewise.
6364 (*floatsidf2_vfp): Likewise.
6365 (floatunssisf2): Likewise.
6366 (floatunssidf2): Likewise.
6367 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
6368 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
6369 * config/arm/arm1020e.md: Update with new attributes.
6370 * config/arm/cortex-a15-neon.md: Update with new attributes.
6371 * config/arm/cortex-a5.md: Update with new attributes.
6372 * config/arm/cortex-a53.md: Update with new attributes.
6373 * config/arm/cortex-a7.md: Update with new attributes.
6374 * config/arm/cortex-a8-neon.md: Update with new attributes.
6375 * config/arm/cortex-a9.md: Update with new attributes.
6376 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6377 * config/arm/cortex-r4f.md: Update with new attributes.
6378 * config/arm/marvell-pj4.md: Update with new attributes.
6379 * config/arm/vfp11.md: Update with new attributes.
6381 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6383 * config/aarch64/arm_neon.h
6384 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
6385 (vqtbx<1,2,3,4><q>_s8): Likewise.
6387 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6389 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
6390 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
6391 (add<mode>3): Add type attribute.
6392 (add<mode>3): Likewise.
6393 (usadd<mode>3): Likewise.
6394 (ssadd<mode>3): Likewise.
6395 (sub<mode>3): Likewise.
6396 (sub<mode>3): Likewise.
6397 (ussub<mode>3): Likewise.
6398 (sssub<mode>3): Likewise.
6399 (ssmulsa3): Likewise.
6400 (usmulusa3): Likewise.
6401 (arm_usatsihi): Likewise.
6403 (*movdi_vfp): Add types for all instructions.
6404 (*movdi_vfp_cortexa8): Likewise.
6405 (*movhf_vfp_neon): Likewise.
6406 (*movhf_vfp): Likewise.
6407 (*movdf_vfp): Likewise.
6408 (*thumb2_movdf_vfp): Likewise.
6409 (*thumb2_movdfcc_vfp): Likewise.
6410 * config/arm/arm.md: Add type attribute to all insn patterns.
6411 (*thumb1_adddi3): Add type attribute.
6412 (*arm_adddi3): Likewise.
6413 (*adddi_sesidi_di): Likewise.
6414 (*adddi_zesidi_di): Likewise.
6415 (*thumb1_addsi3): Likewise.
6416 (addsi3_compare0): Likewise.
6417 (*addsi3_compare0_scratch): Likewise.
6418 (*compare_negsi_si): Likewise.
6419 (cmpsi2_addneg): Likewise.
6420 (*addsi3_carryin_<optab>): Likewise.
6421 (*addsi3_carryin_alt2_<optab>): Likewise.
6422 (*addsi3_carryin_clobercc_<optab>): Likewise.
6423 (*subsi3_carryin): Likewise.
6424 (*subsi3_carryin_const): Likewise.
6425 (*subsi3_carryin_compare): Likewise.
6426 (*subsi3_carryin_compare_const): Likewise.
6427 (*arm_subdi3): Likewise.
6428 (*thumb_subdi3): Likewise.
6429 (*subdi_di_zesidi): Likewise.
6430 (*subdi_di_sesidi): Likewise.
6431 (*subdi_zesidi_di): Likewise.
6432 (*subdi_sesidi_di): Likewise.
6433 (*subdi_zesidi_ze): Likewise.
6434 (thumb1_subsi3_insn): Likewise.
6435 (*arm_subsi3_insn): Likewise.
6436 (*anddi3_insn): Likewise.
6437 (*anddi_zesidi_di): Likewise.
6438 (*anddi_sesdi_di): Likewise.
6439 (*ne_zeroextracts): Likewise.
6440 (*ne_zeroextracts): Likewise.
6441 (*ite_ne_zeroextr): Likewise.
6442 (*ite_ne_zeroextr): Likewise.
6443 (*anddi_notdi_di): Likewise.
6444 (*anddi_notzesidi): Likewise.
6445 (*anddi_notsesidi): Likewise.
6446 (andsi_notsi_si): Likewise.
6447 (thumb1_bicsi3): Likewise.
6448 (*iordi3_insn): Likewise.
6449 (*iordi_zesidi_di): Likewise.
6450 (*iordi_sesidi_di): Likewise.
6451 (*thumb1_iorsi3_insn): Likewise.
6452 (*xordi3_insn): Likewise.
6453 (*xordi_zesidi_di): Likewise.
6454 (*xordi_sesidi_di): Likewise.
6455 (*arm_xorsi3): Likewise.
6456 (*andsi_iorsi3_no): Likewise.
6457 (*smax_0): Likewise.
6458 (*smax_m1): Likewise.
6459 (*arm_smax_insn): Likewise.
6460 (*smin_0): Likewise.
6461 (*arm_smin_insn): Likewise.
6462 (*arm_umaxsi3): Likewise.
6463 (*arm_uminsi3): Likewise.
6464 (*minmax_arithsi): Likewise.
6465 (*minmax_arithsi_): Likewise.
6466 (*satsi_<SAT:code>): Likewise.
6467 (arm_ashldi3_1bit): Likewise.
6468 (arm_ashrdi3_1bit): Likewise.
6469 (arm_lshrdi3_1bit): Likewise.
6470 (*arm_negdi2): Likewise.
6471 (*thumb1_negdi2): Likewise.
6472 (*arm_negsi2): Likewise.
6473 (*thumb1_negsi2): Likewise.
6474 (*negdi_extendsid): Likewise.
6475 (*negdi_zero_extend): Likewise.
6476 (*arm_abssi2): Likewise.
6477 (*thumb1_abssi2): Likewise.
6478 (*arm_neg_abssi2): Likewise.
6479 (*thumb1_neg_abss): Likewise.
6480 (one_cmpldi2): Likewise.
6481 (extend<mode>di2): Likewise.
6482 (*compareqi_eq0): Likewise.
6483 (*arm_extendhisi2addsi): Likewise.
6484 (*arm_movdi): Likewise.
6485 (*thumb1_movdi_insn): Likewise.
6486 (*arm_movt): Likewise.
6487 (*thumb1_movsi_insn): Likewise.
6488 (pic_add_dot_plus_four): Likewise.
6489 (pic_add_dot_plus_eight): Likewise.
6490 (tls_load_dot_plus_eight): Likewise.
6491 (*thumb1_movhi_insn): Likewise.
6492 (*thumb1_movsf_insn): Likewise.
6493 (*movdf_soft_insn): Likewise.
6494 (*thumb_movdf_insn): Likewise.
6495 (cbranchsi4_insn): Likewise.
6496 (cbranchsi4_scratch): Likewise.
6497 (*negated_cbranchsi4): Likewise.
6498 (*tbit_cbranch): Likewise.
6499 (*tlobits_cbranch): Likewise.
6500 (*tstsi3_cbranch): Likewise.
6501 (*cbranchne_decr1): Likewise.
6502 (*addsi3_cbranch): Likewise.
6503 (*addsi3_cbranch_scratch): Likewise.
6504 (*arm_cmpdi_insn): Likewise.
6505 (*arm_cmpdi_unsig): Likewise.
6506 (*arm_cmpdi_zero): Likewise.
6507 (*thumb_cmpdi_zero): Likewise.
6508 (*deleted_compare): Likewise.
6509 (*mov_scc): Likewise.
6510 (*mov_negscc): Likewise.
6511 (*mov_notscc): Likewise.
6512 (*cstoresi_eq0_thumb1_insn): Likewise.
6513 (cstoresi_nltu_thumb1): Likewise.
6514 (cstoresi_ltu_thu): Likewise.
6515 (thumb1_addsi3_addgeu): Likewise.
6516 (*arm_jump): Likewise.
6517 (*thumb_jump): Likewise.
6518 (*check_arch2): Likewise.
6519 (arm_casesi_internal): Likewise.
6520 (thumb1_casesi_dispatch): Likewise.
6521 (*arm_indirect_jump): Likewise.
6522 (*thumb1_indirect_jump): Likewise.
6524 (*and_scc): Likewise.
6525 (*ior_scc): Likewise.
6526 (*compare_scc): Likewise.
6527 (*cond_move): Likewise.
6528 (*cond_arith): Likewise.
6529 (*cond_sub): Likewise.
6530 (*cmp_ite0): Likewise.
6531 (*cmp_ite1): Likewise.
6532 (*cmp_and): Likewise.
6533 (*cmp_ior): Likewise.
6534 (*ior_scc_scc): Likewise.
6535 (*ior_scc_scc_cmp): Likewise.
6536 (*and_scc_scc): Likewise.
6537 (*and_scc_scc_cmp): Likewise.
6538 (*and_scc_scc_nod): Likewise.
6539 (*negscc): Likewise.
6540 (movcond_addsi): Likewise.
6541 (movcond): Likewise.
6542 (*ifcompare_plus_move): Likewise.
6543 (*if_plus_move): Likewise.
6544 (*ifcompare_move_plus): Likewise.
6545 (*if_move_plus): Likewise.
6546 (*ifcompare_arith_arith): Likewise.
6547 (*if_arith_arith): Likewise.
6548 (*ifcompare_arith_move): Likewise.
6549 (*if_arith_move): Likewise.
6550 (*ifcompare_move_arith): Likewise.
6551 (*if_move_arith): Likewise.
6552 (*ifcompare_move_not): Likewise.
6553 (*if_move_not): Likewise.
6554 (*ifcompare_not_move): Likewise.
6555 (*if_not_move): Likewise.
6556 (*ifcompare_shift_move): Likewise.
6557 (*if_shift_move): Likewise.
6558 (*ifcompare_move_shift): Likewise.
6559 (*if_move_shift): Likewise.
6560 (*ifcompare_shift_shift): Likewise.
6561 (*ifcompare_not_arith): Likewise.
6562 (*ifcompare_arith_not): Likewise.
6563 (*if_arith_not): Likewise.
6564 (*ifcompare_neg_move): Likewise.
6565 (*if_neg_move): Likewise.
6566 (*ifcompare_move_neg): Likewise.
6567 (*if_move_neg): Likewise.
6568 (prologue_thumb1_interwork): Likewise.
6569 (*cond_move_not): Likewise.
6570 (*sign_extract_onebit): Likewise.
6571 (*not_signextract_onebit): Likewise.
6572 (stack_tie): Likewise.
6573 (align_4): Likewise.
6574 (align_8): Likewise.
6575 (consttable_end): Likewise.
6576 (consttable_1): Likewise.
6577 (consttable_2): Likewise.
6578 (consttable_4): Likewise.
6579 (consttable_8): Likewise.
6580 (consttable_16): Likewise.
6581 (*thumb1_tablejump): Likewise.
6582 (prefetch): Likewise.
6583 (force_register_use): Likewise.
6584 (thumb_eh_return): Likewise.
6585 (load_tp_hard): Likewise.
6586 (load_tp_soft): Likewise.
6587 (tlscall): Likewise.
6588 (*arm_movtas_ze): Likewise.
6589 (*arm_rev): Likewise.
6590 (*arm_revsh): Likewise.
6591 (*arm_rev16): Likewise.
6592 * config/arm/thumb2.md
6593 (*thumb2_smaxsi3): Likewise.
6594 (*thumb2_sminsi3): Likewise.
6595 (*thumb32_umaxsi3): Likewise.
6596 (*thumb2_uminsi3): Likewise.
6597 (*thumb2_negdi2): Likewise.
6598 (*thumb2_abssi2): Likewise.
6599 (*thumb2_neg_abss): Likewise.
6600 (*thumb2_movsi_insn): Likewise.
6601 (tls_load_dot_plus_four): Likewise.
6602 (*thumb2_movhi_insn): Likewise.
6603 (*thumb2_mov_scc): Likewise.
6604 (*thumb2_mov_negs): Likewise.
6605 (*thumb2_mov_negs): Likewise.
6606 (*thumb2_mov_nots): Likewise.
6607 (*thumb2_mov_nots): Likewise.
6608 (*thumb2_movsicc_): Likewise.
6609 (*thumb2_movsfcc_soft_insn): Likewise.
6610 (*thumb2_indirect_jump): Likewise.
6611 (*thumb2_and_scc): Likewise.
6612 (*thumb2_ior_scc): Likewise.
6613 (*thumb2_ior_scc_strict_it): Likewise.
6614 (*thumb2_cond_move): Likewise.
6615 (*thumb2_cond_arith): Likewise.
6616 (*thumb2_cond_ari): Likewise.
6617 (*thumb2_cond_sub): Likewise.
6618 (*thumb2_negscc): Likewise.
6619 (*thumb2_movcond): Likewise.
6620 (thumb2_casesi_internal): Likewise.
6621 (thumb2_casesi_internal_pic): Likewise.
6622 (*thumb2_alusi3_short): Likewise.
6623 (*thumb2_mov<mode>_shortim): Likewise.
6624 (*thumb2_addsi_short): Likewise.
6625 (*thumb2_subsi_short): Likewise.
6626 (thumb2_addsi3_compare0): Likewise.
6627 (*thumb2_cbz): Likewise.
6628 (*thumb2_cbnz): Likewise.
6629 (*thumb2_one_cmplsi2_short): Likewise.
6630 (*thumb2_negsi2_short): Likewise.
6631 (*orsi_notsi_si): Likewise.
6632 * config/arm/arm1020e.md: Update with new attributes.
6633 * config/arm/arm1026ejs.md: Update with new attributes.
6634 * config/arm/arm1136jfs.md: Update with new attributes.
6635 * config/arm/arm926ejs.md: Update with new attributes.
6636 * config/arm/cortex-a15.md: Update with new attributes.
6637 * config/arm/cortex-a5.md: Update with new attributes.
6638 * config/arm/cortex-a53.md: Update with new attributes.
6639 * config/arm/cortex-a7.md: Update with new attributes.
6640 * config/arm/cortex-a8.md: Update with new attributes.
6641 * config/arm/cortex-a9.md: Update with new attributes.
6642 * config/arm/cortex-m4.md: Update with new attributes.
6643 * config/arm/cortex-r4.md: Update with new attributes.
6644 * config/arm/fa526.md: Update with new attributes.
6645 * config/arm/fa606te.md: Update with new attributes.
6646 * config/arm/fa626te.md: Update with new attributes.
6647 * config/arm/fa726te.md: Update with new attributes.
6649 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6651 * config/aarch64/aarch64-simd.md
6652 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
6653 <vwx> iterator to ensure correct register choice.
6654 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
6655 (aarch64_sqdmull_n<mode>): Likewise.
6656 (aarch64_sqdmull2_n<mode>_internal): Likewise.
6657 * config/aarch64/arm_neon.h
6658 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
6659 (vml<as><q>_n_<su>16): Likewise.
6660 (vml<as>l_high_lane<q>_<su>16): Likewise.
6661 (vml<as>l_high_n_<su>16): Likewise.
6662 (vml<as>l_lane<q>_<su>16): Likewise.
6663 (vml<as>l_n_<su>16): Likewise.
6664 (vmul<q>_lane<q>_<su>16): Likewise.
6665 (vmul<q>_n_<su>16): Likewise.
6666 (vmull_lane<q>_<su>16): Likewise.
6667 (vmull_n_<su>16): Likewise.
6668 (vmull_high_lane<q>_<su>16): Likewise.
6669 (vmull_high_n_<su>16): Likewise.
6670 (vqrdmulh<q>_n_s16): Likewise.
6672 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
6674 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
6675 have the correct lane parameter.
6677 2013-09-06 Richard Biener <rguenther@suse.de>
6679 * cfganal.c (control_dependences::~control_dependences):
6680 Properly free all of the vector.
6682 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
6685 * config/i386/i386.c (ix86_conditional_register_usage):
6686 Proper initialize extended SSE registers.
6688 2013-09-06 Jan Hubicka <jh@suse.cz>
6690 PR tree-optimization/58311
6691 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
6693 2013-09-06 Jan Hubicka <jh@suse.cz>
6695 * Makefile.in (tree-sra.o): Update dependencies.
6696 * tree-sra.c: Include ipa-utils.h
6697 (scan_function): Use recursive_call_p.
6698 (has_caller_p): New function.
6699 (cgraph_for_node_and_aliases): Count also callers of aliases.
6701 2013-09-06 Jan Hubicka <jh@suse.cz>
6704 * cgraph.h (symtab_semantically_equivalent_p): Declare.
6705 * tree-tailcall.c: Include ipa-utils.h.
6706 (find_tail_calls): Use it.
6707 * ipa-pure-const.c (check_call): Likewise.
6708 * ipa-utils.c (recursive_call_p): New function.
6709 * ipa-utils.h (recursive_call_p): Dclare.
6710 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
6711 (symtab_semantically_equivalent_p): New function.
6712 * Makefile.in (tree-tailcall.o): Update dependencies.
6714 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6716 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
6719 2013-09-06 Richard Biener <rguenther@suse.de>
6721 * lto-streamer.h (lto_global_var_decls): Remove.
6722 * Makefile.in (OBJS): Remove lto-symtab.o.
6723 (lto-symtab.o): Remove.
6724 (GTFILES): Remove lto-symtab.c
6725 * lto-symtab.c: Move to lto/
6727 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6729 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
6730 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
6731 (UNSPEC_FPINT_RINT): New constant definitions.
6732 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
6733 definition with 2 attributes.
6734 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
6735 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
6738 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6740 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
6741 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
6742 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
6743 attribute to "z196_alone".
6744 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
6747 2013-09-06 Richard Biener <rguenther@suse.de>
6749 * basic-block.h (class control_dependences): New.
6750 * tree-ssa-dce.c (control_dependence_map): Remove.
6752 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
6753 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
6754 find_pdom, find_control_dependence, find_all_control_dependences):
6756 (mark_control_dependent_edges_necessary,
6757 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
6758 tree_dce_done, perform_tree_ssa_dce): Adjust.
6759 * cfganal.c (set_control_dependence_map_bit,
6760 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
6761 find_all_control_dependences): Move from tree-ssa-dce.c and
6762 implement as methods of control_dependences class.
6763 (control_dependences::control_dependences): New.
6764 (control_dependences::~control_dependences): Likewise.
6765 (control_dependences::get_edges_dependent_on): Likewise.
6766 (control_dependences::get_edge): Likewise.
6768 2013-09-04 Jan Hubicka <jh@suse.cz>
6770 * tree.c (types_same_for_odr): Drop overactive check.
6771 * ipa-devirt.c (hash_type_name): Likewise.
6773 2013-09-04 Jan Hubicka <jh@suse.cz>
6775 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
6776 (analyze_functions): ... here.
6778 2013-09-04 Jan Hubicka <jh@suse.cz>
6781 * cgraphunit.c (analyze_functions): Clear AUX fields
6782 after processing; initialize assembler name has.
6784 2013-09-05 Jeff Law <law@redhat.com>
6786 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
6787 from thread_around_empty_block. Record threading path into PATH.
6788 Recurse if threading through the initial block is successful.
6789 (thread_across_edge): Corresponding changes to slightly simplify.
6791 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6793 * config/aarch64/aarch64.md
6794 (type): Remove frecpe, frecps, frecpx.
6795 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
6796 fix to be a TARGET_SIMD instruction.
6797 (aarch64_frecps): Remove.
6798 * config/aarch64/aarch64-simd.md
6799 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
6800 (aarch64_frecps<mode>): Handle all float/vector of float modes.
6802 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6803 Sofiane Naci <sofiane.naci@arm.com>
6805 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
6806 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
6807 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
6808 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
6809 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
6810 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
6811 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
6812 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
6813 into "clz, "rbit". Rename "shift" to "shift_imm".
6814 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
6815 changes. Update for attribute changes all occurrences of arlo_* and
6817 * config/arm/arm-fixed.md: Update for attribute changes
6818 all occurrences of arlo_* types.
6819 * config/arm/thumb2.md: Update for attribute changes all occurrences
6821 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
6822 (cortexa7_older_only): Likewise.
6823 (cortexa7_younger): Likewise.
6824 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
6825 (1020alu_shift_op): Likewise.
6826 (1020alu_shift_reg_op): Likewise.
6827 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
6828 (alu_shift_op): Likewise.
6829 (alu_shift_reg_op): Likewise.
6830 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
6831 (11_alu_shift_op): Likewise.
6832 (11_alu_shift_reg_op): Likewise.
6833 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
6834 (9_alu_shift_reg_op): Likewise.
6835 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
6837 (cortex_a15_alu_shift): Likewise.
6838 (cortex_a15_alu_shift_reg): Likewise.
6839 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
6841 (cortex_a5_alu_shift): Likewise.
6842 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
6844 (cortex_a53_alu_shift): Likewise.
6845 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
6847 (cortex_a7_alu_reg): Likewise.
6848 (cortex_a7_alu_shift): Likewise.
6849 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
6851 (cortex_a8_alu_shift): Likewise.
6852 (cortex_a8_alu_shift_reg): Likewise.
6853 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
6854 (cortex_a9_dp_shift): Likewise.
6855 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
6857 * config/arm/cortex-r4.md
6858 (cortex_r4_alu): Update for attribute changes.
6859 (cortex_r4_mov): Likewise.
6860 (cortex_r4_alu_shift_reg): Likewise.
6861 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
6862 (526_alu_shift_op): Likewise.
6863 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
6864 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
6865 (626te_alu_shift_op): Likewise.
6866 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
6867 (726te_alu_shift_op): Likewise.
6868 (726te_alu_shift_reg_op): Likewise.
6869 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
6870 (mp626_alu_shift_op): Likewise.
6871 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
6872 (pj4_alu_conds): Likewise.
6873 (pj4_shift): Likewise.
6874 (pj4_shift_conds): Likewise.
6875 (pj4_alu_shift): Likewise.
6876 (pj4_alu_shift_conds): Likewise.
6877 * config/aarch64/aarch64.md: Update for attribute change
6878 all occurrences of arlo_* and shift* types.
6880 2013-09-05 Mike Stump <mikestump@comcast.net>
6882 * tree.h: Move documentation for tree_function_decl to tree-core.h
6883 with the declaration.
6885 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
6888 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
6889 looking for widest mode.
6891 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
6893 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
6895 2013-09-05 Richard Biener <rguenther@suse.de>
6897 PR tree-optimization/58137
6898 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
6899 Do not create vectors of pointers.
6900 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
6901 types for the components of the vector initializer.
6902 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
6903 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
6905 2013-09-05 Martin Jambor <mjambor@suse.cz>
6907 * ipa-prop.c (remove_described_reference): Accept missing references,
6908 return false if that hppens, otherwise return true.
6909 (cgraph_node_for_jfunc): New function.
6910 (try_decrement_rdesc_refcount): Likewise.
6911 (try_make_edge_direct_simple_call): Use them.
6912 (ipa_edge_removal_hook): Remove references from rdescs.
6913 (ipa_edge_duplication_hook): Clone rdescs and their references
6914 when the new edge has the same caller as the old one.
6915 * cgraph.c (cgraph_resolve_speculation): Remove speculative
6916 reference before removing any edges.
6918 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
6920 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
6922 * thumb2.md (thumb2_storewb_parisi): New pattern.
6924 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
6926 * config/aarch64/aarch64-option-extensions.def: Add
6927 AARCH64_OPT_EXTENSION of 'crc'.
6928 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
6929 (AARCH64_ISA_CRC): Ditto.
6930 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
6931 description of the CRC extension.
6933 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
6935 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
6936 * config/rs6000/linux.h: Ditto.
6937 * alpha/linux.h: Ditto.
6938 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
6939 no_c99_libc_has_function.
6940 * config/c6x/uclinux-elf.h: Ditto.
6941 * config/lm32/uclinux-elf.h: Ditto.
6942 * config/m68k/uclinux.h: Ditto.
6943 * config/moxie/uclinux.h: Ditto.
6944 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
6945 (crisv32-*-linux*, cris-*-linux*): Ditto.
6946 * config/bfin/bfin.c: Include "tm_p.h".
6948 2013-09-05 Richard Biener <rguenther@suse.de>
6950 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
6951 check for a definition without a basic-block.
6953 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6954 Sofiane Naci <sofiane.naci@arm.com>
6956 * config/aarch64/aarch64.md
6957 (*movti_aarch64): Rename r_2_f and f_2_r.
6958 (*movsf_aarch64): Likewise.
6959 (*movdf_aarch64): Likewise.
6960 (*movtf_aarch64): Likewise.
6961 (aarch64_movdi_<mode>low): Likewise.
6962 (aarch64_movdi_<mode>high): Likewise.
6963 (aarch64_mov<mode>high_di): Likewise.
6964 (aarch64_mov<mode>low_di): Likewise.
6965 (aarch64_movtilow_tilow): Likewise.
6966 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
6967 values to config/arm/types.md
6968 (attribute "conds"): Update for attribute change.
6969 (anddi3_insn): Likewise.
6970 (iordi3_insn): Likewise.
6971 (xordi3_insn): Likewise.
6972 (one_cmpldi2): Likewise.
6973 * config/arm/types.md (type): Add Neon types.
6974 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
6975 use "type" attribute.
6976 (movmisalign<mode>_neon_store): Likewise.
6977 (movmisalign<mode>_neon_load): Likewise.
6978 (vec_set<mode>_internal): Likewise.
6979 (vec_setv2di_internal): Likewise.
6980 (vec_extract<mode>): Likewise.
6981 (vec_extractv2di): Likewise.
6982 (add<mode>3_neon): Likewise.
6983 (adddi3_neon): Likewise.
6984 (sub<mode>3_neon): Likewise.
6985 (subdi3_neon): Likewise.
6986 (mul<mode>3_neon): Likewise.
6987 (mul<mode>3add<mode>_neon): Likewise.
6988 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
6989 (fma<VCVTF:mode>4)): Likewise.
6990 (fma<VCVTF:mode>4_intrinsic): Likewise.
6991 (fmsub<VCVTF:mode>4)): Likewise.
6992 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
6993 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
6994 (ior<mode>3): Likewise.
6995 (and<mode>3): Likewise.
6996 (anddi3_neon): Likewise.
6997 (orn<mode>3_neon): Likewise.
6998 (orndi3_neon): Likewise.
6999 (bic<mode>3_neon): Likewise.
7000 (bicdi3_neon): Likewise.
7001 (xor<mode>3): Likewise.
7002 (one_cmpl<mode>2): Likewise.
7003 (abs<mode>2): Likewise.
7004 (neg<mode>2): Likewise.
7005 (umin<mode>3_neon): Likewise.
7006 (umax<mode>3_neon): Likewise.
7007 (smin<mode>3_neon): Likewise.
7008 (smax<mode>3_neon): Likewise.
7009 (vashl<mode>3): Likewise.
7010 (vashr<mode>3_imm): Likewise.
7011 (vlshr<mode>3_imm): Likewise.
7012 (ashl<mode>3_signed): Likewise.
7013 (ashl<mode>3_unsigned): Likewise.
7014 (neon_load_count): Likewise.
7015 (ashldi3_neon_noclobber): Likewise.
7016 (signed_shift_di3_neon): Likewise.
7017 (unsigned_shift_di3_neon): Likewise.
7018 (ashrdi3_neon_imm_noclobber): Likewise.
7019 (lshrdi3_neon_imm_noclobber): Likewise.
7020 (widen_ssum<mode>3): Likewise.
7021 (widen_usum<mode>3): Likewise.
7022 (quad_halves_<code>v4si): Likewise.
7023 (quad_halves_<code>v4sf): Likewise.
7024 (quad_halves_<code>v8hi): Likewise.
7025 (quad_halves_<code>v16qi): Likewise.
7026 (reduc_splus_v2di): Likewise.
7027 (neon_vpadd_internal<mode>): Likewise.
7028 (neon_vpsmin<mode>): Likewise.
7029 (neon_vpsmax<mode>): Likewise.
7030 (neon_vpumin<mode>): Likewise.
7031 (neon_vpumax<mode>): Likewise.
7032 (ss_add<mode>_neon): Likewise.
7033 (us_add<mode>_neon): Likewise.
7034 (ss_sub<mode>_neon): Likewise.
7035 (us_sub<mode>_neon): Likewise.
7036 (neon_vadd<mode>_unspec): Likewise.
7037 (neon_vaddl<mode>): Likewise.
7038 (neon_vaddw<mode>): Likewise.
7039 (neon_vhadd<mode>): Likewise.
7040 (neon_vqadd<mode>): Likewise.
7041 (neon_vaddhn<mode>): Likewise.
7042 (neon_vmul<mode>): Likewise.
7043 (neon_vmla<mode>): Likewise.
7044 (neon_vmlal<mode>): Likewise.
7045 (neon_vmls<mode>): Likewise.
7046 (neon_vmlsl<mode>): Likewise.
7047 (neon_vqdmulh<mode>): Likewise.
7048 (neon_vqdmlal<mode>): Likewise.
7049 (neon_vqdmlsl<mode>): Likewise.
7050 (neon_vmull<mode>): Likewise.
7051 (neon_vqdmull<mode>): Likewise.
7052 (neon_vsub<mode>_unspec): Likewise.
7053 (neon_vsubl<mode>): Likewise.
7054 (neon_vsubw<mode>): Likewise.
7055 (neon_vqsub<mode>): Likewise.
7056 (neon_vhsub<mode>): Likewise.
7057 (neon_vsubhn<mode>): Likewise.
7058 (neon_vceq<mode>): Likewise.
7059 (neon_vcge<mode>): Likewise.
7060 (neon_vcgeu<mode>): Likewise.
7061 (neon_vcgt<mode>): Likewise.
7062 (neon_vcgtu<mode>): Likewise.
7063 (neon_vcle<mode>): Likewise.
7064 (neon_vclt<mode>): Likewise.
7065 (neon_vcage<mode>): Likewise.
7066 (neon_vcagt<mode>): Likewise.
7067 (neon_vtst<mode>): Likewise.
7068 (neon_vabd<mode>): Likewise.
7069 (neon_vabdl<mode>): Likewise.
7070 (neon_vaba<mode>): Likewise.
7071 (neon_vabal<mode>): Likewise.
7072 (neon_vmax<mode>): Likewise.
7073 (neon_vmin<mode>): Likewise.
7074 (neon_vpaddl<mode>): Likewise.
7075 (neon_vpadal<mode>): Likewise.
7076 (neon_vpmax<mode>): Likewise.
7077 (neon_vpmin<mode>): Likewise.
7078 (neon_vrecps<mode>): Likewise.
7079 (neon_vrsqrts<mode>): Likewise.
7080 (neon_vqabs<mode>): Likewise.
7081 (neon_vqneg<mode>): Likewise.
7082 (neon_vcls<mode>): Likewise.
7083 (clz<mode>2): Likewise.
7084 (popcount<mode>2): Likewise.
7085 (neon_vrecpe): Likewise.
7086 (neon_vrsqrte): Likewise.
7087 (neon_vget_lane<mode>_sext_internal): Likewise.
7088 (neon_vget_lane<mode>_zext_internal): Likewise.
7089 (neon_vdup_n<mode>): Likewise.
7090 (neon_vdup_nv2di): Likewise.
7091 (neon_vdpu_lane<mode>_internal): Likewise.
7092 (neon_vswp<mode>): Likewise.
7093 (float<mode><V_cvtto>2): Likewise.
7094 (floatuns<mode><V_cvtto>2): Likewise.
7095 (fix_trunc<mode><V_cvtto>)2): Likewise
7096 (fixuns_trunc<mode><V_cvtto)2): Likewise.
7097 (neon_vcvt<mode>): Likewise.
7098 (neon_vcvtv4sfv4hf): Likewise.
7099 (neon_vcvtv4hfv4sf): Likewise.
7100 (neon_vcvt_n<mode>): Likewise.
7101 (neon_vmovn<mode>): Likewise.
7102 (neon_vqmovn<mode>): Likewise.
7103 (neon_vqmovun<mode>): Likewise.
7104 (neon_vmovl<mode>): Likewise.
7105 (neon_vmul_lane<mode>): Likewise.
7106 (neon_vmull_lane<mode>): Likewise.
7107 (neon_vqdmull_lane<mode>): Likewise.
7108 (neon_vqdmulh_lane<mode>): Likewise.
7109 (neon_vmla_lane<mode>): Likewise.
7110 (neon_vmlal_lane<mode>): Likewise.
7111 (neon_vqdmlal_lane<mode>): Likewise.
7112 (neon_vmls_lane<mode>): Likewise.
7113 (neon_vmlsl_lane<mode>): Likewise.
7114 (neon_vqdmlsl_lane<mode>): Likewise.
7115 (neon_vext<mode>): Likewise.
7116 (neon_vrev64<mode>): Likewise.
7117 (neon_vrev32<mode>): Likewise.
7118 (neon_vrev16<mode>): Likewise.
7119 (neon_vbsl<mode>_internal): Likewise.
7120 (neon_vshl<mode>): Likewise.
7121 (neon_vqshl<mode>): Likewise.
7122 (neon_vshr_n<mode>): Likewise.
7123 (neon_vshrn_n<mode>): Likewise.
7124 (neon_vqshrn_n<mode>): Likewise.
7125 (neon_vqshrun_n<mode>): Likewise.
7126 (neon_vshl_n<mode>): Likewise.
7127 (neon_vqshl_n<mode>): Likewise.
7128 (neon_vqshlu_n<mode>): Likewise.
7129 (neon_vshll_n<mode>): Likewise.
7130 (neon_vsra_n<mode>): Likewise.
7131 (neon_vsri_n<mode>): Likewise.
7132 (neon_vsli_n<mode>): Likewise.
7133 (neon_vtbl1v8qi): Likewise.
7134 (neon_vtbl2v8qi): Likewise.
7135 (neon_vtbl3v8qi): Likewise.
7136 (neon_vtbl4v8qi): Likewise.
7137 (neon_vtbx1v8qi): Likewise.
7138 (neon_vtbx2v8qi): Likewise.
7139 (neon_vtbx3v8qi): Likewise.
7140 (neon_vtbx4v8qi): Likewise.
7141 (neon_vtrn<mode>_internal): Likewise.
7142 (neon_vzip<mode>_internal): Likewise.
7143 (neon_vuzp<mode>_internal): Likewise.
7144 (neon_vld1<mode>): Likewise.
7145 (neon_vld1_lane<mode>): Likewise.
7146 (neon_vld1_dup<mode>): Likewise.
7147 (neon_vld1_dupv2di): Likewise.
7148 (neon_vst1<mode>): Likewise.
7149 (neon_vst1_lane<mode>): Likewise.
7150 (neon_vld2<mode>): Likewise.
7151 (neon_vld2_lane<mode>): Likewise.
7152 (neon_vld2_dup<mode>): Likewise.
7153 (neon_vst2<mode>): Likewise.
7154 (neon_vst2_lane<mode>): Likewise.
7155 (neon_vld3<mode>): Likewise.
7156 (neon_vld3qa<mode>): Likewise.
7157 (neon_vld3qb<mode>): Likewise.
7158 (neon_vld3_lane<mode>): Likewise.
7159 (neon_vld3_dup<mode>): Likewise.
7160 (neon_vst3<mode>): Likewise.
7161 (neon_vst3qa<mode>): Likewise.
7162 (neon_vst3qb<mode>): Likewise.
7163 (neon_vst3_lane<mode>): Likewise.
7164 (neon_vld4<mode>): Likewise.
7165 (neon_vld4qa<mode>): Likewise.
7166 (neon_vld4qb<mode>): Likewise.
7167 (neon_vld4_lane<mode>): Likewise.
7168 (neon_vld4_dup<mode>): Likewise.
7169 (neon_vst4<mode>): Likewise.
7170 (neon_vst4qa<mode>): Likewise.
7171 (neon_vst4qb<mode>): Likewise.
7172 (neon_vst4_lane<mode>): Likewise.
7173 (neon_vec_unpack<US>_lo_<mode>): Likewise.
7174 (neon_vec_unpack<US>_hi_<mode>): Likewise.
7175 (neon_vec_<US>mult_lo_<mode>): Likewise.
7176 (neon_vec_<US>mult_hi_<mode>): Likewise.
7177 (neon_vec_<US>shiftl_<mode>): Likewise.
7178 (neon_unpack<US>_<mode>): Likewise.
7179 (neon_vec_<US>mult_<mode>): Likewise.
7180 (vec_pack_trunc_<mode>): Likewise.
7181 (neon_vec_pack_trunk_<mode>): Likewise.
7182 (neon_vabd<mode>_2): Likewise.
7183 (neon_vabd<mode>_3): Likewise.
7184 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
7185 (thumb2_movsi_vfp): Likewise.
7186 (movdi_vfp): Likewise.
7187 (movdi_vfp_cortexa8): Likewise.
7188 (movhf_vfp_neon): Likewise.
7189 (movhf_vfp): Likewiwse.
7190 (movsf_vfp): Likewiwse.
7191 (thumb2_movsf_vfp): Likewiwse.
7192 (movdf_vfp): Likewise.
7193 (thumb2_movdf_vfp): Likewise.
7194 (movsfcc_vfp): Likewise.
7195 (thumb2_movsfcc_vfp): Likewise.
7196 (movdfcc_vfp): Likewise.
7197 (thumb2_movdfcc_vfp): Likewise.
7198 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
7199 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
7200 (v10_v2c): Likewise.
7201 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
7203 (cortex_a15_neon_int_2): Likewise.
7204 (cortex_a15_neon_int_3): Likewise.
7205 (cortex_a15_neon_int_4): Likewise.
7206 (cortex_a15_neon_int_5): Likewise.
7207 (cortex_a15_neon_vqneg_vqabs): Likewise.
7208 (cortex_a15_neon_vmov): Likewise.
7209 (cortex_a15_neon_vaba): Likewise.
7210 (cortex_a15_neon_vaba_qqq): Likewise.
7211 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7212 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7213 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
7214 scalar_64_32_long_scalar): Likewise.
7215 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7216 (cortex_a15_neon_mla_qqq_8_16): Likewise.
7217 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7218 lotype_qdd_64_32_long): Likewise.
7219 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7220 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7221 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
7222 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7223 (cortex_a15_neon_shift_1): Likewise.
7224 (cortex_a15_neon_shift_2): Likewise.
7225 (cortex_a15_neon_shift_3): Likewise.
7226 (cortex_a15_neon_vshl_ddd): Likewise.
7227 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7228 (cortex_a15_neon_vsra_vrsra): Likewise.
7229 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
7230 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
7231 (cortex_a15_neon_fp_vmul_ddd): Likewise.
7232 (cortex_a15_neon_fp_vmul_qqd): Likewise.
7233 (cortex_a15_neon_fp_vmla_ddd): Likewise.
7234 (cortex_a15_neon_fp_vmla_qqq): Likewise.
7235 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
7236 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
7237 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7238 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7239 (cortex_a15_neon_bp_simple): Likewise.
7240 (cortex_a15_neon_bp_2cycle): Likewise.
7241 (cortex_a15_neon_bp_3cycle): Likewise.
7242 (cortex_a15_neon_vld1_1_2_regs): Likewise.
7243 (cortex_a15_neon_vld1_3_4_regs): Likewise.
7244 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7245 (cortex_a15_neon_vld2_4_regs): Likewise.
7246 (cortex_a15_neon_vld3_vld4): Likewise.
7247 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7248 (cortex_a15_neon_vst1_3_4_regs): Likewise.
7249 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
7250 (cortex_a15_neon_vst3_vst4): Likewise.
7251 (cortex_a15_neon_vld1_vld2_lane): Likewise.
7252 (cortex_a15_neon_vld3_vld4_lane" 10
7253 (cortex_a15_neon_vst1_vst2_lane): Likewise.
7254 (cortex_a15_neon_vst3_vst4_lane): Likewise.
7255 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
7256 (cortex_a15_neon_ldm_2): Likewise.0
7257 (cortex_a15_neon_stm_2): Likewise.
7258 (cortex_a15_neon_mcr): Likewise.
7259 (cortex_a15_neon_mcr_2_mcrr): Likewise.
7260 (cortex_a15_neon_mrc): Likewise.
7261 (cortex_a15_neon_mrrc): Likewise.
7262 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
7264 (cortex_a15_alu_shift): Likewise.
7265 (cortex_a15_alu_shift_reg): Likewise.
7266 (cortex_a15_mult32): Likewise.
7267 (cortex_a15_mult64): Likewise.
7268 (cortex_a15_block): Likewise.
7269 (cortex_a15_branch): Likewise.
7270 (cortex_a15_load1): Likewise.
7271 (cortex_a15_load3): Likewise.
7272 (cortex_a15_store1): Likewise.
7273 (cortex_a15_store3): Likewise.
7274 (cortex_a15_call): Likewise.
7275 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
7276 (cortex_a5_f2r): Likewise.
7277 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
7279 (cortex_a53_f2r): Likewise.
7280 * config/arm/cortex-a7.md
7281 (cortex_a7_branch): Update for attribute change.
7282 (cortex_a7_call): Likewise.
7283 (cortex_a7_alu_imm): Likewise.
7284 (cortex_a7_alu_reg): Likewise.
7285 (cortex_a7_alu_shift): Likewise.
7286 (cortex_a7_mul): Likewise.
7287 (cortex_a7_load1): Likewise.
7288 (cortex_a7_store1): Likewise.
7289 (cortex_a7_load2): Likewise.
7290 (cortex_a7_store2): Likewise.
7291 (cortex_a7_load3): Likewise.
7292 (cortex_a7_store3): Likewise.
7293 (cortex_a7_load4): Likewise.
7294 (cortex_a7_store4): Likewise.
7295 (cortex_a7_fpalu): Likewise.
7296 (cortex_a7_fconst): Likewise.
7297 (cortex_a7_fpmuls): Likewise.
7298 (cortex_a7_neon_mul): Likewise.
7299 (cortex_a7_fpmacs): Likewise.
7300 (cortex_a7_neon_mla: Likewise.
7301 (cortex_a7_fpmuld: Likewise.
7302 (cortex_a7_fpmacd: Likewise.
7303 (cortex_a7_fpfmad: Likewise.
7304 (cortex_a7_fdivs: Likewise.
7305 (cortex_a7_fdivd: Likewise.
7306 (cortex_a7_r2f: Likewise.
7307 (cortex_a7_f2r: Likewise.
7308 (cortex_a7_f_flags: Likewise.
7309 (cortex_a7_f_loads: Likewise.
7310 (cortex_a7_f_loadd: Likewise.
7311 (cortex_a7_f_stores: Likewise.
7312 (cortex_a7_f_stored: Likewise.
7313 (cortex_a7_neon): Likewise.
7314 * config/arm/cortex-a8-neon.md
7315 (cortex_a8_neon_mrc): Update for attribute change.
7316 (cortex_a8_neon_mrrc): Likewise.
7317 (cortex_a8_neon_int_1): Likewise.
7318 (cortex_a8_neon_int_2): Likewise.
7319 (cortex_a8_neon_int_3): Likewise.
7320 (cortex_a8_neon_int_4): Likewise.
7321 (cortex_a8_neon_int_5): Likewise.
7322 (cortex_a8_neon_vqneg_vqabs): Likewise.
7323 (cortex_a8_neon_vmov): Likewise.
7324 (cortex_a8_neon_vaba): Likewise.
7325 (cortex_a8_neon_vaba_qqq): Likewise.
7326 (cortex_a8_neon_vsma): Likewise.
7327 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7328 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7329 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7330 long_scalar): Likewise.
7331 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7332 (cortex_a8_neon_mla_qqq_8_16): Likewise.
7333 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7334 long_scalar_qdd_64_32_long): Likewise.
7335 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7336 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7337 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
7338 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7339 (cortex_a8_neon_shift_1): Likewise.
7340 (cortex_a8_neon_shift_2): Likewise.
7341 (cortex_a8_neon_shift_3): Likewise.
7342 (cortex_a8_neon_vshl_ddd): Likewise.
7343 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7344 (cortex_a8_neon_vsra_vrsra): Likewise.
7345 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
7346 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
7347 (cortex_a8_neon_fp_vsum): Likewise.
7348 (cortex_a8_neon_fp_vmul_ddd): Likewise.
7349 (cortex_a8_neon_fp_vmul_qqd): Likewise.
7350 (cortex_a8_neon_fp_vmla_ddd): Likewise.
7351 (cortex_a8_neon_fp_vmla_qqq): Likewise.
7352 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
7353 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
7354 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7355 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7356 (cortex_a8_neon_bp_simple): Likewise.
7357 (cortex_a8_neon_bp_2cycle): Likewise.
7358 (cortex_a8_neon_bp_3cycle): Likewise.
7359 (cortex_a8_neon_ldr): Likewise.
7360 (cortex_a8_neon_str): Likewise.
7361 (cortex_a8_neon_vld1_1_2_regs): Likewise.
7362 (cortex_a8_neon_vld1_3_4_regs): Likewise.
7363 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7364 (cortex_a8_neon_vld2_4_regs): Likewise.
7365 (cortex_a8_neon_vld3_vld4): Likewise.
7366 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7367 (cortex_a8_neon_vst1_3_4_regs): Likewise.
7368 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
7369 (cortex_a8_neon_vst3_vst4): Likewise.
7370 (cortex_a8_neon_vld1_vld2_lane): Likewise.
7371 (cortex_a8_neon_vld3_vld4_lane): Likewise.
7372 (cortex_a8_neon_vst1_vst2_lane): Likewise.
7373 (cortex_a8_neon_vst3_vst4_lane): Likewise.
7374 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
7375 (cortex_a8_neon_mcr): Likewise.
7376 (cortex_a8_neon_mcr_2_mcrr): Likewise.
7377 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
7378 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
7380 (ca9_neon_mrrc): Likewise.
7381 (cortex_a9_neon_int_1): Likewise.
7382 (cortex_a9_neon_int_2): Likewise.
7383 (cortex_a9_neon_int_3): Likewise.
7384 (cortex_a9_neon_int_4): Likewise.
7385 (cortex_a9_neon_int_5): Likewise.
7386 (cortex_a9_neon_vqneg_vqabs): Likewise.
7387 (cortex_a9_neon_vmov): Likewise.
7388 (cortex_a9_neon_vaba): Likewise.
7389 (cortex_a9_neon_vaba_qqq): Likewise.
7390 (cortex_a9_neon_vsma): Likewise.
7391 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7392 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7393 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7394 long_scalar): Likewise.
7395 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7396 (cortex_a9_neon_mla_qqq_8_16): Likewise.
7397 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7398 long_scalar_qdd_64_32_long): Likewise.
7399 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7400 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7401 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
7402 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7403 (cortex_a9_neon_shift_1): Likewise.
7404 (cortex_a9_neon_shift_2): Likewise.
7405 (cortex_a9_neon_shift_3): Likewise.
7406 (cortex_a9_neon_vshl_ddd): Likewise.
7407 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7408 (cortex_a9_neon_vsra_vrsra): Likewise.
7409 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
7410 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
7411 (cortex_a9_neon_fp_vsum): Likewise.
7412 (cortex_a9_neon_fp_vmul_ddd): Likewise.
7413 (cortex_a9_neon_fp_vmul_qqd): Likewise.
7414 (cortex_a9_neon_fp_vmla_ddd): Likewise.
7415 (cortex_a9_neon_fp_vmla_qqq): Likewise.
7416 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
7417 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
7418 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7419 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7420 (cortex_a9_neon_bp_simple): Likewise.
7421 (cortex_a9_neon_bp_2cycle): Likewise.
7422 (cortex_a9_neon_bp_3cycle): Likewise.
7423 (cortex_a9_neon_ldr): Likewise.
7424 (cortex_a9_neon_str): Likewise.
7425 (cortex_a9_neon_vld1_1_2_regs): Likewise.
7426 (cortex_a9_neon_vld1_3_4_regs): Likewise.
7427 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7428 (cortex_a9_neon_vld2_4_regs): Likewise.
7429 (cortex_a9_neon_vld3_vld4): Likewise.
7430 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7431 (cortex_a9_neon_vst1_3_4_regs): Likewise.
7432 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
7433 (cortex_a9_neon_vst3_vst4): Likewise.
7434 (cortex_a9_neon_vld1_vld2_lane): Likewise.
7435 (cortex_a9_neon_vld3_vld4_lane): Likewise.
7436 (cortex_a9_neon_vst1_vst2_lane): Likewise.
7437 (cortex_a9_neon_vst3_vst4_lane): Likewise.
7438 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
7439 (cortex_a9_neon_mcr): Likewise.
7440 (cortex_a9_neon_mcr_2_mcrr): Likewise.
7441 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
7442 (cortex_a9_fps): Likewise.
7443 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
7445 (cortex_m4_fmuls): Likewise.
7446 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
7448 (cortex_r4_mrc): Likewise.
7449 * config/arm/iterators.md: Update comment referring to neon_type.
7450 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
7451 (iwmmxt_movsi_insn): Likewise.
7452 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
7454 (pj4_core_to_vfp): Likewise.
7455 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
7457 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
7458 (vfp_fstore): Likewise.
7459 * doc/md.texi: Change references to neon_type to refer to type.
7461 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7463 * tree.h (DECL_BUILT_IN): Fix typo in comment.
7465 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
7467 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
7470 2013-09-04 Easwaran Raman <eraman@google.com>
7473 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
7474 (build_and_add_sum): Use it.
7475 (appears_later_in_bb): Simplify code.
7477 2013-09-04 Teresa Johnson <tejohnson@google.com>
7479 * dumpfile.c (dump_finish): Don't close stderr/stdout.
7481 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
7483 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
7485 2013-09-04 Jan Hubicka <jh@suse.cz>
7487 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
7488 * ipa-devirt.c: Include diganostic.h
7489 (odr_type_d): Add types and types_set.
7490 (hash_type_name): Work for types with vtables during LTO.
7491 (odr_hasher::remove): Fix comment; destroy types_set.
7492 (add_type_duplicate): New function,
7493 (get_odr_type): Use it.
7494 (dump_type_inheritance_graph): Dump type duplicates.
7495 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
7497 * tree.c (types_same_for_odr): Give exact answers on types with
7500 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7502 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
7503 explaining their differences.
7505 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
7507 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
7509 2013-09-03 Jeff Law <law@redhat.com>
7511 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
7512 when not threading through a joiner block. Pass joiner/no joiner
7513 state to register_jump_thread.
7514 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
7515 state from argument rather than implying on path length.
7516 Dump the entire jump thread path into debugging dump.
7517 * tree-flow.h (register_jump_thread): Update prototype.
7519 2013-08-29 Xinliang David Li <davidxl@google.com>
7521 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7522 Remove a trivial gcc_assert.
7524 2013-08-29 Xinliang David Li <davidxl@google.com>
7526 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
7527 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
7528 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7529 Delay base decl alignment adjustment.
7530 * tree-vectorizer.c (vect_destroy_datarefs): New function.
7531 * tree-vectorizer.h: New data structure.
7532 (set_dr_misalignment): New function.
7533 (dr_misalignment): Ditto.
7534 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
7535 (vectorizable_load): Ditto.
7536 (ensure_base_align): New function.
7537 (vectorize_loops): Add dbg_cnt support.
7538 (execute_vect_slp): Ditto.
7539 * dbgcnt.def: New debug counter.
7540 * Makefile: New dependency.
7542 2013-09-03 Meador Inge <meadori@codesourcery.com>
7546 2013-08-30 Meador Inge <meadori@codesourcery.com>
7548 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7550 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
7552 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
7553 function descriptor.
7555 2013-09-03 Richard Biener <rguenther@suse.de>
7557 * tree-affine.c (add_elt_to_tree): Fix association issue,
7558 avoid useless converts and make sure to always return a
7559 properly typed result.
7561 2013-09-03 Richard Biener <rguenther@suse.de>
7564 * fold-const.c (negate_expr_p): Fix division case.
7565 (negate_expr): Likewise.
7567 2013-09-03 Richard Biener <rguenther@suse.de>
7570 * tree-streamer-out.c: Include tm.h.
7571 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
7573 2013-09-03 Jan Hubicka <jh@suse.cz>
7575 * tree-profile.c (tree_profiling): Cleanup CFG when done.
7577 2013-09-03 Alan Modra <amodra@gmail.com>
7579 * config.gcc (powerpc*-*-linux*): Add support for little-endian
7580 multilibs to big-endian target and vice versa.
7581 * config/rs6000/t-linux64: Use := assignment on all vars.
7582 (MULTILIB_EXTRA_OPTS): Remove fPIC.
7583 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
7584 * config/rs6000/t-linux64le: New file.
7585 * config/rs6000/t-linux64bele: New file.
7586 * config/rs6000/t-linux64lebe: New file.
7588 2013-09-02 Jan Hubicka <jh@suse.cz>
7590 * ipa-inline-transform.c (inline_transform): Do not
7591 optimize_inline_calls when not optimizing.
7593 2013-09-02 Jan Hubicka <jh@suse.cz>
7595 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
7596 duplicated nodes for assembler names.
7597 * symtab.c (symtab_unregister_node): Do not attempt to unlink
7598 hard registers from assembler name hash.
7600 2013-09-02 Jan Hubicka <jh@suse.cz>
7602 * ipa-split.c (execute_split_functions): Split externally visible
7603 functions called once.
7605 2013-09-02 Martin Jambor <mjambor@suse.cz>
7608 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
7609 linked list. When finding the correct duplicate, also consider also
7610 the caller in additon to its inlined_to node.
7612 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
7614 * config/aarch64/aarch64-simd-builtins.def
7615 (dup_lane_scalar): Remove.
7616 * config/aarch64/aarch64-simd.md
7617 (aarch64_simd_dup): Add 'w->w' alternative.
7618 (aarch64_dup_lane<mode>): Allow for VALL.
7619 (aarch64_dup_lane_scalar<mode>): Remove.
7620 (aarch64_dup_lane_<vswap_width_name><mode>): New.
7621 (aarch64_get_lane_signed<mode>): Add w->w altenative.
7622 (aarch64_get_lane_unsigned<mode>): Likewise.
7623 (aarch64_get_lane<mode>): Likewise.
7624 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
7625 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
7626 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
7627 (VCON): Change container of V2SF.
7628 (vswap_width_name): Likewise.
7629 * config/aarch64/arm_neon.h
7630 (__aarch64_vdup_lane_any): New.
7631 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7632 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
7633 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7635 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
7638 * expr.c (emit_move_complex): Do not move complex FP values as parts if
7639 the source or the destination is a single hard register.
7641 2013-09-02 Richard Biener <rguenther@suse.de>
7644 * tree-scalar-evolution.c (instantiate_scev_name): Allow
7647 2013-09-02 Richard Biener <rguenther@suse.de>
7649 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
7650 arithmetic to sizetype.
7652 2013-09-02 Bin Cheng <bin.cheng@arm.com>
7654 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
7655 Find auto-increment use both before and after candidate.
7657 2013-09-02 Marek Polacek <polacek@redhat.com>
7659 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
7661 2013-09-01 Jan Hubicka <jh@suse.cz>
7663 * Makefile.in: Add ipa-profile.o
7664 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
7665 * cgraph.c (struct cgraph_propagate_frequency_data,
7666 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
7667 ipa-profile.c; replace cgraph_ by ipa_ prefix.
7668 * cgraph.h (cgraph_propagate_frequency): Remove.
7669 * ipa-inline-analysis.c: Include ipa-utils.h;
7670 drop duplicated cfgloop.h.
7671 (inline_update_callee_summaries): Update.
7672 * ipa-profile.c: New file.
7673 * ipa-utils.h (ipa_propagate_frequency): Declare.
7674 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
7675 data-streamer.h, value-prof.h.
7676 (symtab_remove_unreachable_nodes): Update profile.
7677 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
7678 account_time_size, cmp_counts, dump_histogram,
7679 ipa_profile_generate_summary, ipa_profile_write_summary,
7680 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
7681 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
7682 Move to ipa-profile.c.
7684 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
7686 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
7688 2013-09-01 Jan Hubicka <jh@suse.cz>
7690 * common.opt (fdevirtualize-speculatively): New function.
7691 * invoke.texi (fdevirtualize-speculatively): Document.
7692 * ipa-devirt.c: Include ipa-inline.h
7693 (likely_target_p): New function.
7694 (ipa_devirt): New function.
7695 (gate_ipa_devirt): New function.
7696 (pass_data_ipa_devirt): New static var.
7697 (pass_ipa_devirt): Likewise.
7698 (make_pass_ipa_devirt): New function.
7699 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
7700 (common_handle_option): Disable devirtualization when
7701 value range profiling is available.
7702 * passes.def (pass_ipa_devirt): Add.
7703 * timever.def (TV_IPA_DEVIRT): New timevar.
7704 * tree-pass.h (make_pass_ipa_devirt):
7706 2013-09-01 Iain Sandoe <iain@codesourcery.com>
7708 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
7709 include sanitize(undefined).
7711 2013-08-31 Diego Novillo <dnovillo@google.com>
7713 * Makefile.in (TREE_CORE_H): Define.
7715 (GTFILES): Add tree-core.h.
7716 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
7718 * tree-core.h: New file.
7719 Move all data structures, enum, typedefs, global
7720 declarations and constants from ...
7723 2013-08-31 Jan Hubicka <jh@suse.cz>
7725 * bulitins.c (expand_builtin): Do not early exit for gcov
7726 instrumented functions.
7728 2013-08-31 Marek Polacek <polacek@redhat.com>
7730 * ubsan.c: Include tm_p.h.
7732 2013-08-31 Jan Hubicka <jh@suse.cz>
7734 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
7737 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
7738 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
7739 * tree-cfg.c (verify_gimple_label): ... here.
7740 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
7741 (ipa_merge_profiles): New function.
7742 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
7743 (lto_input_function_body): Likewise.
7744 * ipa-utils.h (ipa_merge_profiles): Declare.
7745 * lto-streamer.h (lto_input_function_body): Update prototype.
7746 (emit_label_in_global_context_p): Remove.
7747 * lto-symtab.c: Include ipa-utils.h
7748 (lto_cgraph_replace_node): Use ipa_merge_profiles.
7750 2013-08-31 Jan Hubicka <jh@suse.cz>
7752 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
7754 2013-08-31 Jan Hubicka <jh@suse.cz>
7756 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
7758 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
7760 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
7761 "cmp" RTX before signed_comparison_operator check to account
7764 2013-08-30 Jan Hubicka <jh@suse.cz>
7766 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
7767 (detect_type_change_1): Rename to ...
7768 (detect_type_change): ... this one; early return on non-polymorphic
7770 (detect_type_change_ssa): Add comp_type parameter; update
7771 use of detect_type_change.
7772 (compute_complex_assign_jump_func): Add param_type parameter;
7773 update use of detect_type_change_ssa.
7774 (compute_complex_ancestor_jump_func): Likewise.
7775 (ipa_get_callee_param_type): New function.
7776 (ipa_compute_jump_functions_for_edge): Compute parameter type;
7777 update calls to the jump function computation functions.
7779 2013-08-30 Teresa Johnson <tejohnson@google.com>
7780 Steven Bosscher <steven@gcc.gnu.org>
7782 * cfgrtl.c (fixup_new_cold_bb): New routine.
7783 (commit_edge_insertions): Invoke fixup_partitions.
7784 (find_partition_fixes): New routine.
7785 (fixup_partitions): Ditto.
7786 (verify_hot_cold_block_grouping): Update comments.
7787 (rtl_verify_edges): Invoke find_partition_fixes.
7788 (rtl_verify_bb_pointers): Update comments.
7789 (rtl_verify_bb_layout): Ditto.
7790 * basic-block.h (probably_never_executed_edge_p): Declare.
7791 (fixup_partitions): Ditto.
7792 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
7793 * bb-reorder.c (sanitize_hot_paths): New function.
7794 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
7796 * predict.c (probably_never_executed_edge_p): New routine.
7797 * cfg.c (check_bb_profile): Add partition insanity warnings.
7799 2013-08-30 Meador Inge <meadori@codesourcery.com>
7801 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7803 2013-08-30 Marek Polacek <polacek@redhat.com>
7805 * Makefile.in (ubsan.o): Add.
7806 (c-family/c-ubsan.o): Add.
7807 (builtins.o): Add ubsan.h dependency.
7808 * ubsan.h: New file.
7809 * ubsan.c: New file.
7810 * common.opt: Add -fsanitize=undefined option.
7811 (flag_sanitize): Add variable.
7812 (fsanitize=): Add option. Add Driver.
7813 (fsanitize=thread): Remove option.
7814 (fsanitize=address): Likewise.
7815 (static-libubsan): New option.
7816 * doc/invoke.texi: Document the new flag and -static-libubsan.
7817 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
7818 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
7819 * builtin-attrs.def (ATTR_COLD): Define.
7820 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7821 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
7822 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
7823 * flag-types.h (sanitize_code): New enum.
7824 * opts.c (common_handle_option): Parse command line arguments
7825 of -fsanitize=. Add -fsanitize=unreachable option.
7826 * varasm.c (get_variable_section): Adjust.
7827 (assemble_noswitch_variable): Likewise.
7828 (assemble_variable): Likewise.
7829 (output_constant_def_contents): Likewise.
7830 (categorize_decl_for_section): Likewise.
7831 (place_block_symbol): Likewise.
7832 (output_object_block): Likewise.
7833 * builtins.def: Likewise.
7834 * toplev.c (compile_file): Likewise.
7835 (process_options): Likewise.
7836 * cppbuiltin.c: Likewise.
7837 * tsan.c (tsan_pass): Likewise.
7838 (tsan_gate): Likewise.
7839 (tsan_gate_O0): Likewise.
7840 * cfgexpand.c (partition_stack_vars): Likewise.
7841 (expand_stack_vars): Likewise.
7842 (defer_stack_allocation): Likewise.
7843 (expand_used_vars): Likewise.
7844 * cfgcleanup.c (old_insns_match_p): Likewise.
7845 * asan.c (asan_finish_file): Likewise.
7846 (asan_instrument): Likewise.
7847 (gate_asan): Likewise.
7848 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
7849 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7850 (asan_global_struct): Use pointer_sized_int_node instead
7851 calling build_nonstandard_integer_type.
7852 (initialize_sanitizer_builtins): Likewise.
7853 (asan_finish_file): Likewise.
7854 * gcc.c: Document %{%:function(args):X}.
7855 (static_spec_functions): Add sanitize.
7856 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
7857 store funcval != NULL there.
7858 (do_spec_1): Adjust handle_spec_function caller.
7859 (handle_braces): Allow %:function(args) as condition.
7860 (sanitize_spec_function): New function.
7861 (ADD_STATIC_LIBUBSAN_LIBS): Define.
7862 (LIBUBSAN_SPEC): Likewise.
7863 (LIBUBSAN_EARLY_SPEC): Likewise.
7864 (SANITIZER_SPEC): Handle libubsan.
7865 (SANITIZER_EARLY_SPEC): Likewise.
7866 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
7867 instead of fsanitize=address.
7868 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
7869 instead of fsanitize=address*.
7870 * builtins.c: Include ubsan.h.
7871 (fold_builtin_0): Instrument __builtin_unreachable.
7872 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
7873 instead of flag_asan.
7874 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
7875 (pointer_sized_int_node): Define.
7876 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
7878 2013-08-30 Mike Stump <mikestump@comcast.net>
7880 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
7883 2013-08-29 Jan Hubicka <jh@suse.cz>
7885 * cgraph.c (cgraph_function_body_availability): Handle weakref
7887 * passes.def: Remove pass_fixup_cfg.
7888 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
7889 track when we need to remove functions.
7890 (gate_ipa_inline): Execute inlining always; add comment why.
7891 (pass_data_ipa_inline): Remove TODO_remove_functions.
7892 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
7893 do not produce summaries.
7894 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
7895 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
7896 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
7897 constant pool and vtable.
7899 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
7901 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
7902 New arm_neon.h's internal macros to specify 64-bit constants.
7903 Avoid using stdint.h's macros.
7905 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
7907 * recog.c (verify_changes): Verify that changes[i].old is non-zero
7908 before applying REG_P.
7910 2013-08-30 Jakub Jelinek <jakub@redhat.com>
7912 PR tree-optimization/58277
7913 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
7914 after seeing too many stmts with vdef in between dombb and current
7915 bb, invalidate everything.
7917 2013-08-30 Richard Biener <rguenther@suse.de>
7919 * fold-const.c (fold_single_bit_test): Fix overflow test.
7921 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
7923 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
7924 and which can live in a register, always retrieve the value on entry.
7925 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
7926 passed by invisible reference specially.
7927 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
7928 (vt_add_function_parameter): Correctly deal with a parameter passed by
7929 invisible reference.
7931 2013-08-30 Jan Hubicka <jh@suse.cz>
7933 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
7935 2013-08-30 Richard Biener <rguenther@suse.de>
7937 PR tree-optimization/58228
7938 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
7939 allow invariant loads in nested loop vectorization.
7941 2013-08-30 Richard Biener <rguenther@suse.de>
7943 PR tree-optimization/58223
7944 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
7945 (has_anti_or_output_dependence): ... this and adjust to also
7946 look for output dependences.
7947 (mark_nodes_having_upstream_mem_writes): Adjust.
7948 (rdg_flag_uses): Likewise.
7950 2013-08-30 Richard Biener <rguenther@suse.de>
7952 PR tree-optimization/58010
7953 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
7954 assert that we have a loop-closed PHI.
7956 2013-08-29 Jan Hubicka <jh@suse.cz>
7958 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
7959 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
7960 * lto-section-in.c (lto_free_function_in_decl_state): New function.
7961 (lto_free_function_in_decl_state_for_node): New function.
7963 2013-08-29 Xinliang David Li <davidxl@google.com>
7965 * loop-unroll.c (report_unroll_peel): Minor message change.
7966 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
7967 Emit alignment peeling message with default -fopt-info.
7968 (vect_loop_versioning): Emit loop version info message.
7969 * tree-vectorizer.c (vectorize_loops): Minor message change.
7970 (execute_vect_slp): Ditto.
7972 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
7974 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
7975 of the clone from the DECL_NAME of the original function.
7977 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
7979 * passes.c (register_pass): Add overload.
7980 * tree-pass.h (register_pass): Forward declare it. Add comment.
7982 2013-08-29 Jan Hubicka <jh@suse.cz>
7984 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
7985 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
7986 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
7987 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7988 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
7989 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
7990 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7991 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
7993 2013-08-29 Teresa Johnson <tejohnson@google.com>
7995 * dumpfile.c (dump_loc): Output column number.
7996 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
7997 * doc/invoke.texi: Document optall -fopt-info flag.
7998 * profile.c (read_profile_edge_counts): Use new dump framework.
7999 (compute_branch_probabilities): Ditto.
8000 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
8001 when pass not in any opt group.
8002 * pass_manager.h (pass_manager::get_pass_profile): New method.
8003 * value-prof.c (check_counter): Use new dump framework.
8004 (check_ic_target): Ditto.
8005 * coverage.c (get_coverage_counts): Ditto.
8006 (coverage_init): Setup new dump framework.
8008 2013-08-29 Richard Biener <rguenther@suse.de>
8010 PR tree-optimization/58246
8011 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
8012 handle the dominance check inside a basic-block.
8014 2013-08-29 Richard Biener <rguenther@suse.de>
8017 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
8018 of default defs that appear in abnormal PHI nodes.
8020 2013-08-29 Richard Biener <rguenther@suse.de>
8022 PR tree-optimization/57685
8023 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
8024 single-use operands to avoid exponential complexity.
8026 2013-08-28 Dehao Chen <dehao@google.com>
8028 * ipa-inline.c (edge_badness): Fix integer underflow.
8030 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
8032 * gtm-builtins.def (_ITM_free): Declare leaf.
8034 2013-08-28 Jakub Jelinek <jakub@redhat.com>
8037 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
8038 (*tls_local_dynamic_base_64_largepic): Likewise.
8039 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
8040 Remove predicate from call operand.
8041 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
8042 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
8044 2013-08-28 Jeff Law <law@redhat.com>
8046 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
8047 checks for the number of predecessors and successors allowed.
8048 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
8049 which require copying a joiner block if there is a request which
8050 is a subpath that requires no joiner block copying.
8052 2013-08-28 Jan Hubicka <jh@suse.cz>
8054 * lto-streamer-out.c (DFS_write_tree_body): Drop
8055 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
8056 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
8057 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
8058 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
8059 Do not read DECL_ERROR_ISSUED.
8060 (unpack_ts_decl_with_vis_value_fields): Do not read
8062 (lto_input_ts_binfo_tree_pointers): Do not read
8063 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
8064 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
8065 write DECL_ERROR_ISSUED..
8066 (pack_ts_decl_with_vis_value_fields): Do not write
8068 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
8069 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
8070 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
8071 * tree.h (tree_decl_common): Update comment.
8072 (DECL_ERROR_ISSUED): Remove.
8074 2013-08-28 Jakub Jelinek <jakub@redhat.com>
8077 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
8079 2013-08-28 Jan Hubicka <jh@suse.cz>
8081 * builtins.def (free): Declare leaf.
8083 2013-08-27 David Malcolm <dmalcolm@redhat.com>
8086 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
8087 * configure: Regenerate.
8089 2013-08-27 Martin Jambor <mjambor@suse.cz>
8091 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
8092 (ipa_ancestor_jf_data): Likewise.
8093 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
8094 (ipa_get_jf_pass_through_type_preserved): New function.
8095 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
8096 (ipa_get_jf_ancestor_type_preserved): New function.
8097 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
8098 (ipa_get_jf_ancestor_result): Likewise.
8099 (propagate_vals_accross_pass_through): Use
8100 ipa_get_jf_pass_through_result to do all the value mappings.
8101 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
8102 type_preserved flag.
8103 (ipa_set_jf_cst_copy): New function.
8104 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
8105 (ipa_set_jf_arith_pass_through): Likewise.
8106 (ipa_set_ancestor_jf): Likewise.
8107 (compute_complex_assign_jump_func): Set type_preserved instead of
8109 (ipa_compute_jump_functions_for_edge): Likewise.
8110 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
8111 (update_jump_functions_after_inlining): Update type_preserved.
8112 Explicitely create jump functions when combining one with pass_through.
8113 (ipa_write_jump_function): Stream the type_preserved flags.
8114 (ipa_read_jump_function): Likewise.
8116 2013-08-27 Jakub Jelinek <jakub@redhat.com>
8117 Aldy Hernandez <aldyh@redhat.com>
8119 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
8120 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
8121 * function.h (struct function): Add has_force_vect_loops and
8123 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
8124 * gimple.c (gimple_build_omp_critical): Add KIND argument and
8126 * gimple.def: Update CLAUSES comments.
8127 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
8128 (gimple_build_omp_for): Add argument to prototype.
8129 (gimple_omp_for_kind): New.
8130 (gimple_omp_for_set_kind): New.
8131 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
8132 GOVD_DATA_SHARE_CLASS.
8133 (enum omp_region_type): Add ORT_SIMD.
8134 (gimple_add_tmp_var): Handle ORT_SIMD.
8135 (gimplify_var_or_parm_decl): Same.
8136 (is_gimple_stmt): Same.
8137 (omp_firstprivatize_variable): Same.
8138 (omp_add_variable): Only use splay_tree_insert if lookup failed.
8139 (omp_notice_variable): Handle ORT_SIMD.
8140 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
8141 (omp_check_private): Handle ORT_SIMD.
8142 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
8144 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
8145 Handle OMP_CLAUSE_LASTPRIVATE.
8146 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
8148 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
8149 (gimplify_expr): Handle OMP_SIMD.
8150 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
8151 (expand_GOMP_SIMD_VF): New.
8152 (expand_GOMP_SIMD_LAST_LANE): New.
8153 * internal-fn.def (GOMP_SIMD_LANE): New.
8154 (GOMP_SIMD_VF): New.
8155 (GOMP_SIMD_LAST_LANE): New.
8156 * omp-low.c: Include target.h.
8157 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
8159 (check_omp_nesting_restrictions): Same.
8161 (lower_rec_simd_input_clauses): New.
8162 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
8164 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
8165 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
8166 (expand_omp_build_assign): New.
8167 (expand_omp_for_init_counts): New.
8168 (expand_omp_for_init_vars): New.
8169 (extract_omp_for_update_vars): New.
8170 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
8171 and rewrite accordingly.
8172 (expand_omp_simd): New.
8173 (expand_omp_for): Use expand_omp_simd.
8174 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
8175 (lower_omp_for): Do not lower the body.
8176 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
8178 * tree-flow.h (find_omp_clause): Remove prototype.
8179 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
8180 forcing vectorization of the loop, or if flag_tree_vectorize.
8181 (gate_tree_if_conversion): Similarly.
8182 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
8183 gimple_build_omp_for.
8184 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
8185 * tree-parloops (create_parallel_loop): Pass kind argument to
8186 gimple_build_omp_for.
8187 * tree-pretty-print.c (dump_omp_clause): Add cases for
8188 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
8189 OMP_CLAUSE__SIMDUID_.
8190 (dump_generic_node): Handle OMP_SIMD.
8191 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
8192 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
8193 unroll OMP_SIMD loops here.
8194 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
8195 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
8197 (vect_analyze_data_refs): Handle simd loops.
8198 * tree-vect-loop.c (vectorizable_live_operation): Handle
8200 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
8201 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
8202 (vectorizable_load): Same.
8203 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
8204 (struct simduid_to_vf): New.
8205 (simduid_to_vf::hash): New.
8206 (simduid_to-vf::equal): New.
8207 (struct simd_array_to_simduid): New.
8208 (simd_array_to_simduid::hash): New.
8209 (simd_array_to_simduid::equal): New.
8210 (adjust_simduid_builtins): New.
8211 (struct note_simd_array_uses_struct): New.
8212 (note_simd_array_uses_cb): New.
8213 (note_simd_array_uses): New.
8214 (vectorize_loops): Handle simd hints and adjust simd builtins
8216 * tree-vectorizer.h (struct _stmt_vec_info): Add
8217 simd_lane_access_p field.
8218 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
8219 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
8220 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
8221 (omp_clause_code_name): Same.
8222 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
8223 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
8224 * tree.def (OMP_SIMD): New entry.
8225 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
8226 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
8227 (OMP_CLAUSE_DECL): Adjust range for new clauses.
8228 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
8229 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
8230 (OMP_CLAUSE_LINEAR_STEP): New.
8231 (OMP_CLAUSE_SAFELEN_EXPR): New.
8232 (OMP_CLAUSE__SIMDUID__DECL): New.
8233 (find_omp_clause): New prototype.
8235 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
8237 * config/i386/driver-i386.c (host_detect_local_cpu): Update
8238 Haswell processor detection.
8240 2013-08-27 Christian Widmer <shadow@umbrox.de>
8243 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
8244 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
8245 AVX2 capable processors.
8247 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
8249 * config/aarch64/arm_neon.h: Replace all inline asm implementations
8250 of vget_low_* with implementations in terms of other intrinsics.
8252 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
8255 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
8256 values to -1, 0 and 1.
8258 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
8260 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
8261 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
8262 (<optab><mode>3_insn): Remove.
8263 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
8264 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
8265 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
8266 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
8267 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
8268 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
8269 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
8270 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
8271 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
8272 (ror<mode>3_insn): Likewise.
8273 * config/aarch64/predicates.md (aarch64_simd_register): New.
8275 2013-08-27 Richard Biener <rguenther@suse.de>
8277 PR tree-optimization/57521
8278 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
8279 one edge is non-critical.
8280 (find_phi_replacement_condition): Make sure to use a non-critical
8281 edge. Cleanup and remove old bug workarounds.
8282 (bb_postdominates_preds): Remove.
8283 (if_convertible_loop_p_1): Do not compute post-dominators.
8284 (combine_blocks): Do not free post-dominators.
8285 (main_tree_if_conversion): Likewise.
8286 (pass_data_if_conversion): Add TODO_verify_ssa.
8288 2013-08-27 DJ Delorie <dj@redhat.com>
8290 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
8292 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
8294 * function.c (assign_parm_find_data_types): Set passed_mode and
8295 nominal_mode to the TYPE_MODE of nominal_type for the built
8296 pointer type in case of the struct-pass-by-reference.
8298 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8300 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
8301 (UINT16_TYPE): Change default to "unsigned int".
8303 * config/avr/avr.opt (mfract-convert-truncate): New option.
8304 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
8305 is set, round negative fractional integers according to n1169
8306 when converting to integer types.
8308 2013-08-26 Jan Hubicka <jh@suse.cz>
8310 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
8311 methods can not be called indirectly when their address is not taken.
8313 2013-08-26 Jan Hubicka <jh@suse.cz>
8315 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
8318 2013-08-26 Jan Hubicka <jh@suse.cz>
8320 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
8323 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8325 * reload.c (find_valid_class): Allow classes that do not include
8326 FIRST_PSEUDO_REGISTER - 1.
8328 2013-08-26 Jan Hubicka <jh@suse.cz>
8330 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
8331 fix edge count/frequency when speculation failed; fix type check
8332 for the direct call.
8334 2013-08-26 Jan Hubicka <jh@suse.cz>
8336 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
8338 2013-08-26 Jan Hubicka <jh@suse.cz>
8340 * ipa-inline-transform.c (inline_transform): Be ready for basic block
8341 to be changed by edge redirection.
8343 2013-08-26 Jan Hubicka <jh@suse.cz>
8345 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
8346 formating; add sanity check.
8347 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
8348 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
8349 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
8350 (resolve_noninline_speculation): Update callee keys, too.
8352 2013-08-26 Jan Hubicka <jh@suse.cz>
8354 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
8355 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
8357 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8359 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
8362 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
8364 * config/i386/i386.c (ix86_debug_options): Remove prototype.
8365 (x86_64_elf_select_section): Ditto.
8366 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
8368 (ix86_pass_by_reference): Ditto.
8369 (output_set_got): Ditto.
8370 (ix86_unary_operator_ok): Ditto.
8371 (ix86_expand_builtin): Ditto.
8373 2013-08-23 Jan Hubicka <jh@suse.cz>
8375 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
8377 2013-08-23 Jan Hubicka <jh@suse.cz>
8379 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
8380 (tree_decl_with_vis): Add FINAL field.
8382 2013-08-23 Jeff Law <law@redhat.com>
8384 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
8385 the debugging dump when the expression is fully redundant.
8387 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8389 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
8390 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
8391 * pretty-print.c (pp_formatted_text_data): Likewise.
8392 (pp_write_text_to_stream): Likewise.
8393 (pp_write_text_as_dot_label_to_stream): Likewise.
8394 (pp_append_r): Likewise.
8395 (pp_format): Likewise.
8396 (pp_flush): Likewise.
8397 (pp_clear_output_area): Likewise.
8398 (pp_append_text): Likewise.
8399 (pp_formatted_text): Likewise.
8400 (pp_remaining_character_count_for_line): Likewise.
8401 (pp_newline): Likewise.
8402 (pp_character): Likewise.
8403 (output_buffer::~output_buffer): Define.
8404 (pretty_printer::~pretty_printer): Destruct output buffer.
8405 * pretty-print.h (output_buffer::~output_buffer): Declare.
8406 (pretty_printer::~pretty_printer): Declare virtual.
8408 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
8411 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
8412 HOST_WIDE_INT_M1U): New macros.
8413 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
8414 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
8415 unsigned -1 for lshift.
8416 * cse.c (cse_insn): Likewise.
8417 * double-int.c (rshift_double, lshift_double): Likewise.
8418 * builtins.c (fold_builtin_bitop): Likewise.
8419 * combine.c (force_to_mode): Likewise.
8420 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
8421 * simplify-rtx.c (simplify_const_unary_operation,
8422 simplify_const_binary_operation): Likewise.
8423 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
8424 check_va_list_escapes): Likewise.
8425 * rtlanal.c (nonzero_bits1): Likewise.
8426 * expmed.c (expand_smod_pow2): Likewise.
8427 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
8429 2013-08-23 Jan Hubicka <jh@suse.cz>
8431 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
8432 as having address taken.
8434 2013-08-23 Jan Hubicka <jh@suse.cz>
8436 * ipa-utils.h (method_class_type): Declare.
8437 * ipa-devirt.c (method_class_type): Export.
8439 * cgraphunit.c (analyze_functions): Do basic devirtualization;
8440 do not walk base classes of anonymous types.
8442 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
8444 PR rtl-optimization/58220
8446 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
8447 handle SEQUENCE insns properly.
8449 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8451 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
8453 (pp_newline_and_indent): Likewise.
8454 (pp_separate_with): Likewise.
8455 * pretty-print.c (pp_newline_and_flush): Define.
8456 (pp_newline_and_indent): Likewise.
8457 (pp_separate_with): Likewise.
8459 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8462 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
8463 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
8465 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
8467 * config/i386/predicates.md (ext_sse_reg_operand): New.
8468 * config/i386/i386.md (*movti_internal): Use
8469 predicate to determine if EVEX is needed.
8470 (*movsi_internal): Ditto.
8471 (*movdf_internal): Ditto.
8472 (*movsf_internal): Ditto.
8473 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
8475 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8477 PR tree-optimization/58209
8478 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
8479 (find_tail_calls): Give up for pointer result types if m is non-NULL.
8480 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
8481 emit POINTER_PLUS_EXPR.
8482 (create_tailcall_accumulator): For pointer result type accumulate in
8485 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
8487 * configure.ac: Add backslashes missing from the last change.
8488 * configure: Regenerate.
8490 2013-08-22 Jan Hubicka <jh@susue.cz>
8492 * ipa.c (function_and_variable_visibility): First remember function
8493 was global and then make it local.
8495 2013-08-22 Julian Brown <julian@codesourcery.com>
8497 * configure.ac: Add aarch64 to list of arches which use "nop" in
8499 * configure: Regenerate.
8501 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8503 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
8504 gnu_libc_has_function.
8505 * config/s390/tpf.h: Likewise.
8507 2013-08-22 Jan Hubicka <jh@susue.cz>
8509 * timevar.c (validate_phases): Add cast.
8511 2013-08-22 Jan Hubicka <jh@susue.cz>
8513 * timevar.c (validate_phases): Use size_t for memory.
8514 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
8516 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
8518 * pretty-print.h (output_buffer::output_buffer): Declare.
8519 (pretty_printer::pretty_printer): Likewise.
8520 (pp_construct): Remove.
8521 * pretty-print.c (output_buffer::output_buffer): Define.
8522 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
8523 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
8524 (print_gimple_expr): Likewise.
8525 (print_gimple_seq): Likewise.
8526 (gimple_dump_bb): Likewise.
8527 * sched-vis.c (dump_value_slim): Likewise.
8528 (dump_insn_slim): Likewise.
8529 (dump_rtl_slim): Likewise.
8530 (str_pattern_slim): Likewise.
8531 * tree-mudflap.c (mf_varname_tree): Likewise.
8532 * graph.c (print_graph_cfg): Likewise.
8533 (start_graph_dump): Likewise.
8534 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
8536 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
8537 pretty printer initialization.
8538 * coretypes.h (diagnostic_context): Remove superflous type alias
8540 (pretty_printer): Likewise. Declare directly as a class.
8541 (pretty_print_info): Remove declaration as class.
8542 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
8543 and pp_clear_output_area.
8544 (asan_add_global): Likewise.
8546 2013-08-22 Jan Hubicka <jh@suse.cz>
8548 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
8549 * ipa-utils.h (update_type_inheritance_graph): Declare.
8550 (possible_polymorphic_call_target_p): Declare.
8551 (possible_polymorphic_call_target_p): New.
8552 * ipa-devirt.c: Update toplevel comments.
8553 (cached_polymorphic_call_targets): Move up.
8554 (odr_type_d): Move ID down.
8555 (polymorphic_type_binfo_p): Update comment.
8556 (odr_hasher::remove): Likewise;
8557 (get_odr_type): Set anonymous_namespace.
8558 (dump_odr_type): Dump it.
8559 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
8560 (maybe_record_node): Record node in cached_polymorphic_call_targets.
8561 (record_binfo): Add comment.
8562 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
8564 (devirt_node_removal_hook): Do not iCE when cache is freed.
8565 (possible_polymorphic_call_target_p): New predicate.
8566 (update_type_inheritance_graph): New function.
8568 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
8569 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8570 Sergey Lega <sergey.s.lega@intel.com>
8571 Anna Tikhonova <anna.tikhonova@intel.com>
8572 Ilya Tocar <ilya.tocar@intel.com>
8573 Andrey Turetskiy <andrey.turetskiy@intel.com>
8574 Ilya Verbin <ilya.verbin@intel.com>
8575 Kirill Yukhin <kirill.yukhin@intel.com>
8576 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8578 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
8579 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
8580 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
8581 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
8582 (OPTION_MASK_ISA_AVX2_UNSET): Update.
8583 (OPTION_MASK_ISA_AVX512F_UNSET): New.
8584 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
8585 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
8586 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
8587 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
8588 OPT_mavx512pf, OPT_mavx512er cases.
8589 * config/i386/constraints.md (v): New constraint.
8590 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
8591 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
8592 (bit_AVX512CD): New.
8593 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8594 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
8595 * config/i386/i386-c.c (ix86_target_macros_internal):
8596 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
8598 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
8599 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
8600 * config/i386/i386.c (regclass_map, dbx_register_map)
8601 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
8602 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
8603 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
8605 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
8606 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
8607 -mavx512pf options. Fix formatting.
8608 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
8609 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
8610 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
8611 -mavx512cd, -mavx512pf options.
8612 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
8613 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
8614 (ix86_preferred_output_reload_class): Replace SSE_REGS with
8616 (ix86_hard_regno_mode_ok): Support 512-bit registers.
8617 (ix86_set_reg_reg_cost): Ditto.
8618 (x86_order_regs_for_local_alloc): Ditto.
8619 (MAX_VECT_LEN): Extend to 64-byte.
8620 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
8621 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
8622 (TARGET_AVX512ER, TARGET_AVX512CD): New.
8623 (BIGGEST_ALIGNMENT): Extend to 512-bits.
8624 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
8625 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
8626 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
8627 (SSE_REG_MODE_P): Support new modes.
8628 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
8629 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
8630 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
8631 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
8632 (REG_CLASS_CONTENTS): Add new registers.
8633 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
8634 (EXT_REX_SSE_REGNO_P): New.
8635 (HI_REGISTER_NAMES): Add new registers.
8636 * config/i386/i386.md: Define constants for new registers.
8637 (mode): Add new 512-bit modes.
8638 (prefix): Support evex prefix.
8639 (isa): Support avx512f, noavx512f, fma_avx512f.
8640 (ssemodesuffix): Add new 512-bit modes.
8642 (*movxi_internal_avx512f): Ditto.
8643 (*movdi_internal): Replace constraint "x" with the new constraint "v".
8645 (*movsi_internal): Likewise.
8646 (*movdf_internal): Likewise.
8647 (*movsf_internal): Likewise.
8648 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
8649 (<code><mode>3): Likewise.
8650 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
8652 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
8653 with the new constraint "v".
8654 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
8656 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
8657 with the new constraint "v".
8658 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
8659 (<sse2>_storedqu<avxsizesuffix>): Likewise.
8660 (*<plusminus_insn><mode>3): Likewise.
8661 (<sse>_vm<plusminus_insn><mode>3): Likewise.
8662 (*mul<mode>3): Likewise.
8663 (<sse>_vmmul<mode>3): Likewise.
8664 (<sse>_div<mode>3): Likewise.
8665 (<sse>_vmdiv<mode>3): Likewise.
8666 (<sse>_sqrt<mode>2): Likewise.
8667 (<sse>_vmsqrt<mode>2): Likewise.
8668 (*<code><mode>3_finite): Likewise.
8669 (*<code><mode>3) <smaxmin>: Likewise.
8670 (<sse>_vm<code><mode>3): Likewise.
8671 (*<code><mode>3) <any_logic>: Likewise.
8672 (*fma_fmadd_<mode>): Likewise.
8673 (*fma_fmsub_<mode>): Likewise.
8674 (*fma_fnmadd_<mode>): Likewise.
8675 (*fma_fnmsub_<mode>): Likewise.
8676 (*fma_fmaddsub_<mode>): Likewise.
8677 (*fma_fmsubadd_<mode>): Likewise.
8678 (*fmai_fmadd_<mode>): Likewise.
8679 (*fmai_fmsub_<mode>): Likewise.
8680 (*fmai_fnmadd_<mode>): Likewise.
8681 (*fmai_fnmsub_<mode>): Likewise.
8682 (sse_cvtsi2ss): Likewise.
8683 (sse_cvtsi2ssq): Likewise.
8684 (sse_cvtss2si): Likewise.
8685 (sse_cvtss2si_2): Likewise.
8686 (sse_cvtss2siq): Likewise.
8687 (sse_cvtss2siq_2): Likewise.
8688 (sse_cvttss2si): Likewise.
8689 (sse_cvtss2siq_2): Likewise.
8690 (float<sseintvecmodelower><mode>2): Likewise.
8691 (sse2_cvtsd2si_2): Likewise.
8692 (sse2_cvtsd2siq_2): Likewise.
8693 (*<plusminus_insn><mode>3): Likewise.
8694 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
8695 (*<sse4_1_avx2>_mul<mode>3): Likewise.
8696 (ashr<mode>3): Likewise.
8697 (<shift_insn><mode>3): Likewise.
8698 (avx2_<code><mode>3): Likewise.
8699 (*avx2_<code><mode>3): Likewise.
8700 (*andnot<mode>3): Likewise.
8701 (*<code><mode>3) <any_logic>: Likewise.
8702 (abs<mode>2): Likewise.
8703 (avx2_permvar<mode>): Likewise.
8704 (avx2_perm<mode>_1): Likewise.
8705 (*avx_vpermilp<mode>): Likewise.
8706 (avx_vpermilvar<mode>3): Likewise.
8707 (avx2_ashrv<mode>): Likewise.
8708 (avx2_<shift_insn>v<mode>): Likewise.
8709 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
8711 * doc/rtl.texi: Document XImode.
8713 2013-08-21 Jeff Law <law@redhat.com>
8715 * tree-flow.h (register_jump_thread): Pass vector of edges
8716 instead of each important edge.
8717 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
8718 thread path into a vector and pass that to register_jump_thread.
8719 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
8720 passed in edge vector to the current 3-edge form.
8723 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8725 * dce.c (fini_dce): Call df_analyze again just in case
8726 delete_unmarked_insns removed anything.
8728 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
8730 * reload.h (struct reg_equivs): Rename to ..
8731 (struct reg_equivs_s): .. this.
8733 2013-08-20 Martin Liska <marxin.liska@gmail.com>
8735 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
8737 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8739 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
8741 2013-08-21 Jeff Law <law@redhat.com>
8743 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
8744 simplify assignments too. If the RHS collapses to a singleton
8745 range, then return the value for the range.
8747 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
8749 * config/i386/sse.md (V16): Rename to...
8751 (mov<mode>): Update iterator name.
8752 (*mov<mode>_internal): Ditto.
8753 (push<mode>1): Ditto.
8754 (movmisalign<mode>): Ditto.
8756 2013-08-20 Jan Hubicka <jh@suse.cz>
8759 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
8760 entry for direct edges.
8761 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
8763 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8765 Revert my last two changes, r201865 and r201864:
8768 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8770 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8771 instances can own GC refs.
8773 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8774 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8775 (gcc::context::gt_pch_nx): Likewise.
8776 (gcc::context::gt_pch_nx): Likewise.
8777 * ggc.h (gt_ggc_mx <T>): New.
8778 (gt_pch_nx_with_op <T>): New.
8779 (gt_pch_nx <T>): New.
8780 * passes.c (opt_pass::gt_ggc_mx): New.
8781 (opt_pass::gt_pch_nx): New.
8782 (opt_pass::gt_pch_nx_with_op): New.
8783 (pass_manager::gt_ggc_mx): New.
8784 (pass_manager::gt_pch_nx): New.
8785 (pass_manager::gt_pch_nx_with_op): New.
8786 (pass_manager::operator new): Use
8787 ggc_internal_cleared_alloc_stat rather than xcalloc.
8788 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8789 (pass_manager::gt_ggc_mx): New.
8790 (pass_manager::gt_pch_nx): New.
8791 (pass_manager::gt_pch_nx_with_op): New.
8792 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8793 (opt_pass::operator new): New.
8794 (opt_pass::gt_ggc_mx): New.
8795 (opt_pass::gt_pch_nx): New.
8796 (opt_pass::gt_pch_nx_with_op): New.
8799 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8801 * Makefile.in (GTFILES): Add context.h.
8802 * context.c (gcc::context::operator new): New.
8803 (gcc::context::gt_ggc_mx): New.
8804 (gcc::context::gt_pch_nx): New.
8805 (gcc::context::gt_pch_nx): New.
8806 * context.h (gcc::context): Add GTY((user)) marking.
8807 (gcc::context::operator new): New.
8808 (gcc::context::gt_ggc_mx): New.
8809 (gcc::context::gt_pch_nx): New.
8810 (gcc::context::gt_pch_nx): New.
8811 (g): Add GTY marking.
8812 (gt_ggc_mx (gcc::context *)): New.
8813 (gt_pch_nx (gcc::context *)): New.
8814 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8815 void *cookie)): New.
8816 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8818 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8820 * dce.c (fini_dce): Call df_analyze again just in case
8821 delete_unmarked_insns removed anything.
8823 2013-08-20 Teresa Johnson <tejohnson@google.com>
8825 PR rtl-optimizations/57451
8826 * final.c (reemit_insn_block_notes): Prevent lexical blocks
8827 from crossing split section boundaries.
8829 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8831 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
8832 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
8833 with MULTLIB_DEFAULTS.
8835 2013-08-20 Nick Clifton <nickc@redhat.com>
8837 * target.def (narrow_volatile_bitfield): Note that the default
8838 value is false, not !TARGET_STRICT_ALIGN.
8839 * doc/tm.texi: Regenerate.
8841 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
8843 Fix LIB_SPEC for systems without libpthread.
8845 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
8846 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
8848 * config/i386/linux-common.h: Likewise.
8849 * config/mips/linux-common.h: Likewise.
8851 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
8853 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
8856 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8858 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8859 instances can own GC refs.
8861 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8862 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8863 (gcc::context::gt_pch_nx): Likewise.
8864 (gcc::context::gt_pch_nx): Likewise.
8865 * ggc.h (gt_ggc_mx <T>): New.
8866 (gt_pch_nx_with_op <T>): New.
8867 (gt_pch_nx <T>): New.
8868 * passes.c (opt_pass::gt_ggc_mx): New.
8869 (opt_pass::gt_pch_nx): New.
8870 (opt_pass::gt_pch_nx_with_op): New.
8871 (pass_manager::gt_ggc_mx): New.
8872 (pass_manager::gt_pch_nx): New.
8873 (pass_manager::gt_pch_nx_with_op): New.
8874 (pass_manager::operator new): Use
8875 ggc_internal_cleared_alloc_stat rather than xcalloc.
8876 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8877 (pass_manager::gt_ggc_mx): New.
8878 (pass_manager::gt_pch_nx): New.
8879 (pass_manager::gt_pch_nx_with_op): New.
8880 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8881 (opt_pass::operator new): New.
8882 (opt_pass::gt_ggc_mx): New.
8883 (opt_pass::gt_pch_nx): New.
8884 (opt_pass::gt_pch_nx_with_op): New.
8886 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8888 * Makefile.in (GTFILES): Add context.h.
8889 * context.c (gcc::context::operator new): New.
8890 (gcc::context::gt_ggc_mx): New.
8891 (gcc::context::gt_pch_nx): New.
8892 (gcc::context::gt_pch_nx): New.
8893 * context.h (gcc::context): Add GTY((user)) marking.
8894 (gcc::context::operator new): New.
8895 (gcc::context::gt_ggc_mx): New.
8896 (gcc::context::gt_pch_nx): New.
8897 (gcc::context::gt_pch_nx): New.
8898 (g): Add GTY marking.
8899 (gt_ggc_mx (gcc::context *)): New.
8900 (gt_pch_nx (gcc::context *)): New.
8901 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8902 void *cookie)): New.
8903 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8905 2013-08-20 Alan Modra <amodra@gmail.com>
8908 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
8909 (rs6000_emit_epilogue): Likewise.
8911 2013-08-19 Dehao Chen <dehao@google.com>
8913 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
8915 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
8916 Jakub Jelinek <jakub@redhat.com>
8918 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
8919 (BUILT_IN_FABSD64): Likewise.
8920 (BUILT_IN_FABSD128): Likewise.
8921 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
8922 (fold_builtin_1): Likewise.
8923 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
8924 destination and source operands.
8925 (*abstd2_fpr): Likewise.
8926 (*nabstd2_fpr): Likewise.
8928 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
8930 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
8933 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
8935 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
8938 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
8940 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
8941 * builtins.c (default_libc_has_function): New.
8942 (gnu_libc_has_function): Ditto.
8943 (no_c99_libc_has_function): Ditto.
8944 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
8945 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
8946 (fold_builtin_sincos): Likewise.
8947 (fold_builtin_cexp): Likewise.
8948 * builtins.def (DEF_C94_BUILTIN): Likewise.
8949 (DEF_C99_BUILTIN): Likewise.
8950 (DEF_C99_C90RES_BUILTIN): Likewise.
8951 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
8952 definitions to using this define.
8953 * config/darwin-protos.h (darwin_libc_has_function): New.
8954 * config/darwin.c (darwin_libc_has_function): Ditto.
8955 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
8956 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
8957 * config/darwin.h: Ditto.
8958 * config/elfos.h: Ditto.
8959 * config/freebsd.h: Ditto.
8960 * config/i386/cygming.h: Ditto.
8961 * config/i386/djgpp.h: Ditto.
8962 * config/i386/i386-interix.h: Ditto.
8963 * config/microblaze/microblaze.h: Ditto.
8964 * config/mmix/mmix.h: Ditto.
8965 * config/gnu-user.h: Ditto.
8966 * config/ia64/hpux.h: Ditto.
8967 * config/pa/pa-hpux.h: Ditto.
8968 * config/pdp11/pdp11.h: Ditto.
8969 * config/picochip/picochip.h: Ditto.
8970 * config/linux.h: Ditto.
8971 * config/netbsd.h: Ditto.
8972 * config/openbsd.h: Ditto.
8973 * config/rs6000/aix43.h: Ditto.
8974 * config/rs6000/aix51.h: Ditto.
8975 * config/rs6000/aix52.h: Ditto.
8976 * config/rs6000/aix53.h: Ditto.
8977 * config/rs6000/aix61.h: Ditto.
8978 * config/rs6000/darwin.h: Ditto.
8979 * config/rs6000/linux.h: Ditto.
8980 * config/rs6000/linux64.h: Ditto.
8981 * config/s390/tpf.h: Ditto.
8982 * config/sol2-10.h: Ditto.
8983 * config/sol2.h: Ditto.
8984 * config/vms/vms.h: Ditto.
8985 * config/vxworks.h: Ditto.
8986 * config/linux-android.c (linux_android_libc_has_function):
8987 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
8988 * config/linux-protos.h (linux_android_libc_has_function):
8990 * config/i386/i386.c (ix86_libc_has_function): New.
8991 * config/i386/i386-protos.h
8992 (ix86_libc_has_function): New declaration.
8993 * config/i386/i386.md
8994 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
8995 ("isinf<mode>2): Likewise.
8996 * convert.c (convert_to_integer): Using new target hook
8997 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
8998 TARGET_C99_FUNCTIONS.
8999 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
9000 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
9001 * coretypes.h (function_class): New enum for different
9002 classes of functions.
9003 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
9004 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
9005 (TARGET_HAS_SINCOS): Likewise.
9006 (TARGET_LIBC_HAS_FUNCTION): New.
9007 * doc/tm.texi: Regenerated.
9008 * targhooks.h (default_libc_has_function): New declaration.
9009 (no_c99_libc_has_function): Ditto.
9010 (gnu_libc_has_function): Ditto.
9011 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
9012 and TARGET_HAS_SINCOS.
9014 2013-08-18 Jan Hubicka <jh@suse.cz>
9016 * Makeifle-in (ipa-devirt.o): New.
9017 (GTFILES): Add ipa-utils.h and ipa-devirt.c
9018 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
9019 (analyze_functions): Look into possible targets of polymorphic call.
9020 * dumpfile.c (dump_files): Add type-inheritance dump.
9021 * dumpfile.h (TDI_inheritance): New.
9022 * ipa-devirt.c: New file.
9023 * ipa-utils.h (odr_type_d): Forward declare.
9024 (odr_type): New type.
9025 (build_type_inheritance_graph): Declare.
9026 (possible_polymorphic_call_targets): Declare and introduce inline
9027 variant when only edge is pased.
9028 (dump_possible_polymorphic_call_targets): Likewise.
9029 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
9030 * tree.c (type_in_anonymous_namespace_p): Break out from ...
9031 (types_same_for_odr): ... here.
9032 * tree.h (type_in_anonymous_namespace_p): Declare.
9034 2013-08-18 Jakub Jelinek <jakub@redhat.com>
9036 PR tree-optimization/58006
9037 * tree-parloops.c (take_address_of): Don't ICE if get_name
9039 (eliminate_local_variables_stmt): Remove clobber stmts.
9041 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
9043 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
9044 error message is issued for an alias to undefined symbol.
9046 2013-08-18 Jan Hubicka <jh@suse.cz>
9048 * cgraph.c (cgraph_create_indirect_edge): Discover
9049 polymorphic calls and record basic info into indirect_info.
9050 * gimple-fold.c (gimple_fold_call): When doing BINFO based
9051 devirtualization, ignore objc function calls.
9052 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
9053 call with no parm index info.
9054 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
9055 * tree.c (virtual_method_call_p): New function.
9056 * tree.h (virtual_method_call_p): Declare.
9058 2013-08-16 Jan Hubicka <jh@suse.cz>
9061 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
9063 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
9065 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
9068 2013-08-16 David Malcolm <dmalcolm@redhat.com>
9070 * gengtype.c (type_for_name): Add special-case support for
9071 locating types within the "gcc::" namespace.
9072 (open_base_files): Emit a "using namespace gcc" directive.
9074 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9077 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
9078 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
9080 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
9081 array instead of each individual operand as a separate argument.
9082 (emit_fusion_gpr_load): Likewise.
9083 (expand_fusion_gpr_load): Add new function declaration.
9085 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
9086 signature to have the operands passed as an array, instead of as
9087 separate arguments. Allow ZERO_EXTEND to be in the memory
9088 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
9089 depend on the register live/dead flags when peepholes are run.
9090 (expand_fusion_gpr_load): New function to be called from the
9091 peephole2 pass, to change the register that addis sets to be the
9093 (emit_fusion_gpr_load): Change the calling signature to have the
9094 operands passed as an array, instead of as separate arguments.
9095 Allow ZERO_EXTEND to be in the memory address, and also
9096 SIGN_EXTEND if -mpower8-fusion-sign.
9098 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
9100 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
9101 adjust the register addis loads up in the peephole2 pass. Do not
9102 depend on the register live/dead state when the peephole pass is done.
9104 2013-08-16 David Malcolm <dmalcolm@redhat.com>
9106 * gengtype.c (create_user_defined_type): Ensure that the kind
9107 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
9108 declaration is seen before the GTY((user)) marking.
9110 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
9113 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
9115 2013-08-16 Jan Hubicka <jh@suse.cz>
9117 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
9118 arugment expected_type.
9119 (gimple_fold_call): Use it.
9120 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
9121 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
9122 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
9123 (try_make_edge_direct_virtual_call): Likewise.
9124 * tree.c (obj_type_ref_class): New.
9125 * tree.h (obj_type_ref_class): Use it.
9127 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
9129 * sched-vis.c (rtl_slim_pp_initialized): Remove.
9130 (rtl_slim_pp): Likewise.
9131 (init_rtl_slim_pretty_print): Likewise.
9132 (dump_value_slim): Don't call it. Use local pretty printer.
9133 (dump_insn_slim): Likewise.
9134 (dump_rtl_slim): Likewise.
9135 (str_pattern_slim): Likewise.
9136 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
9139 2013-08-16 Jakub Jelinek <jakub@redhat.com>
9141 PR tree-optimization/58164
9142 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
9143 walk gimple_goto_dest of GIMPLE_GOTO.
9145 PR tree-optimization/58165
9146 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
9147 bi_call must be the last stmt in a bb, don't split_block, instead
9148 use fallthru edge from it and give up if there is none.
9149 Release conds vector when returning early.
9151 2013-08-14 Xinliang David Li <davidxl@google.com>
9153 * config/i386/i386.c (ix86_option_override_internal):
9154 Remove unused variable and field.
9156 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9159 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
9160 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
9161 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
9162 and Linux, correct BLKmode alignment when 128-bit alignment is
9163 required and compatibility flag is not set.
9164 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
9165 for zero-size arguments when compatibility flag is not set.
9167 2013-08-14 Jakub Jelinek <jakub@redhat.com>
9169 PR tree-optimization/58145
9170 * tree-sra.c (build_ref_for_offset): If prev_base has
9171 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
9173 2013-08-14 Xinliang David Li <davidxl@google.com>
9175 * config/i386/i386.c (ix86_option_override_internal):
9176 Fix uninitialized variable error.
9178 2013-08-14 Xinliang David Li <davidxl@google.com>
9180 * config/i386/i386.opt: Define two new options.
9181 * config/i386/x86-tune.def: Add arch selector field in macros.
9182 * config/i386/i386.h: Adjust macro definition.
9183 * config/i386/i386.c (ix86_option_override_internal):
9185 (parse_mtune_ctrl_str): New function.
9186 (set_ix86_tune_features): New function.
9187 (ix86_function_specific_restore): Call the new helper function.
9189 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
9191 PR rtl-optimization/57662
9192 * sel-sched.c (code_motion_process_successors): When the current insn
9193 is removed after the recursive traversal, break from the loop.
9194 Add comments and debug printouts.
9196 2013-08-14 Jakub Jelinek <jakub@redhat.com>
9197 Alexandre Oliva <aoliva@redhat.com>
9200 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
9201 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
9202 there also UNSPEC_PLTOFF.
9204 2013-08-14 Marek Polacek <polacek@redhat.com>
9206 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
9207 NUM_CONDITIONS bit positions.
9209 2013-08-13 Cary Coutant <ccoutant@google.com>
9211 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
9212 (attr_checksum): Hash vector contents instead of pointer.
9213 (attr_checksum_ordered): Likewise.
9215 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
9217 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
9218 when Pmode != word_mode. Add length_address attribute.
9219 (sse3_monitor_<mode>): Merge from sse3_monitor and
9220 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
9221 Pmode != word_mode. Update insn length attribute.
9222 * config/i386/i386.c (ix86_option_override_internal): Update
9223 ix86_gen_monitor selection for merged sse3_monitor insn.
9225 2013-08-13 Julian Brown <julian@codesourcery.com>
9227 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
9228 perform invalid legitimization on greater-than-word-size modes for
9231 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
9233 * ira.c (setup_class_translate_array): Use aclass instead of cl
9234 for classes not fully covered by allocno classes.
9236 2013-08-13 Jakub Jelinek <jakub@redhat.com>
9238 PR tree-optimization/57661
9239 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
9240 * tree-inline.c (tree_function_versioning): Initialize it.
9241 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
9242 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
9243 that is not being copied.
9246 * asan.c (instrument_strlen_call): Fix typo in comment.
9247 Use char * type even for the lhs of POINTER_PLUS_EXPR.
9249 2013-08-13 Steve Ellcey <sellcey@mips.com>
9251 * config/mips/mips.md (prefetch): Use lw instead of ld on
9252 loongson in 32bit mode.
9254 2013-08-13 Nick Clifton <nickc@redhat.com>
9256 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
9258 2013-08-13 Jan Hubicka <jh@suse.cz>
9260 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
9261 introduced edge; fix typo in sanity check.
9262 (cgraph_resolve_speculation): Export; improve diagnostic.
9263 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
9264 speculation at type mismatch.
9265 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
9266 (cgraph_resolve_speculation): Declare.
9267 (symtab_can_be_discarded): New function.
9268 * value-prof.c (gimple_ic_transform): Remove actual transform code.
9269 * ipa-inline-transform.c (speculation_removed): New global var.
9270 (clone_inlined_nodes): See if speculation can be removed.
9271 (inline_call): If speculations was removed, we growths may not match.
9272 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
9273 (speculation_useful_p): New function.
9274 (resolve_noninline_speculation): New function.
9275 (inline_small_functions): Resolve useless speculations.
9276 * ipa-inline.h (speculation_useful_p): Declare
9277 * ipa.c (can_replace_by_local_alias): Simplify.
9278 (ipa_profile): Produce speculative calls in non-lto, too;
9279 add simple cost model; produce local aliases.
9281 2013-08-13 David Malcolm <dmalcolm@redhat.com>
9283 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
9286 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
9288 * config/i386/i386.c (ix86_function_versions): Use error + inform.
9290 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
9292 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
9293 iterator instead of X87MODEF.
9295 2013-08-12 Perez Read <netfirewall@gmail.com>
9298 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
9299 operand 0 for intel asm alternative.
9300 (*movabs<mode>_2): Ditto for operand 1.
9302 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
9304 * config/aarch64/arm_none.h
9305 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
9307 2013-08-12 Nick Clifton <nickc@redhat.com>
9309 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
9311 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
9313 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
9314 expand for QI/HImode operand to produce more effictive code for
9315 unsigned char(short) --> float(double) conversion.
9317 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
9319 * doc/invoke.texi: Mention that -ftls-model does not force the final
9322 2013-08-12 Marek Polacek <polacek@redhat.com>
9323 Marc Glisse <marc.glisse@inria.fr>
9325 PR tree-optimization/57980
9326 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
9327 when creating -1 constant.
9329 2013-08-10 Jan Hubicka <jh@suse.cz>
9331 Workaround binutils PR14342.
9332 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
9333 (gimple_init_edge_profiler): Likewise.
9334 (gimple_gen_ic_func_profiler): Likewise.
9336 2013-08-09 Jan Hubicka <jh@suse.cz>
9338 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
9340 2013-08-09 Xinliang David Li <davidxl@google.com>
9342 * config/i386/stringop.def: New file.
9343 * config/i386/stringop.opt: New file.
9344 * config/i386/i386-opts.h: Include stringopt.def.
9345 * config/i386/i386.opt: Include stringopt.opt.
9346 * config/i386/i386.c (ix86_option_override_internal):
9347 Override default size based stringop inline strategies with options.
9348 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
9351 2013-08-09 Jan Hubicka <jh@suse.cz>
9353 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
9355 2013-08-09 Jan Hubicka <jh@suse.cz>
9357 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
9359 (dump_cgraph_node): Dump profile-id.
9360 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
9361 and common_target_probability.
9362 * lto-cgraph.c (lto_output_edge): Stream common targets.
9363 (lto_output_node): Stream profile ids.
9364 (input_node): Stream profile ids.
9365 (input_edge): Stream common targets.
9366 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
9367 * ipa.c: Include value-prof.h
9368 (ipa_profile_generate_summary): Turn indirect call statement histograms
9369 into common targets.
9370 (ipa_profile): Turn common targets into speculative edges.
9372 2013-08-09 Jan Hubicka <jh@suse.cz>
9374 * cgraph.h (cgraph_node): Add profile_id.
9375 * value-prof.c (cgraph_node_map): Turn into pointer_map.
9376 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
9377 (del_node_map): Update.
9378 (find_func_by_funcdef_no): Replace by ...
9379 (find_func_by_profile_id): ... this one.
9380 (gimple_ic_transform): Do not remove useful histograms when
9381 speculation is not done; dump info when indirect call removal
9383 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
9384 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
9385 (__gcov_indirect_call_profiler_v2): .. this one.
9386 * profile.h (init_node_map): Update.
9387 * coverage.c (coverage_compute_profile_id): New function.
9388 * coverage.h (coverage_compute_profile_id): Declare.
9389 * tree-profile.c (init_ic_make_global_vars): Make
9390 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
9391 (gimple_init_edge_profiler): Update prototype of
9392 __gcov_indirect_call_profiler.
9393 (gimple_gen_ic_func_profiler): Simplify.
9394 (tree_profiling): Use init_node_map
9396 2013-08-09 Jan Hubicka <jh@suse.cz>
9398 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
9399 non-speculative refs.
9400 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
9401 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
9402 (cgraph_set_call_stmt): Likewise.
9403 (cgraph_create_edge_1): Fix release checking compilatoin;
9405 (cgraph_free_edge): Free indirect info.
9406 (cgraph_turn_edge_to_speculative): New function.
9407 (cgraph_speculative_call_info): New function.
9408 (cgraph_make_edge_direct): Return direct edge; handle speculation.
9409 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
9410 (dump_cgraph_node): Dump speculation.
9411 (verify_edge_count_and_frequency): Accept speculative edges.
9412 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
9413 (verify_cgraph_node): Handle speculation.
9414 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
9415 (cgraph_set_call_stmt): Update prototype.
9416 (cgraph_make_edge_direct): Update prototype.
9417 (cgraph_speculative_call_info): Declare.
9418 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
9419 to change; update call of ipa_find_references.
9420 * ipa-ref.c (ipa_record_reference): Fix return value; clear
9421 lto_stmt_uid and speculative flags.
9422 (ipa_dump_references): Dump speculation.
9423 (ipa_clone_references): Clone speculative flag.
9424 (ipa_clone_referring): Likewise.
9425 (ipa_clone_ref): New function.
9426 (ipa_find_reference): Look into lto_stmt_uids
9427 (ipa_clear_stmts_in_references): Do not clear speculative calls.
9428 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
9429 (ipa_find_reference): Update declaration.
9430 (ipa_clone_ref): Declare.
9431 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
9432 stream speculative flag.
9433 (lto_output_ref): Stream statements uids and speculation.
9434 (input_ref): Likewise.
9435 (input_edge): Stream speuclation.
9436 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
9437 (cgraph_set_call_stmt_including_clones): Handle speculation.
9438 * ipa-inline.c (heap_edge_removal_hook): New function.
9439 (inline_small_functions): Register it.
9440 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
9441 also initialize refs.
9442 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
9444 (try_make_edge_direct_simple_call): Likewise.
9445 (try_make_edge_direct_simple_call): Likewise.
9446 (update_indirect_edges_after_inlining): Likewise.
9447 (remove_described_reference): Look proper lto_stmt_uid.
9448 (propagate_controlled_uses): Likewise.
9449 (propagate_controlled_uses): Liekwise.
9450 * tree-inline.c (copy_bb): Copy speculative edges.
9451 (redirect_all_calls): New function.
9452 (copy_cfg_body): Do redirection after loop info is updated.
9453 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
9455 2013-08-09 Jan Hubicka <jh@suse.cz>
9457 * lto-streamer-out.c (output_function): Renumber PHIs.
9458 * lto-streamer-in.c (input_function): Likewise.
9460 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
9462 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
9463 (get_lane_unsigned): Likewise.
9464 (dup_lane_scalar): Likewise.
9465 (get_lane): enable for VALL.
9466 * config/aarch64/aarch64-simd.md
9467 (aarch64_dup_lane_scalar<mode>): Remove.
9468 (aarch64_get_lane_signed<mode>): Likewise.
9469 (aarch64_get_lane_unsigned<mode>): Likewise.
9470 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
9471 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
9472 (aarch64_get_lane<mode>): Enable for all vector modes.
9473 (aarch64_get_lanedi): Remove misleading constraints.
9474 * config/aarch64/arm_neon.h
9475 (__aarch64_vget_lane_any): Define.
9476 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
9477 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
9478 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
9479 * config/aarch64/iterators.md (VDQQH): New.
9483 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
9485 * configure.ac: Add GAS check for LEON instructions on SPARC.
9486 * configure: Regenerate.
9487 * config.in: Likewise.
9488 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
9490 * config/sparc/sparc.opt (LEON, LEON3): New masks.
9491 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
9493 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
9494 (AS_LEON_FLAG): New macro.
9495 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
9496 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
9497 Deal with LEON and LEON3 for the memory model.
9498 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
9499 (atomic_compare_and_swap<mode>_1): Likewise.
9500 (*atomic_compare_and_swap<mode>_1): Likewise.
9502 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9504 * config/arm/neon.md (vcond): Fix floating-point vector
9505 comparisons against 0.
9507 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
9509 * lra-constraints.c (emit_spill_move): Remove assert.
9510 (process_alt_operands): Add more debugging
9511 output. Increase reject for spilling into memory. Decrease
9512 reject for reloading scratch.
9513 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
9515 2013-08-08 Steve Ellcey <sellcey@mips.com>
9517 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
9518 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
9519 micromips incompatible. Add nan2008.
9520 (MULTILIB_DIRNAMES): Add nan2008.
9521 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9522 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
9523 and micromips incompatible. Add nan2008.
9524 (MULTILIB_DIRNAMES): Add nan2008.
9525 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9527 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
9529 PR rtl-optimization/58079
9530 * combine.c (combine_simplify_rtx): Avoid using SUBST if
9531 simplify_comparison has widened a comparison with an integer.
9533 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9535 * config/arm/neon.md (movmisalign<mode>): Disable when we
9536 don't allow unaligned accesses.
9537 (*movmisalign<mode>_neon_store): Likewise.
9538 (*movmisalign<mode>_neon_load): Likewise.
9539 (*movmisalign<mode>_neon_store): Likewise.
9540 (*movmisalign<mode>_neon_load): Likewise.
9542 2013-08-08 Jan Hubicka <jh@suse.cz>
9544 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
9545 (make_pass_rebuild_cgraph_edges): Also clear references.
9546 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
9547 * ipa-inline-transform.c (inline_transform): Remove all references
9549 * cgraphunit.c (expand_function): Remove all references after
9551 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
9552 (ipa_find_reference): Rewrite to iterator.
9553 (remove_stmt_references): Likewise.
9554 (ipa_clear_stmts_in_references): New function.
9555 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
9556 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
9558 * ipa-split.c (split_function): Remove references in split function.
9560 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9563 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
9564 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
9566 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9569 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
9570 suggested mode for the assignment isn't compatible with the
9573 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9576 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
9578 2013-08-07 Xinliang David Li <davidxl@google.com>
9580 * config/i386/i386.opt: New option -mtune-ctrl=.
9581 * config/i386/x86-tune.def: New file.
9582 * config/i386/i386.h: include x86-tune.def.
9583 * config/i386/i386.c (ix86_option_override_internal):
9584 Parsing -mtune-ctrl= option and set tune features.
9586 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
9589 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
9590 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
9593 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9595 * rtl.h (update_alignments): Declare.
9596 * final.c (grow_label_align): New function extracted from...
9597 (shorten_branches): ...here. Call it.
9598 (update_alignments): New function.
9599 * reorg.c (sibling_labels): New variable.
9600 (get_label_before): Add SIBLING parameter. If it is non-zero, push
9601 the new label along with it onto the sibling_labels vector.
9602 (fill_simple_delay_slots): Adjust call to get_label_before.
9603 (fill_slots_from_thread): Likewise.
9604 (relax_delay_slots): Likewise.
9605 (make_return_insns): Likewise.
9606 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
9608 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9610 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
9611 document its semantics.
9612 (diagnostic_report_diagnostic): Adjust accordingly.
9614 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9616 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
9617 (sparc_option_override): ...and port to new C++ pass API.
9618 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
9620 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
9622 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
9624 2013-08-06 Caroline Tice <cmtice@google.com>
9626 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
9627 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
9628 * tree-pass.h: Add pass_vtable_verify.
9629 * varasm.c (assemble_variable): Add code to properly set the comdat
9630 section and name for the .vtable_map_vars section.
9631 (assemble_vtyv_preinit_initializer): New function.
9632 (default_sectin_type_flags): Make sure .vtable_map_vars section has
9634 * output.h: Add function decl for assemble_vtv_preinit_initializer.
9635 * vtable-verify.c: New file.
9636 * vtable-verify.h: New file.
9637 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
9638 initialiation levels.
9639 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
9640 * passes.def: Insert pass_vtable_verify.
9641 * aclocal.m4: Reorder includes.
9642 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
9643 -fvtv-counts options.
9644 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
9645 as appropriate, if -fvtable-verify=... is used.
9646 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
9647 -fvtable-verify=... is used.
9648 * Makefile.in (OBJS): Add vtable-verify.o to list.
9649 (vtable-verify.o): Add new build rule.
9650 (GTFILES): Add vtable-verify.c to list.
9651 * common.opt (fvtable-verify=): New flag.
9652 (vtv_priority): Values for fvtable-verify= flag.
9653 (fvtv-counts): New flag.
9654 (fvtv-debug): New flag.
9655 * tree.h (save_vtable_map_decl): New extern function decl.
9657 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9659 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
9660 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
9661 (pass_data_rl78_devirt): ...new pass_data instance and...
9662 (make_pass_rl78_devirt): ...new function.
9663 (rl78_asm_file_start): Port pass registration to new C++ API.
9665 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9667 * coretypes.h (rtl_opt_pass): Add.
9668 (gcc::context): Add.
9669 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
9670 (epiphany_init): Port to new C++ pass API.
9671 (epiphany_optimize_mode_switching): Likewise.
9672 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
9673 (pass_manager::get_pass_mode_switching): New.
9674 (pass_manager::get_pass_peephole2): New.
9675 * mode-switching.c (pass_mode_switching): Add clone method.
9676 * recog.c (pass_peephole2): Add clone method.
9677 (pass_split_all_insns): Add clone method.
9679 2013-08-06 David Malcolm <dmalcolm@redhat.com>
9681 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
9682 (mips_option_override): ...here, porting to new C++ API for passes.
9684 2013-08-06 Jan Hubicka <jh@suse.cz>
9686 * cgraph.c (cgraph_get_body): New function based on lto.c
9688 * cgraph.h (cgraph_get_body): Declare.
9689 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
9691 * cgraphunit.c (expand_function): Get body prior expanding.
9692 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
9693 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
9695 * passes.c (do_per_function_toporder): Get body.
9696 * tree-inline.c (expand_call_inline): Get body prior inlining it.
9697 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
9699 2013-08-06 Martin Jambor <mjambor@suse.cz>
9702 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
9703 is not re-finalized. Rename second parameter to no_collect.
9705 2013-08-06 Martin Jambor <mjambor@suse.cz>
9708 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
9709 MEM_REF has proper alignment information.
9711 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
9714 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
9716 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
9717 rtx (*) (rtx, ...) with insn_gen_fn.
9718 * genoutput.c (output_insn_data): Cast gen_? function pointers to
9719 insn_gen_fn::stored_funcptr. Add initializer braces.
9721 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9723 Rewrite how instances of passes are cloned to remove assumptions
9724 about their sizes (thus allowing pass subclasses to have
9725 additional data fields, albeit non-GC-managed ones at this point).
9727 * passes.c (make_pass_instance): Now that passes have clone
9728 methods, rewrite this function to eliminate XNEW and memcpy
9729 calls that used hardcoded sizes. Since this function no longer
9730 creates pass instances, rename it to...
9731 (add_pass_instance): ...this. Document the old way that passes were
9732 numbered and flagged, and rework this function to continue using it.
9733 (next_pass_1): Add an initial_pass argument for use by
9735 (position_pass): When adding multiple instances of a pass, use
9736 the pass's clone method, rather than relying on the XNEW/memcpy
9737 within the former make_pass_instance (now add_pass_instance).
9738 (pass_manager::pass_manager): When invoking next_pass_1, also supply
9739 the initial instance of the current pass within the pass manager.
9741 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9743 This is the automated part of the conversion of passes from C
9744 structs to C++ classes.
9746 Patch autogenerated by refactor_passes.py from
9747 https://github.com/davidmalcolm/gcc-refactoring-scripts
9748 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
9750 * asan.c (pass_asan): Convert from a global struct to a subclass of
9751 gimple_opt_pass along with...
9752 (pass_data_asan): ...new pass_data instance and...
9753 (make_pass_asan): ...new function.
9754 (pass_asan_O0): Convert from a global struct to a subclass of
9755 gimple_opt_pass along with...
9756 (pass_data_asan_O0): ...new pass_data instance and...
9757 (make_pass_asan_O0): ...new function.
9758 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
9759 subclass of rtl_opt_pass along with...
9760 (pass_data_inc_dec): ...new pass_data instance and...
9761 (make_pass_inc_dec): ...new function.
9762 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
9763 a subclass of rtl_opt_pass along with...
9764 (pass_data_reorder_blocks): ...new pass_data instance and...
9765 (make_pass_reorder_blocks): ...new function.
9766 (pass_duplicate_computed_gotos): Convert from a global struct to a
9767 subclass of rtl_opt_pass along with...
9768 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
9769 (make_pass_duplicate_computed_gotos): ...new function.
9770 (pass_partition_blocks): Convert from a global struct to a subclass of
9771 rtl_opt_pass along with...
9772 (pass_data_partition_blocks): ...new pass_data instance and...
9773 (make_pass_partition_blocks): ...new function.
9774 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
9775 struct to a subclass of rtl_opt_pass along with...
9776 (pass_data_branch_target_load_optimize1): ...new pass_data instance
9778 (make_pass_branch_target_load_optimize1): ...new function.
9779 (pass_branch_target_load_optimize2): Convert from a global struct to a
9780 subclass of rtl_opt_pass along with...
9781 (pass_data_branch_target_load_optimize2): ...new pass_data instance
9783 (make_pass_branch_target_load_optimize2): ...new function.
9784 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
9785 of rtl_opt_pass along with...
9786 (pass_data_jump): ...new pass_data instance and...
9787 (make_pass_jump): ...new function.
9788 (pass_jump2): Convert from a global struct to a subclass of
9789 rtl_opt_pass along with...
9790 (pass_data_jump2): ...new pass_data instance and...
9791 (make_pass_jump2): ...new function.
9792 * cfgexpand.c (pass_expand): Convert from a global struct to a
9793 subclass of rtl_opt_pass along with...
9794 (pass_data_expand): ...new pass_data instance and...
9795 (make_pass_expand): ...new function.
9796 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
9797 of rtl_opt_pass along with...
9798 (pass_data_free_cfg): ...new pass_data instance and...
9799 (make_pass_free_cfg): ...new function.
9800 (pass_into_cfg_layout_mode): Convert from a global struct to a
9801 subclass of rtl_opt_pass along with...
9802 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
9803 (make_pass_into_cfg_layout_mode): ...new function.
9804 (pass_outof_cfg_layout_mode): Convert from a global struct to a
9805 subclass of rtl_opt_pass along with...
9806 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
9807 (make_pass_outof_cfg_layout_mode): ...new function.
9808 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
9809 struct to a subclass of gimple_opt_pass along with...
9810 (pass_data_build_cgraph_edges): ...new pass_data instance and...
9811 (make_pass_build_cgraph_edges): ...new function.
9812 (pass_rebuild_cgraph_edges): Convert from a global struct to a
9813 subclass of gimple_opt_pass along with...
9814 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
9815 (make_pass_rebuild_cgraph_edges): ...new function.
9816 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
9817 subclass of gimple_opt_pass along with...
9818 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
9820 (make_pass_remove_cgraph_callee_edges): ...new function.
9821 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
9822 struct to a subclass of rtl_opt_pass along with...
9823 (pass_data_stack_adjustments): ...new pass_data instance and...
9824 (make_pass_stack_adjustments): ...new function.
9825 * combine.c (pass_combine): Convert from a global struct to a subclass
9826 of rtl_opt_pass along with...
9827 (pass_data_combine): ...new pass_data instance and...
9828 (make_pass_combine): ...new function.
9829 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
9830 global struct to a subclass of rtl_opt_pass along with...
9831 (pass_data_compare_elim_after_reload): ...new pass_data instance
9833 (make_pass_compare_elim_after_reload): ...new function.
9834 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
9835 of rtl_opt_pass along with...
9836 (pass_data_rtl_cprop): ...new pass_data instance and...
9837 (make_pass_rtl_cprop): ...new function.
9838 * cse.c (pass_cse): Convert from a global struct to a subclass of
9839 rtl_opt_pass along with...
9840 (pass_data_cse): ...new pass_data instance and...
9841 (make_pass_cse): ...new function.
9842 (pass_cse2): Convert from a global struct to a subclass of
9843 rtl_opt_pass along with...
9844 (pass_data_cse2): ...new pass_data instance and...
9845 (make_pass_cse2): ...new function.
9846 (pass_cse_after_global_opts): Convert from a global struct to a
9847 subclass of rtl_opt_pass along with...
9848 (pass_data_cse_after_global_opts): ...new pass_data instance and...
9849 (make_pass_cse_after_global_opts): ...new function.
9850 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
9851 of rtl_opt_pass along with...
9852 (pass_data_ud_rtl_dce): ...new pass_data instance and...
9853 (make_pass_ud_rtl_dce): ...new function.
9854 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
9855 rtl_opt_pass along with...
9856 (pass_data_fast_rtl_dce): ...new pass_data instance and...
9857 (make_pass_fast_rtl_dce): ...new function.
9858 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
9859 a subclass of rtl_opt_pass along with...
9860 (pass_data_df_initialize_opt): ...new pass_data instance and...
9861 (make_pass_df_initialize_opt): ...new function.
9862 (pass_df_initialize_no_opt): Convert from a global struct to a
9863 subclass of rtl_opt_pass along with...
9864 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
9865 (make_pass_df_initialize_no_opt): ...new function.
9866 (pass_df_finish): Convert from a global struct to a subclass of
9867 rtl_opt_pass along with...
9868 (pass_data_df_finish): ...new pass_data instance and...
9869 (make_pass_df_finish): ...new function.
9870 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
9871 rtl_opt_pass along with...
9872 (pass_data_rtl_dse1): ...new pass_data instance and...
9873 (make_pass_rtl_dse1): ...new function.
9874 (pass_rtl_dse2): Convert from a global struct to a subclass of
9875 rtl_opt_pass along with...
9876 (pass_data_rtl_dse2): ...new pass_data instance and...
9877 (make_pass_rtl_dse2): ...new function.
9878 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
9879 subclass of rtl_opt_pass along with...
9880 (pass_data_dwarf2_frame): ...new pass_data instance and...
9881 (make_pass_dwarf2_frame): ...new function.
9882 * except.c (pass_set_nothrow_function_flags): Convert from a global
9883 struct to a subclass of rtl_opt_pass along with...
9884 (pass_data_set_nothrow_function_flags): ...new pass_data instance
9886 (make_pass_set_nothrow_function_flags): ...new function.
9887 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
9888 subclass of rtl_opt_pass along with...
9889 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
9891 (make_pass_convert_to_eh_region_ranges): ...new function.
9892 * final.c (pass_compute_alignments): Convert from a global struct to a
9893 subclass of rtl_opt_pass along with...
9894 (pass_data_compute_alignments): ...new pass_data instance and...
9895 (make_pass_compute_alignments): ...new function.
9896 (pass_final): Convert from a global struct to a subclass of
9897 rtl_opt_pass along with...
9898 (pass_data_final): ...new pass_data instance and...
9899 (make_pass_final): ...new function.
9900 (pass_shorten_branches): Convert from a global struct to a subclass of
9901 rtl_opt_pass along with...
9902 (pass_data_shorten_branches): ...new pass_data instance and...
9903 (make_pass_shorten_branches): ...new function.
9904 (pass_clean_state): Convert from a global struct to a subclass of
9905 rtl_opt_pass along with...
9906 (pass_data_clean_state): ...new pass_data instance and...
9907 (make_pass_clean_state): ...new function.
9908 * function.c (pass_instantiate_virtual_regs): Convert from a global
9909 struct to a subclass of rtl_opt_pass along with...
9910 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
9911 (make_pass_instantiate_virtual_regs): ...new function.
9912 (pass_leaf_regs): Convert from a global struct to a subclass of
9913 rtl_opt_pass along with...
9914 (pass_data_leaf_regs): ...new pass_data instance and...
9915 (make_pass_leaf_regs): ...new function.
9916 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
9917 subclass of rtl_opt_pass along with...
9918 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
9920 (make_pass_thread_prologue_and_epilogue): ...new function.
9921 (pass_match_asm_constraints): Convert from a global struct to a
9922 subclass of rtl_opt_pass along with...
9923 (pass_data_match_asm_constraints): ...new pass_data instance and...
9924 (make_pass_match_asm_constraints): ...new function.
9925 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
9926 subclass of rtl_opt_pass along with...
9927 (pass_data_rtl_fwprop): ...new pass_data instance and...
9928 (make_pass_rtl_fwprop): ...new function.
9929 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
9930 rtl_opt_pass along with...
9931 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
9932 (make_pass_rtl_fwprop_addr): ...new function.
9933 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
9934 rtl_opt_pass along with...
9935 (pass_data_rtl_pre): ...new pass_data instance and...
9936 (make_pass_rtl_pre): ...new function.
9937 (pass_rtl_hoist): Convert from a global struct to a subclass of
9938 rtl_opt_pass along with...
9939 (pass_data_rtl_hoist): ...new pass_data instance and...
9940 (make_pass_rtl_hoist): ...new function.
9941 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
9942 subclass of gimple_opt_pass along with...
9943 (pass_data_lower_cf): ...new pass_data instance and...
9944 (make_pass_lower_cf): ...new function.
9945 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
9946 from a global struct to a subclass of gimple_opt_pass along with...
9947 (pass_data_strength_reduction): ...new pass_data instance and...
9948 (make_pass_strength_reduction): ...new function.
9949 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
9950 of rtl_opt_pass along with...
9951 (pass_data_rtl_ifcvt): ...new pass_data instance and...
9952 (make_pass_rtl_ifcvt): ...new function.
9953 (pass_if_after_combine): Convert from a global struct to a subclass of
9954 rtl_opt_pass along with...
9955 (pass_data_if_after_combine): ...new pass_data instance and...
9956 (make_pass_if_after_combine): ...new function.
9957 (pass_if_after_reload): Convert from a global struct to a subclass of
9958 rtl_opt_pass along with...
9959 (pass_data_if_after_reload): ...new pass_data instance and...
9960 (make_pass_if_after_reload): ...new function.
9961 * init-regs.c (pass_initialize_regs): Convert from a global struct to
9962 a subclass of rtl_opt_pass along with...
9963 (pass_data_initialize_regs): ...new pass_data instance and...
9964 (make_pass_initialize_regs): ...new function.
9965 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
9966 of ipa_opt_pass_d along with...
9967 (pass_data_ipa_cp): ...new pass_data instance and...
9968 (make_pass_ipa_cp): ...new function.
9969 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
9970 global struct to a subclass of gimple_opt_pass along with...
9971 (pass_data_inline_parameters): ...new pass_data instance and...
9972 (make_pass_inline_parameters): ...new function.
9973 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
9974 subclass of gimple_opt_pass along with...
9975 (pass_data_early_inline): ...new pass_data instance and...
9976 (make_pass_early_inline): ...new function.
9977 (pass_ipa_inline): Convert from a global struct to a subclass of
9978 ipa_opt_pass_d along with...
9979 (pass_data_ipa_inline): ...new pass_data instance and...
9980 (make_pass_ipa_inline): ...new function.
9981 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
9982 struct to a subclass of gimple_opt_pass along with...
9983 (pass_data_local_pure_const): ...new pass_data instance and...
9984 (make_pass_local_pure_const): ...new function.
9985 (pass_ipa_pure_const): Convert from a global struct to a subclass of
9986 ipa_opt_pass_d along with...
9987 (pass_data_ipa_pure_const): ...new pass_data instance and...
9988 (make_pass_ipa_pure_const): ...new function.
9989 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
9990 to a subclass of ipa_opt_pass_d along with...
9991 (pass_data_ipa_reference): ...new pass_data instance and...
9992 (make_pass_ipa_reference): ...new function.
9993 * ipa-split.c (pass_split_functions): Convert from a global struct to
9994 a subclass of gimple_opt_pass along with...
9995 (pass_data_split_functions): ...new pass_data instance and...
9996 (make_pass_split_functions): ...new function.
9997 (pass_feedback_split_functions): Convert from a global struct to a
9998 subclass of gimple_opt_pass along with...
9999 (pass_data_feedback_split_functions): ...new pass_data instance and...
10000 (make_pass_feedback_split_functions): ...new function.
10001 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
10002 global struct to a subclass of simple_ipa_opt_pass along with...
10003 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
10005 (make_pass_ipa_function_and_variable_visibility): ...new function.
10006 (pass_ipa_free_inline_summary): Convert from a global struct to a
10007 subclass of simple_ipa_opt_pass along with...
10008 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
10009 (make_pass_ipa_free_inline_summary): ...new function.
10010 (pass_ipa_whole_program_visibility): Convert from a global struct to a
10011 subclass of ipa_opt_pass_d along with...
10012 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
10014 (make_pass_ipa_whole_program_visibility): ...new function.
10015 (pass_ipa_profile): Convert from a global struct to a subclass of
10016 ipa_opt_pass_d along with...
10017 (pass_data_ipa_profile): ...new pass_data instance and...
10018 (make_pass_ipa_profile): ...new function.
10019 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
10020 ipa_opt_pass_d along with...
10021 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
10022 (make_pass_ipa_cdtor_merge): ...new function.
10023 * ira.c (pass_ira): Convert from a global struct to a subclass of
10024 rtl_opt_pass along with...
10025 (pass_data_ira): ...new pass_data instance and...
10026 (make_pass_ira): ...new function.
10027 (pass_reload): Convert from a global struct to a subclass of
10028 rtl_opt_pass along with...
10029 (pass_data_reload): ...new pass_data instance and...
10030 (make_pass_reload): ...new function.
10031 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
10032 subclass of rtl_opt_pass along with...
10033 (pass_data_cleanup_barriers): ...new pass_data instance and...
10034 (make_pass_cleanup_barriers): ...new function.
10035 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
10036 of rtl_opt_pass along with...
10037 (pass_data_loop2): ...new pass_data instance and...
10038 (make_pass_loop2): ...new function.
10039 (pass_rtl_loop_init): Convert from a global struct to a subclass of
10040 rtl_opt_pass along with...
10041 (pass_data_rtl_loop_init): ...new pass_data instance and...
10042 (make_pass_rtl_loop_init): ...new function.
10043 (pass_rtl_loop_done): Convert from a global struct to a subclass of
10044 rtl_opt_pass along with...
10045 (pass_data_rtl_loop_done): ...new pass_data instance and...
10046 (make_pass_rtl_loop_done): ...new function.
10047 (pass_rtl_move_loop_invariants): Convert from a global struct to a
10048 subclass of rtl_opt_pass along with...
10049 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
10050 (make_pass_rtl_move_loop_invariants): ...new function.
10051 (pass_rtl_unswitch): Convert from a global struct to a subclass of
10052 rtl_opt_pass along with...
10053 (pass_data_rtl_unswitch): ...new pass_data instance and...
10054 (make_pass_rtl_unswitch): ...new function.
10055 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
10056 subclass of rtl_opt_pass along with...
10057 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
10059 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10060 (pass_rtl_doloop): Convert from a global struct to a subclass of
10061 rtl_opt_pass along with...
10062 (pass_data_rtl_doloop): ...new pass_data instance and...
10063 (make_pass_rtl_doloop): ...new function.
10064 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
10065 a subclass of rtl_opt_pass along with...
10066 (pass_data_lower_subreg): ...new pass_data instance and...
10067 (make_pass_lower_subreg): ...new function.
10068 (pass_lower_subreg2): Convert from a global struct to a subclass of
10069 rtl_opt_pass along with...
10070 (pass_data_lower_subreg2): ...new pass_data instance and...
10071 (make_pass_lower_subreg2): ...new function.
10072 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
10073 struct to a subclass of ipa_opt_pass_d along with...
10074 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
10075 (make_pass_ipa_lto_gimple_out): ...new function.
10076 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
10077 of ipa_opt_pass_d along with...
10078 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
10079 (make_pass_ipa_lto_finish_out): ...new function.
10080 * mode-switching.c (pass_mode_switching): Convert from a global struct
10081 to a subclass of rtl_opt_pass along with...
10082 (pass_data_mode_switching): ...new pass_data instance and...
10083 (make_pass_mode_switching): ...new function.
10084 * modulo-sched.c (pass_sms): Convert from a global struct to a
10085 subclass of rtl_opt_pass along with...
10086 (pass_data_sms): ...new pass_data instance and...
10087 (make_pass_sms): ...new function.
10088 * omp-low.c (pass_expand_omp): Convert from a global struct to a
10089 subclass of gimple_opt_pass along with...
10090 (pass_data_expand_omp): ...new pass_data instance and...
10091 (make_pass_expand_omp): ...new function.
10092 (pass_lower_omp): Convert from a global struct to a subclass of
10093 gimple_opt_pass along with...
10094 (pass_data_lower_omp): ...new pass_data instance and...
10095 (make_pass_lower_omp): ...new function.
10096 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
10097 of gimple_opt_pass along with...
10098 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
10099 (make_pass_diagnose_omp_blocks): ...new function.
10100 * passes.c (pass_early_local_passes): Convert from a global struct to
10101 a subclass of simple_ipa_opt_pass along with...
10102 (pass_data_early_local_passes): ...new pass_data instance and...
10103 (make_pass_early_local_passes): ...new function.
10104 (pass_all_early_optimizations): Convert from a global struct to a
10105 subclass of gimple_opt_pass along with...
10106 (pass_data_all_early_optimizations): ...new pass_data instance and...
10107 (make_pass_all_early_optimizations): ...new function.
10108 (pass_all_optimizations): Convert from a global struct to a subclass
10109 of gimple_opt_pass along with...
10110 (pass_data_all_optimizations): ...new pass_data instance and...
10111 (make_pass_all_optimizations): ...new function.
10112 (pass_all_optimizations_g): Convert from a global struct to a subclass
10113 of gimple_opt_pass along with...
10114 (pass_data_all_optimizations_g): ...new pass_data instance and...
10115 (make_pass_all_optimizations_g): ...new function.
10116 (pass_rest_of_compilation): Convert from a global struct to a subclass
10117 of rtl_opt_pass along with...
10118 (pass_data_rest_of_compilation): ...new pass_data instance and...
10119 (make_pass_rest_of_compilation): ...new function.
10120 (pass_postreload): Convert from a global struct to a subclass of
10121 rtl_opt_pass along with...
10122 (pass_data_postreload): ...new pass_data instance and...
10123 (make_pass_postreload): ...new function.
10124 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
10125 subclass of rtl_opt_pass along with...
10126 (pass_data_gcse2): ...new pass_data instance and...
10127 (make_pass_gcse2): ...new function.
10128 * postreload.c (pass_postreload_cse): Convert from a global struct to
10129 a subclass of rtl_opt_pass along with...
10130 (pass_data_postreload_cse): ...new pass_data instance and...
10131 (make_pass_postreload_cse): ...new function.
10132 * predict.c (pass_profile): Convert from a global struct to a subclass
10133 of gimple_opt_pass along with...
10134 (pass_data_profile): ...new pass_data instance and...
10135 (make_pass_profile): ...new function.
10136 (pass_strip_predict_hints): Convert from a global struct to a subclass
10137 of gimple_opt_pass along with...
10138 (pass_data_strip_predict_hints): ...new pass_data instance and...
10139 (make_pass_strip_predict_hints): ...new function.
10140 * recog.c (pass_peephole2): Convert from a global struct to a subclass
10141 of rtl_opt_pass along with...
10142 (pass_data_peephole2): ...new pass_data instance and...
10143 (make_pass_peephole2): ...new function.
10144 (pass_split_all_insns): Convert from a global struct to a subclass of
10145 rtl_opt_pass along with...
10146 (pass_data_split_all_insns): ...new pass_data instance and...
10147 (make_pass_split_all_insns): ...new function.
10148 (pass_split_after_reload): Convert from a global struct to a subclass
10149 of rtl_opt_pass along with...
10150 (pass_data_split_after_reload): ...new pass_data instance and...
10151 (make_pass_split_after_reload): ...new function.
10152 (pass_split_before_regstack): Convert from a global struct to a
10153 subclass of rtl_opt_pass along with...
10154 (pass_data_split_before_regstack): ...new pass_data instance and...
10155 (make_pass_split_before_regstack): ...new function.
10156 (pass_split_before_sched2): Convert from a global struct to a subclass
10157 of rtl_opt_pass along with...
10158 (pass_data_split_before_sched2): ...new pass_data instance and...
10159 (make_pass_split_before_sched2): ...new function.
10160 (pass_split_for_shorten_branches): Convert from a global struct to a
10161 subclass of rtl_opt_pass along with...
10162 (pass_data_split_for_shorten_branches): ...new pass_data instance
10164 (make_pass_split_for_shorten_branches): ...new function.
10165 * ree.c (pass_ree): Convert from a global struct to a subclass of
10166 rtl_opt_pass along with...
10167 (pass_data_ree): ...new pass_data instance and...
10168 (make_pass_ree): ...new function.
10169 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
10170 subclass of rtl_opt_pass along with...
10171 (pass_data_stack_regs): ...new pass_data instance and...
10172 (make_pass_stack_regs): ...new function.
10173 (pass_stack_regs_run): Convert from a global struct to a subclass of
10174 rtl_opt_pass along with...
10175 (pass_data_stack_regs_run): ...new pass_data instance and...
10176 (make_pass_stack_regs_run): ...new function.
10177 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
10178 subclass of rtl_opt_pass along with...
10179 (pass_data_cprop_hardreg): ...new pass_data instance and...
10180 (make_pass_cprop_hardreg): ...new function.
10181 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
10182 subclass of rtl_opt_pass along with...
10183 (pass_data_reginfo_init): ...new pass_data instance and...
10184 (make_pass_reginfo_init): ...new function.
10185 * regmove.c (pass_regmove): Convert from a global struct to a subclass
10186 of rtl_opt_pass along with...
10187 (pass_data_regmove): ...new pass_data instance and...
10188 (make_pass_regmove): ...new function.
10189 * regrename.c (pass_regrename): Convert from a global struct to a
10190 subclass of rtl_opt_pass along with...
10191 (pass_data_regrename): ...new pass_data instance and...
10192 (make_pass_regrename): ...new function.
10193 * reorg.c (pass_delay_slots): Convert from a global struct to a
10194 subclass of rtl_opt_pass along with...
10195 (pass_data_delay_slots): ...new pass_data instance and...
10196 (make_pass_delay_slots): ...new function.
10197 (pass_machine_reorg): Convert from a global struct to a subclass of
10198 rtl_opt_pass along with...
10199 (pass_data_machine_reorg): ...new pass_data instance and...
10200 (make_pass_machine_reorg): ...new function.
10201 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
10202 of rtl_opt_pass along with...
10203 (pass_data_sched): ...new pass_data instance and...
10204 (make_pass_sched): ...new function.
10205 (pass_sched2): Convert from a global struct to a subclass of
10206 rtl_opt_pass along with...
10207 (pass_data_sched2): ...new pass_data instance and...
10208 (make_pass_sched2): ...new function.
10209 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
10210 to a subclass of rtl_opt_pass along with...
10211 (pass_data_stack_ptr_mod): ...new pass_data instance and...
10212 (make_pass_stack_ptr_mod): ...new function.
10213 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
10214 to a subclass of rtl_opt_pass along with...
10215 (pass_data_rtl_store_motion): ...new pass_data instance and...
10216 (make_pass_rtl_store_motion): ...new function.
10217 * tracer.c (pass_tracer): Convert from a global struct to a subclass
10218 of gimple_opt_pass along with...
10219 (pass_data_tracer): ...new pass_data instance and...
10220 (make_pass_tracer): ...new function.
10221 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
10222 to a subclass of gimple_opt_pass along with...
10223 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
10224 (make_pass_diagnose_tm_blocks): ...new function.
10225 (pass_lower_tm): Convert from a global struct to a subclass of
10226 gimple_opt_pass along with...
10227 (pass_data_lower_tm): ...new pass_data instance and...
10228 (make_pass_lower_tm): ...new function.
10229 (pass_tm_init): Convert from a global struct to a subclass of
10230 gimple_opt_pass along with...
10231 (pass_data_tm_init): ...new pass_data instance and...
10232 (make_pass_tm_init): ...new function.
10233 (pass_tm_mark): Convert from a global struct to a subclass of
10234 gimple_opt_pass along with...
10235 (pass_data_tm_mark): ...new pass_data instance and...
10236 (make_pass_tm_mark): ...new function.
10237 (pass_tm_edges): Convert from a global struct to a subclass of
10238 gimple_opt_pass along with...
10239 (pass_data_tm_edges): ...new pass_data instance and...
10240 (make_pass_tm_edges): ...new function.
10241 (pass_tm_memopt): Convert from a global struct to a subclass of
10242 gimple_opt_pass along with...
10243 (pass_data_tm_memopt): ...new pass_data instance and...
10244 (make_pass_tm_memopt): ...new function.
10245 (pass_ipa_tm): Convert from a global struct to a subclass of
10246 simple_ipa_opt_pass along with...
10247 (pass_data_ipa_tm): ...new pass_data instance and...
10248 (make_pass_ipa_tm): ...new function.
10249 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
10250 subclass of gimple_opt_pass along with...
10251 (pass_data_call_cdce): ...new pass_data instance and...
10252 (make_pass_call_cdce): ...new function.
10253 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
10254 subclass of gimple_opt_pass along with...
10255 (pass_data_build_cfg): ...new pass_data instance and...
10256 (make_pass_build_cfg): ...new function.
10257 (pass_split_crit_edges): Convert from a global struct to a subclass of
10258 gimple_opt_pass along with...
10259 (pass_data_split_crit_edges): ...new pass_data instance and...
10260 (make_pass_split_crit_edges): ...new function.
10261 (pass_warn_function_return): Convert from a global struct to a
10262 subclass of gimple_opt_pass along with...
10263 (pass_data_warn_function_return): ...new pass_data instance and...
10264 (make_pass_warn_function_return): ...new function.
10265 (pass_warn_function_noreturn): Convert from a global struct to a
10266 subclass of gimple_opt_pass along with...
10267 (pass_data_warn_function_noreturn): ...new pass_data instance and...
10268 (make_pass_warn_function_noreturn): ...new function.
10269 (pass_warn_unused_result): Convert from a global struct to a subclass
10270 of gimple_opt_pass along with...
10271 (pass_data_warn_unused_result): ...new pass_data instance and...
10272 (make_pass_warn_unused_result): ...new function.
10273 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
10274 a subclass of gimple_opt_pass along with...
10275 (pass_data_merge_phi): ...new pass_data instance and...
10276 (make_pass_merge_phi): ...new function.
10277 * tree-complex.c (pass_lower_complex): Convert from a global struct to
10278 a subclass of gimple_opt_pass along with...
10279 (pass_data_lower_complex): ...new pass_data instance and...
10280 (make_pass_lower_complex): ...new function.
10281 (pass_lower_complex_O0): Convert from a global struct to a subclass of
10282 gimple_opt_pass along with...
10283 (pass_data_lower_complex_O0): ...new pass_data instance and...
10284 (make_pass_lower_complex_O0): ...new function.
10285 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
10286 subclass of gimple_opt_pass along with...
10287 (pass_data_lower_eh): ...new pass_data instance and...
10288 (make_pass_lower_eh): ...new function.
10289 (pass_refactor_eh): Convert from a global struct to a subclass of
10290 gimple_opt_pass along with...
10291 (pass_data_refactor_eh): ...new pass_data instance and...
10292 (make_pass_refactor_eh): ...new function.
10293 (pass_lower_resx): Convert from a global struct to a subclass of
10294 gimple_opt_pass along with...
10295 (pass_data_lower_resx): ...new pass_data instance and...
10296 (make_pass_lower_resx): ...new function.
10297 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
10298 of gimple_opt_pass along with...
10299 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
10300 (make_pass_lower_eh_dispatch): ...new function.
10301 (pass_cleanup_eh): Convert from a global struct to a subclass of
10302 gimple_opt_pass along with...
10303 (pass_data_cleanup_eh): ...new pass_data instance and...
10304 (make_pass_cleanup_eh): ...new function.
10305 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
10306 to a subclass of simple_ipa_opt_pass along with...
10307 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
10308 (make_pass_ipa_lower_emutls): ...new function.
10309 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
10310 a subclass of gimple_opt_pass along with...
10311 (pass_data_if_conversion): ...new pass_data instance and...
10312 (make_pass_if_conversion): ...new function.
10313 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
10314 subclass of gimple_opt_pass along with...
10315 (pass_data_build_ssa): ...new pass_data instance and...
10316 (make_pass_build_ssa): ...new function.
10317 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
10318 global struct to a subclass of gimple_opt_pass along with...
10319 (pass_data_loop_distribution): ...new pass_data instance and...
10320 (make_pass_loop_distribution): ...new function.
10321 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
10322 subclass of gimple_opt_pass along with...
10323 (pass_data_mudflap_1): ...new pass_data instance and...
10324 (make_pass_mudflap_1): ...new function.
10325 (pass_mudflap_2): Convert from a global struct to a subclass of
10326 gimple_opt_pass along with...
10327 (pass_data_mudflap_2): ...new pass_data instance and...
10328 (make_pass_mudflap_2): ...new function.
10329 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
10330 subclass of gimple_opt_pass along with...
10331 (pass_data_mudflap_1): ...new pass_data instance and...
10332 (make_pass_mudflap_1): ...new function.
10333 (pass_mudflap_2): Convert from a global struct to a subclass of
10334 gimple_opt_pass along with...
10335 (pass_data_mudflap_2): ...new pass_data instance and...
10336 (make_pass_mudflap_2): ...new function.
10337 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
10338 gimple_opt_pass along with...
10339 (pass_data_nrv): ...new pass_data instance and...
10340 (make_pass_nrv): ...new function.
10341 (pass_return_slot): Convert from a global struct to a subclass of
10342 gimple_opt_pass along with...
10343 (pass_data_return_slot): ...new pass_data instance and...
10344 (make_pass_return_slot): ...new function.
10345 * tree-object-size.c (pass_object_sizes): Convert from a global struct
10346 to a subclass of gimple_opt_pass along with...
10347 (pass_data_object_sizes): ...new pass_data instance and...
10348 (make_pass_object_sizes): ...new function.
10349 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
10350 global struct to a subclass of gimple_opt_pass along with...
10351 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
10353 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10354 (pass_fixup_cfg): Convert from a global struct to a subclass of
10355 gimple_opt_pass along with...
10356 (pass_data_fixup_cfg): ...new pass_data instance and...
10357 (make_pass_fixup_cfg): ...new function.
10358 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
10359 (make_pass_mudflap_1): ...new function.
10360 (pass_mudflap_2): Replace declaration with that of...
10361 (make_pass_mudflap_2): ...new function.
10362 (pass_asan): Replace declaration with that of...
10363 (make_pass_asan): ...new function.
10364 (pass_asan_O0): Replace declaration with that of...
10365 (make_pass_asan_O0): ...new function.
10366 (pass_tsan): Replace declaration with that of...
10367 (make_pass_tsan): ...new function.
10368 (pass_tsan_O0): Replace declaration with that of...
10369 (make_pass_tsan_O0): ...new function.
10370 (pass_lower_cf): Replace declaration with that of...
10371 (make_pass_lower_cf): ...new function.
10372 (pass_refactor_eh): Replace declaration with that of...
10373 (make_pass_refactor_eh): ...new function.
10374 (pass_lower_eh): Replace declaration with that of...
10375 (make_pass_lower_eh): ...new function.
10376 (pass_lower_eh_dispatch): Replace declaration with that of...
10377 (make_pass_lower_eh_dispatch): ...new function.
10378 (pass_lower_resx): Replace declaration with that of...
10379 (make_pass_lower_resx): ...new function.
10380 (pass_build_cfg): Replace declaration with that of...
10381 (make_pass_build_cfg): ...new function.
10382 (pass_early_tree_profile): Replace declaration with that of...
10383 (make_pass_early_tree_profile): ...new function.
10384 (pass_cleanup_eh): Replace declaration with that of...
10385 (make_pass_cleanup_eh): ...new function.
10386 (pass_sra): Replace declaration with that of...
10387 (make_pass_sra): ...new function.
10388 (pass_sra_early): Replace declaration with that of...
10389 (make_pass_sra_early): ...new function.
10390 (pass_early_ipa_sra): Replace declaration with that of...
10391 (make_pass_early_ipa_sra): ...new function.
10392 (pass_tail_recursion): Replace declaration with that of...
10393 (make_pass_tail_recursion): ...new function.
10394 (pass_tail_calls): Replace declaration with that of...
10395 (make_pass_tail_calls): ...new function.
10396 (pass_tree_loop): Replace declaration with that of...
10397 (make_pass_tree_loop): ...new function.
10398 (pass_tree_loop_init): Replace declaration with that of...
10399 (make_pass_tree_loop_init): ...new function.
10400 (pass_lim): Replace declaration with that of...
10401 (make_pass_lim): ...new function.
10402 (pass_tree_unswitch): Replace declaration with that of...
10403 (make_pass_tree_unswitch): ...new function.
10404 (pass_predcom): Replace declaration with that of...
10405 (make_pass_predcom): ...new function.
10406 (pass_iv_canon): Replace declaration with that of...
10407 (make_pass_iv_canon): ...new function.
10408 (pass_scev_cprop): Replace declaration with that of...
10409 (make_pass_scev_cprop): ...new function.
10410 (pass_empty_loop): Replace declaration with that of...
10411 (make_pass_empty_loop): ...new function.
10412 (pass_record_bounds): Replace declaration with that of...
10413 (make_pass_record_bounds): ...new function.
10414 (pass_graphite): Replace declaration with that of...
10415 (make_pass_graphite): ...new function.
10416 (pass_graphite_transforms): Replace declaration with that of...
10417 (make_pass_graphite_transforms): ...new function.
10418 (pass_if_conversion): Replace declaration with that of...
10419 (make_pass_if_conversion): ...new function.
10420 (pass_loop_distribution): Replace declaration with that of...
10421 (make_pass_loop_distribution): ...new function.
10422 (pass_vectorize): Replace declaration with that of...
10423 (make_pass_vectorize): ...new function.
10424 (pass_slp_vectorize): Replace declaration with that of...
10425 (make_pass_slp_vectorize): ...new function.
10426 (pass_complete_unroll): Replace declaration with that of...
10427 (make_pass_complete_unroll): ...new function.
10428 (pass_complete_unrolli): Replace declaration with that of...
10429 (make_pass_complete_unrolli): ...new function.
10430 (pass_parallelize_loops): Replace declaration with that of...
10431 (make_pass_parallelize_loops): ...new function.
10432 (pass_loop_prefetch): Replace declaration with that of...
10433 (make_pass_loop_prefetch): ...new function.
10434 (pass_iv_optimize): Replace declaration with that of...
10435 (make_pass_iv_optimize): ...new function.
10436 (pass_tree_loop_done): Replace declaration with that of...
10437 (make_pass_tree_loop_done): ...new function.
10438 (pass_ch): Replace declaration with that of...
10439 (make_pass_ch): ...new function.
10440 (pass_ccp): Replace declaration with that of...
10441 (make_pass_ccp): ...new function.
10442 (pass_phi_only_cprop): Replace declaration with that of...
10443 (make_pass_phi_only_cprop): ...new function.
10444 (pass_build_ssa): Replace declaration with that of...
10445 (make_pass_build_ssa): ...new function.
10446 (pass_build_alias): Replace declaration with that of...
10447 (make_pass_build_alias): ...new function.
10448 (pass_build_ealias): Replace declaration with that of...
10449 (make_pass_build_ealias): ...new function.
10450 (pass_dominator): Replace declaration with that of...
10451 (make_pass_dominator): ...new function.
10452 (pass_dce): Replace declaration with that of...
10453 (make_pass_dce): ...new function.
10454 (pass_dce_loop): Replace declaration with that of...
10455 (make_pass_dce_loop): ...new function.
10456 (pass_cd_dce): Replace declaration with that of...
10457 (make_pass_cd_dce): ...new function.
10458 (pass_call_cdce): Replace declaration with that of...
10459 (make_pass_call_cdce): ...new function.
10460 (pass_merge_phi): Replace declaration with that of...
10461 (make_pass_merge_phi): ...new function.
10462 (pass_split_crit_edges): Replace declaration with that of...
10463 (make_pass_split_crit_edges): ...new function.
10464 (pass_pre): Replace declaration with that of...
10465 (make_pass_pre): ...new function.
10466 (pass_profile): Replace declaration with that of...
10467 (make_pass_profile): ...new function.
10468 (pass_strip_predict_hints): Replace declaration with that of...
10469 (make_pass_strip_predict_hints): ...new function.
10470 (pass_lower_complex_O0): Replace declaration with that of...
10471 (make_pass_lower_complex_O0): ...new function.
10472 (pass_lower_complex): Replace declaration with that of...
10473 (make_pass_lower_complex): ...new function.
10474 (pass_lower_vector): Replace declaration with that of...
10475 (make_pass_lower_vector): ...new function.
10476 (pass_lower_vector_ssa): Replace declaration with that of...
10477 (make_pass_lower_vector_ssa): ...new function.
10478 (pass_lower_omp): Replace declaration with that of...
10479 (make_pass_lower_omp): ...new function.
10480 (pass_diagnose_omp_blocks): Replace declaration with that of...
10481 (make_pass_diagnose_omp_blocks): ...new function.
10482 (pass_expand_omp): Replace declaration with that of...
10483 (make_pass_expand_omp): ...new function.
10484 (pass_expand_omp_ssa): Replace declaration with that of...
10485 (make_pass_expand_omp_ssa): ...new function.
10486 (pass_object_sizes): Replace declaration with that of...
10487 (make_pass_object_sizes): ...new function.
10488 (pass_strlen): Replace declaration with that of...
10489 (make_pass_strlen): ...new function.
10490 (pass_fold_builtins): Replace declaration with that of...
10491 (make_pass_fold_builtins): ...new function.
10492 (pass_stdarg): Replace declaration with that of...
10493 (make_pass_stdarg): ...new function.
10494 (pass_early_warn_uninitialized): Replace declaration with that of...
10495 (make_pass_early_warn_uninitialized): ...new function.
10496 (pass_late_warn_uninitialized): Replace declaration with that of...
10497 (make_pass_late_warn_uninitialized): ...new function.
10498 (pass_cse_reciprocals): Replace declaration with that of...
10499 (make_pass_cse_reciprocals): ...new function.
10500 (pass_cse_sincos): Replace declaration with that of...
10501 (make_pass_cse_sincos): ...new function.
10502 (pass_optimize_bswap): Replace declaration with that of...
10503 (make_pass_optimize_bswap): ...new function.
10504 (pass_optimize_widening_mul): Replace declaration with that of...
10505 (make_pass_optimize_widening_mul): ...new function.
10506 (pass_warn_function_return): Replace declaration with that of...
10507 (make_pass_warn_function_return): ...new function.
10508 (pass_warn_function_noreturn): Replace declaration with that of...
10509 (make_pass_warn_function_noreturn): ...new function.
10510 (pass_cselim): Replace declaration with that of...
10511 (make_pass_cselim): ...new function.
10512 (pass_phiopt): Replace declaration with that of...
10513 (make_pass_phiopt): ...new function.
10514 (pass_forwprop): Replace declaration with that of...
10515 (make_pass_forwprop): ...new function.
10516 (pass_phiprop): Replace declaration with that of...
10517 (make_pass_phiprop): ...new function.
10518 (pass_tree_ifcombine): Replace declaration with that of...
10519 (make_pass_tree_ifcombine): ...new function.
10520 (pass_dse): Replace declaration with that of...
10521 (make_pass_dse): ...new function.
10522 (pass_nrv): Replace declaration with that of...
10523 (make_pass_nrv): ...new function.
10524 (pass_rename_ssa_copies): Replace declaration with that of...
10525 (make_pass_rename_ssa_copies): ...new function.
10526 (pass_sink_code): Replace declaration with that of...
10527 (make_pass_sink_code): ...new function.
10528 (pass_fre): Replace declaration with that of...
10529 (make_pass_fre): ...new function.
10530 (pass_check_data_deps): Replace declaration with that of...
10531 (make_pass_check_data_deps): ...new function.
10532 (pass_copy_prop): Replace declaration with that of...
10533 (make_pass_copy_prop): ...new function.
10534 (pass_vrp): Replace declaration with that of...
10535 (make_pass_vrp): ...new function.
10536 (pass_uncprop): Replace declaration with that of...
10537 (make_pass_uncprop): ...new function.
10538 (pass_return_slot): Replace declaration with that of...
10539 (make_pass_return_slot): ...new function.
10540 (pass_reassoc): Replace declaration with that of...
10541 (make_pass_reassoc): ...new function.
10542 (pass_rebuild_cgraph_edges): Replace declaration with that of...
10543 (make_pass_rebuild_cgraph_edges): ...new function.
10544 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
10545 (make_pass_remove_cgraph_callee_edges): ...new function.
10546 (pass_build_cgraph_edges): Replace declaration with that of...
10547 (make_pass_build_cgraph_edges): ...new function.
10548 (pass_local_pure_const): Replace declaration with that of...
10549 (make_pass_local_pure_const): ...new function.
10550 (pass_tracer): Replace declaration with that of...
10551 (make_pass_tracer): ...new function.
10552 (pass_warn_unused_result): Replace declaration with that of...
10553 (make_pass_warn_unused_result): ...new function.
10554 (pass_diagnose_tm_blocks): Replace declaration with that of...
10555 (make_pass_diagnose_tm_blocks): ...new function.
10556 (pass_lower_tm): Replace declaration with that of...
10557 (make_pass_lower_tm): ...new function.
10558 (pass_tm_init): Replace declaration with that of...
10559 (make_pass_tm_init): ...new function.
10560 (pass_tm_mark): Replace declaration with that of...
10561 (make_pass_tm_mark): ...new function.
10562 (pass_tm_memopt): Replace declaration with that of...
10563 (make_pass_tm_memopt): ...new function.
10564 (pass_tm_edges): Replace declaration with that of...
10565 (make_pass_tm_edges): ...new function.
10566 (pass_split_functions): Replace declaration with that of...
10567 (make_pass_split_functions): ...new function.
10568 (pass_feedback_split_functions): Replace declaration with that of...
10569 (make_pass_feedback_split_functions): ...new function.
10570 (pass_strength_reduction): Replace declaration with that of...
10571 (make_pass_strength_reduction): ...new function.
10572 (pass_ipa_lower_emutls): Replace declaration with that of...
10573 (make_pass_ipa_lower_emutls): ...new function.
10574 (pass_ipa_function_and_variable_visibility): Replace declaration with
10576 (make_pass_ipa_function_and_variable_visibility): ...new function.
10577 (pass_ipa_tree_profile): Replace declaration with that of...
10578 (make_pass_ipa_tree_profile): ...new function.
10579 (pass_early_local_passes): Replace declaration with that of...
10580 (make_pass_early_local_passes): ...new function.
10581 (pass_ipa_whole_program_visibility): Replace declaration with that
10583 (make_pass_ipa_whole_program_visibility): ...new function.
10584 (pass_ipa_lto_gimple_out): Replace declaration with that of...
10585 (make_pass_ipa_lto_gimple_out): ...new function.
10586 (pass_ipa_increase_alignment): Replace declaration with that of...
10587 (make_pass_ipa_increase_alignment): ...new function.
10588 (pass_ipa_inline): Replace declaration with that of...
10589 (make_pass_ipa_inline): ...new function.
10590 (pass_ipa_free_lang_data): Replace declaration with that of...
10591 (make_pass_ipa_free_lang_data): ...new function.
10592 (pass_ipa_free_inline_summary): Replace declaration with that of...
10593 (make_pass_ipa_free_inline_summary): ...new function.
10594 (pass_ipa_cp): Replace declaration with that of...
10595 (make_pass_ipa_cp): ...new function.
10596 (pass_ipa_reference): Replace declaration with that of...
10597 (make_pass_ipa_reference): ...new function.
10598 (pass_ipa_pure_const): Replace declaration with that of...
10599 (make_pass_ipa_pure_const): ...new function.
10600 (pass_ipa_pta): Replace declaration with that of...
10601 (make_pass_ipa_pta): ...new function.
10602 (pass_ipa_lto_finish_out): Replace declaration with that of...
10603 (make_pass_ipa_lto_finish_out): ...new function.
10604 (pass_ipa_tm): Replace declaration with that of...
10605 (make_pass_ipa_tm): ...new function.
10606 (pass_ipa_profile): Replace declaration with that of...
10607 (make_pass_ipa_profile): ...new function.
10608 (pass_ipa_cdtor_merge): Replace declaration with that of...
10609 (make_pass_ipa_cdtor_merge): ...new function.
10610 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
10612 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10613 (pass_init_datastructures): Replace declaration with that of...
10614 (make_pass_init_datastructures): ...new function.
10615 (pass_fixup_cfg): Replace declaration with that of...
10616 (make_pass_fixup_cfg): ...new function.
10617 (pass_expand): Replace declaration with that of...
10618 (make_pass_expand): ...new function.
10619 (pass_instantiate_virtual_regs): Replace declaration with that of...
10620 (make_pass_instantiate_virtual_regs): ...new function.
10621 (pass_rtl_fwprop): Replace declaration with that of...
10622 (make_pass_rtl_fwprop): ...new function.
10623 (pass_rtl_fwprop_addr): Replace declaration with that of...
10624 (make_pass_rtl_fwprop_addr): ...new function.
10625 (pass_jump): Replace declaration with that of...
10626 (make_pass_jump): ...new function.
10627 (pass_jump2): Replace declaration with that of...
10628 (make_pass_jump2): ...new function.
10629 (pass_lower_subreg): Replace declaration with that of...
10630 (make_pass_lower_subreg): ...new function.
10631 (pass_cse): Replace declaration with that of...
10632 (make_pass_cse): ...new function.
10633 (pass_fast_rtl_dce): Replace declaration with that of...
10634 (make_pass_fast_rtl_dce): ...new function.
10635 (pass_ud_rtl_dce): Replace declaration with that of...
10636 (make_pass_ud_rtl_dce): ...new function.
10637 (pass_rtl_dce): Replace declaration with that of...
10638 (make_pass_rtl_dce): ...new function.
10639 (pass_rtl_dse1): Replace declaration with that of...
10640 (make_pass_rtl_dse1): ...new function.
10641 (pass_rtl_dse2): Replace declaration with that of...
10642 (make_pass_rtl_dse2): ...new function.
10643 (pass_rtl_dse3): Replace declaration with that of...
10644 (make_pass_rtl_dse3): ...new function.
10645 (pass_rtl_cprop): Replace declaration with that of...
10646 (make_pass_rtl_cprop): ...new function.
10647 (pass_rtl_pre): Replace declaration with that of...
10648 (make_pass_rtl_pre): ...new function.
10649 (pass_rtl_hoist): Replace declaration with that of...
10650 (make_pass_rtl_hoist): ...new function.
10651 (pass_rtl_store_motion): Replace declaration with that of...
10652 (make_pass_rtl_store_motion): ...new function.
10653 (pass_cse_after_global_opts): Replace declaration with that of...
10654 (make_pass_cse_after_global_opts): ...new function.
10655 (pass_rtl_ifcvt): Replace declaration with that of...
10656 (make_pass_rtl_ifcvt): ...new function.
10657 (pass_into_cfg_layout_mode): Replace declaration with that of...
10658 (make_pass_into_cfg_layout_mode): ...new function.
10659 (pass_outof_cfg_layout_mode): Replace declaration with that of...
10660 (make_pass_outof_cfg_layout_mode): ...new function.
10661 (pass_loop2): Replace declaration with that of...
10662 (make_pass_loop2): ...new function.
10663 (pass_rtl_loop_init): Replace declaration with that of...
10664 (make_pass_rtl_loop_init): ...new function.
10665 (pass_rtl_move_loop_invariants): Replace declaration with that of...
10666 (make_pass_rtl_move_loop_invariants): ...new function.
10667 (pass_rtl_unswitch): Replace declaration with that of...
10668 (make_pass_rtl_unswitch): ...new function.
10669 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
10670 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10671 (pass_rtl_doloop): Replace declaration with that of...
10672 (make_pass_rtl_doloop): ...new function.
10673 (pass_rtl_loop_done): Replace declaration with that of...
10674 (make_pass_rtl_loop_done): ...new function.
10675 (pass_web): Replace declaration with that of...
10676 (make_pass_web): ...new function.
10677 (pass_cse2): Replace declaration with that of...
10678 (make_pass_cse2): ...new function.
10679 (pass_df_initialize_opt): Replace declaration with that of...
10680 (make_pass_df_initialize_opt): ...new function.
10681 (pass_df_initialize_no_opt): Replace declaration with that of...
10682 (make_pass_df_initialize_no_opt): ...new function.
10683 (pass_reginfo_init): Replace declaration with that of...
10684 (make_pass_reginfo_init): ...new function.
10685 (pass_inc_dec): Replace declaration with that of...
10686 (make_pass_inc_dec): ...new function.
10687 (pass_stack_ptr_mod): Replace declaration with that of...
10688 (make_pass_stack_ptr_mod): ...new function.
10689 (pass_initialize_regs): Replace declaration with that of...
10690 (make_pass_initialize_regs): ...new function.
10691 (pass_combine): Replace declaration with that of...
10692 (make_pass_combine): ...new function.
10693 (pass_if_after_combine): Replace declaration with that of...
10694 (make_pass_if_after_combine): ...new function.
10695 (pass_ree): Replace declaration with that of...
10696 (make_pass_ree): ...new function.
10697 (pass_partition_blocks): Replace declaration with that of...
10698 (make_pass_partition_blocks): ...new function.
10699 (pass_match_asm_constraints): Replace declaration with that of...
10700 (make_pass_match_asm_constraints): ...new function.
10701 (pass_regmove): Replace declaration with that of...
10702 (make_pass_regmove): ...new function.
10703 (pass_split_all_insns): Replace declaration with that of...
10704 (make_pass_split_all_insns): ...new function.
10705 (pass_fast_rtl_byte_dce): Replace declaration with that of...
10706 (make_pass_fast_rtl_byte_dce): ...new function.
10707 (pass_lower_subreg2): Replace declaration with that of...
10708 (make_pass_lower_subreg2): ...new function.
10709 (pass_mode_switching): Replace declaration with that of...
10710 (make_pass_mode_switching): ...new function.
10711 (pass_sms): Replace declaration with that of...
10712 (make_pass_sms): ...new function.
10713 (pass_sched): Replace declaration with that of...
10714 (make_pass_sched): ...new function.
10715 (pass_ira): Replace declaration with that of...
10716 (make_pass_ira): ...new function.
10717 (pass_reload): Replace declaration with that of...
10718 (make_pass_reload): ...new function.
10719 (pass_clean_state): Replace declaration with that of...
10720 (make_pass_clean_state): ...new function.
10721 (pass_branch_prob): Replace declaration with that of...
10722 (make_pass_branch_prob): ...new function.
10723 (pass_value_profile_transformations): Replace declaration with that
10725 (make_pass_value_profile_transformations): ...new function.
10726 (pass_postreload_cse): Replace declaration with that of...
10727 (make_pass_postreload_cse): ...new function.
10728 (pass_gcse2): Replace declaration with that of...
10729 (make_pass_gcse2): ...new function.
10730 (pass_split_after_reload): Replace declaration with that of...
10731 (make_pass_split_after_reload): ...new function.
10732 (pass_branch_target_load_optimize1): Replace declaration with that
10734 (make_pass_branch_target_load_optimize1): ...new function.
10735 (pass_thread_prologue_and_epilogue): Replace declaration with that
10737 (make_pass_thread_prologue_and_epilogue): ...new function.
10738 (pass_stack_adjustments): Replace declaration with that of...
10739 (make_pass_stack_adjustments): ...new function.
10740 (pass_peephole2): Replace declaration with that of...
10741 (make_pass_peephole2): ...new function.
10742 (pass_if_after_reload): Replace declaration with that of...
10743 (make_pass_if_after_reload): ...new function.
10744 (pass_regrename): Replace declaration with that of...
10745 (make_pass_regrename): ...new function.
10746 (pass_cprop_hardreg): Replace declaration with that of...
10747 (make_pass_cprop_hardreg): ...new function.
10748 (pass_reorder_blocks): Replace declaration with that of...
10749 (make_pass_reorder_blocks): ...new function.
10750 (pass_branch_target_load_optimize2): Replace declaration with that
10752 (make_pass_branch_target_load_optimize2): ...new function.
10753 (pass_leaf_regs): Replace declaration with that of...
10754 (make_pass_leaf_regs): ...new function.
10755 (pass_split_before_sched2): Replace declaration with that of...
10756 (make_pass_split_before_sched2): ...new function.
10757 (pass_compare_elim_after_reload): Replace declaration with that of...
10758 (make_pass_compare_elim_after_reload): ...new function.
10759 (pass_sched2): Replace declaration with that of...
10760 (make_pass_sched2): ...new function.
10761 (pass_stack_regs): Replace declaration with that of...
10762 (make_pass_stack_regs): ...new function.
10763 (pass_stack_regs_run): Replace declaration with that of...
10764 (make_pass_stack_regs_run): ...new function.
10765 (pass_df_finish): Replace declaration with that of...
10766 (make_pass_df_finish): ...new function.
10767 (pass_compute_alignments): Replace declaration with that of...
10768 (make_pass_compute_alignments): ...new function.
10769 (pass_duplicate_computed_gotos): Replace declaration with that of...
10770 (make_pass_duplicate_computed_gotos): ...new function.
10771 (pass_variable_tracking): Replace declaration with that of...
10772 (make_pass_variable_tracking): ...new function.
10773 (pass_free_cfg): Replace declaration with that of...
10774 (make_pass_free_cfg): ...new function.
10775 (pass_machine_reorg): Replace declaration with that of...
10776 (make_pass_machine_reorg): ...new function.
10777 (pass_cleanup_barriers): Replace declaration with that of...
10778 (make_pass_cleanup_barriers): ...new function.
10779 (pass_delay_slots): Replace declaration with that of...
10780 (make_pass_delay_slots): ...new function.
10781 (pass_split_for_shorten_branches): Replace declaration with that of...
10782 (make_pass_split_for_shorten_branches): ...new function.
10783 (pass_split_before_regstack): Replace declaration with that of...
10784 (make_pass_split_before_regstack): ...new function.
10785 (pass_convert_to_eh_region_ranges): Replace declaration with that
10787 (make_pass_convert_to_eh_region_ranges): ...new function.
10788 (pass_shorten_branches): Replace declaration with that of...
10789 (make_pass_shorten_branches): ...new function.
10790 (pass_set_nothrow_function_flags): Replace declaration with that of...
10791 (make_pass_set_nothrow_function_flags): ...new function.
10792 (pass_dwarf2_frame): Replace declaration with that of...
10793 (make_pass_dwarf2_frame): ...new function.
10794 (pass_final): Replace declaration with that of...
10795 (make_pass_final): ...new function.
10796 (pass_rtl_seqabstr): Replace declaration with that of...
10797 (make_pass_rtl_seqabstr): ...new function.
10798 (pass_release_ssa_names): Replace declaration with that of...
10799 (make_pass_release_ssa_names): ...new function.
10800 (pass_early_inline): Replace declaration with that of...
10801 (make_pass_early_inline): ...new function.
10802 (pass_inline_parameters): Replace declaration with that of...
10803 (make_pass_inline_parameters): ...new function.
10804 (pass_update_address_taken): Replace declaration with that of...
10805 (make_pass_update_address_taken): ...new function.
10806 (pass_convert_switch): Replace declaration with that of...
10807 (make_pass_convert_switch): ...new function.
10808 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
10809 to a subclass of simple_ipa_opt_pass along with...
10810 (pass_data_ipa_tree_profile): ...new pass_data instance and...
10811 (make_pass_ipa_tree_profile): ...new function.
10812 * tree-sra.c (pass_sra_early): Convert from a global struct to a
10813 subclass of gimple_opt_pass along with...
10814 (pass_data_sra_early): ...new pass_data instance and...
10815 (make_pass_sra_early): ...new function.
10816 (pass_sra): Convert from a global struct to a subclass of
10817 gimple_opt_pass along with...
10818 (pass_data_sra): ...new pass_data instance and...
10819 (make_pass_sra): ...new function.
10820 (pass_early_ipa_sra): Convert from a global struct to a subclass of
10821 gimple_opt_pass along with...
10822 (pass_data_early_ipa_sra): ...new pass_data instance and...
10823 (make_pass_early_ipa_sra): ...new function.
10824 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
10825 subclass of gimple_opt_pass along with...
10826 (pass_data_ccp): ...new pass_data instance and...
10827 (make_pass_ccp): ...new function.
10828 (pass_fold_builtins): Convert from a global struct to a subclass of
10829 gimple_opt_pass along with...
10830 (pass_data_fold_builtins): ...new pass_data instance and...
10831 (make_pass_fold_builtins): ...new function.
10832 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
10833 subclass of gimple_opt_pass along with...
10834 (pass_data_copy_prop): ...new pass_data instance and...
10835 (make_pass_copy_prop): ...new function.
10836 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
10837 global struct to a subclass of gimple_opt_pass along with...
10838 (pass_data_rename_ssa_copies): ...new pass_data instance and...
10839 (make_pass_rename_ssa_copies): ...new function.
10840 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
10841 subclass of gimple_opt_pass along with...
10842 (pass_data_dce): ...new pass_data instance and...
10843 (make_pass_dce): ...new function.
10844 (pass_dce_loop): Convert from a global struct to a subclass of
10845 gimple_opt_pass along with...
10846 (pass_data_dce_loop): ...new pass_data instance and...
10847 (make_pass_dce_loop): ...new function.
10848 (pass_cd_dce): Convert from a global struct to a subclass of
10849 gimple_opt_pass along with...
10850 (pass_data_cd_dce): ...new pass_data instance and...
10851 (make_pass_cd_dce): ...new function.
10852 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
10853 subclass of gimple_opt_pass along with...
10854 (pass_data_dominator): ...new pass_data instance and...
10855 (make_pass_dominator): ...new function.
10856 (pass_phi_only_cprop): Convert from a global struct to a subclass of
10857 gimple_opt_pass along with...
10858 (pass_data_phi_only_cprop): ...new pass_data instance and...
10859 (make_pass_phi_only_cprop): ...new function.
10860 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
10861 subclass of gimple_opt_pass along with...
10862 (pass_data_dse): ...new pass_data instance and...
10863 (make_pass_dse): ...new function.
10864 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
10865 a subclass of gimple_opt_pass along with...
10866 (pass_data_forwprop): ...new pass_data instance and...
10867 (make_pass_forwprop): ...new function.
10868 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
10869 struct to a subclass of gimple_opt_pass along with...
10870 (pass_data_tree_ifcombine): ...new pass_data instance and...
10871 (make_pass_tree_ifcombine): ...new function.
10872 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
10873 subclass of gimple_opt_pass along with...
10874 (pass_data_ch): ...new pass_data instance and...
10875 (make_pass_ch): ...new function.
10876 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
10877 subclass of gimple_opt_pass along with...
10878 (pass_data_tree_loop): ...new pass_data instance and...
10879 (make_pass_tree_loop): ...new function.
10880 (pass_tree_loop_init): Convert from a global struct to a subclass of
10881 gimple_opt_pass along with...
10882 (pass_data_tree_loop_init): ...new pass_data instance and...
10883 (make_pass_tree_loop_init): ...new function.
10884 (pass_lim): Convert from a global struct to a subclass of
10885 gimple_opt_pass along with...
10886 (pass_data_lim): ...new pass_data instance and...
10887 (make_pass_lim): ...new function.
10888 (pass_tree_unswitch): Convert from a global struct to a subclass of
10889 gimple_opt_pass along with...
10890 (pass_data_tree_unswitch): ...new pass_data instance and...
10891 (make_pass_tree_unswitch): ...new function.
10892 (pass_predcom): Convert from a global struct to a subclass of
10893 gimple_opt_pass along with...
10894 (pass_data_predcom): ...new pass_data instance and...
10895 (make_pass_predcom): ...new function.
10896 (pass_vectorize): Convert from a global struct to a subclass of
10897 gimple_opt_pass along with...
10898 (pass_data_vectorize): ...new pass_data instance and...
10899 (make_pass_vectorize): ...new function.
10900 (pass_graphite): Convert from a global struct to a subclass of
10901 gimple_opt_pass along with...
10902 (pass_data_graphite): ...new pass_data instance and...
10903 (make_pass_graphite): ...new function.
10904 (pass_graphite_transforms): Convert from a global struct to a subclass
10905 of gimple_opt_pass along with...
10906 (pass_data_graphite_transforms): ...new pass_data instance and...
10907 (make_pass_graphite_transforms): ...new function.
10908 (pass_check_data_deps): Convert from a global struct to a subclass of
10909 gimple_opt_pass along with...
10910 (pass_data_check_data_deps): ...new pass_data instance and...
10911 (make_pass_check_data_deps): ...new function.
10912 (pass_iv_canon): Convert from a global struct to a subclass of
10913 gimple_opt_pass along with...
10914 (pass_data_iv_canon): ...new pass_data instance and...
10915 (make_pass_iv_canon): ...new function.
10916 (pass_scev_cprop): Convert from a global struct to a subclass of
10917 gimple_opt_pass along with...
10918 (pass_data_scev_cprop): ...new pass_data instance and...
10919 (make_pass_scev_cprop): ...new function.
10920 (pass_record_bounds): Convert from a global struct to a subclass of
10921 gimple_opt_pass along with...
10922 (pass_data_record_bounds): ...new pass_data instance and...
10923 (make_pass_record_bounds): ...new function.
10924 (pass_complete_unroll): Convert from a global struct to a subclass of
10925 gimple_opt_pass along with...
10926 (pass_data_complete_unroll): ...new pass_data instance and...
10927 (make_pass_complete_unroll): ...new function.
10928 (pass_complete_unrolli): Convert from a global struct to a subclass of
10929 gimple_opt_pass along with...
10930 (pass_data_complete_unrolli): ...new pass_data instance and...
10931 (make_pass_complete_unrolli): ...new function.
10932 (pass_parallelize_loops): Convert from a global struct to a subclass
10933 of gimple_opt_pass along with...
10934 (pass_data_parallelize_loops): ...new pass_data instance and...
10935 (make_pass_parallelize_loops): ...new function.
10936 (pass_loop_prefetch): Convert from a global struct to a subclass of
10937 gimple_opt_pass along with...
10938 (pass_data_loop_prefetch): ...new pass_data instance and...
10939 (make_pass_loop_prefetch): ...new function.
10940 (pass_iv_optimize): Convert from a global struct to a subclass of
10941 gimple_opt_pass along with...
10942 (pass_data_iv_optimize): ...new pass_data instance and...
10943 (make_pass_iv_optimize): ...new function.
10944 (pass_tree_loop_done): Convert from a global struct to a subclass of
10945 gimple_opt_pass along with...
10946 (pass_data_tree_loop_done): ...new pass_data instance and...
10947 (make_pass_tree_loop_done): ...new function.
10948 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
10949 struct to a subclass of gimple_opt_pass along with...
10950 (pass_data_cse_reciprocals): ...new pass_data instance and...
10951 (make_pass_cse_reciprocals): ...new function.
10952 (pass_cse_sincos): Convert from a global struct to a subclass of
10953 gimple_opt_pass along with...
10954 (pass_data_cse_sincos): ...new pass_data instance and...
10955 (make_pass_cse_sincos): ...new function.
10956 (pass_optimize_bswap): Convert from a global struct to a subclass of
10957 gimple_opt_pass along with...
10958 (pass_data_optimize_bswap): ...new pass_data instance and...
10959 (make_pass_optimize_bswap): ...new function.
10960 (pass_optimize_widening_mul): Convert from a global struct to a
10961 subclass of gimple_opt_pass along with...
10962 (pass_data_optimize_widening_mul): ...new pass_data instance and...
10963 (make_pass_optimize_widening_mul): ...new function.
10964 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
10965 subclass of gimple_opt_pass along with...
10966 (pass_data_phiopt): ...new pass_data instance and...
10967 (make_pass_phiopt): ...new function.
10968 (pass_cselim): Convert from a global struct to a subclass of
10969 gimple_opt_pass along with...
10970 (pass_data_cselim): ...new pass_data instance and...
10971 (make_pass_cselim): ...new function.
10972 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
10973 subclass of gimple_opt_pass along with...
10974 (pass_data_phiprop): ...new pass_data instance and...
10975 (make_pass_phiprop): ...new function.
10976 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
10977 subclass of gimple_opt_pass along with...
10978 (pass_data_pre): ...new pass_data instance and...
10979 (make_pass_pre): ...new function.
10980 (pass_fre): Convert from a global struct to a subclass of
10981 gimple_opt_pass along with...
10982 (pass_data_fre): ...new pass_data instance and...
10983 (make_pass_fre): ...new function.
10984 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
10985 subclass of gimple_opt_pass along with...
10986 (pass_data_reassoc): ...new pass_data instance and...
10987 (make_pass_reassoc): ...new function.
10988 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
10989 subclass of gimple_opt_pass along with...
10990 (pass_data_sink_code): ...new pass_data instance and...
10991 (make_pass_sink_code): ...new function.
10992 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
10993 subclass of gimple_opt_pass along with...
10994 (pass_data_strlen): ...new pass_data instance and...
10995 (make_pass_strlen): ...new function.
10996 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
10997 struct to a subclass of gimple_opt_pass along with...
10998 (pass_data_build_alias): ...new pass_data instance and...
10999 (make_pass_build_alias): ...new function.
11000 (pass_build_ealias): Convert from a global struct to a subclass of
11001 gimple_opt_pass along with...
11002 (pass_data_build_ealias): ...new pass_data instance and...
11003 (make_pass_build_ealias): ...new function.
11004 (pass_ipa_pta): Convert from a global struct to a subclass of
11005 simple_ipa_opt_pass along with...
11006 (pass_data_ipa_pta): ...new pass_data instance and...
11007 (make_pass_ipa_pta): ...new function.
11008 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
11009 subclass of gimple_opt_pass along with...
11010 (pass_data_uncprop): ...new pass_data instance and...
11011 (make_pass_uncprop): ...new function.
11012 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
11013 global struct to a subclass of gimple_opt_pass along with...
11014 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
11015 (make_pass_late_warn_uninitialized): ...new function.
11016 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
11017 to a subclass of gimple_opt_pass along with...
11018 (pass_data_init_datastructures): ...new pass_data instance and...
11019 (make_pass_init_datastructures): ...new function.
11020 (pass_early_warn_uninitialized): Convert from a global struct to a
11021 subclass of gimple_opt_pass along with...
11022 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
11023 (make_pass_early_warn_uninitialized): ...new function.
11024 (pass_update_address_taken): Convert from a global struct to a
11025 subclass of gimple_opt_pass along with...
11026 (pass_data_update_address_taken): ...new pass_data instance and...
11027 (make_pass_update_address_taken): ...new function.
11028 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
11029 struct to a subclass of gimple_opt_pass along with...
11030 (pass_data_release_ssa_names): ...new pass_data instance and...
11031 (make_pass_release_ssa_names): ...new function.
11032 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
11033 subclass of gimple_opt_pass along with...
11034 (pass_data_stdarg): ...new pass_data instance and...
11035 (make_pass_stdarg): ...new function.
11036 * tree-switch-conversion.c (pass_convert_switch): Convert from a
11037 global struct to a subclass of gimple_opt_pass along with...
11038 (pass_data_convert_switch): ...new pass_data instance and...
11039 (make_pass_convert_switch): ...new function.
11040 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
11041 to a subclass of gimple_opt_pass along with...
11042 (pass_data_tail_recursion): ...new pass_data instance and...
11043 (make_pass_tail_recursion): ...new function.
11044 (pass_tail_calls): Convert from a global struct to a subclass of
11045 gimple_opt_pass along with...
11046 (pass_data_tail_calls): ...new pass_data instance and...
11047 (make_pass_tail_calls): ...new function.
11048 * tree-vect-generic.c (pass_lower_vector): Convert from a global
11049 struct to a subclass of gimple_opt_pass along with...
11050 (pass_data_lower_vector): ...new pass_data instance and...
11051 (make_pass_lower_vector): ...new function.
11052 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
11053 gimple_opt_pass along with...
11054 (pass_data_lower_vector_ssa): ...new pass_data instance and...
11055 (make_pass_lower_vector_ssa): ...new function.
11056 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
11057 to a subclass of gimple_opt_pass along with...
11058 (pass_data_slp_vectorize): ...new pass_data instance and...
11059 (make_pass_slp_vectorize): ...new function.
11060 (pass_ipa_increase_alignment): Convert from a global struct to a
11061 subclass of simple_ipa_opt_pass along with...
11062 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
11063 (make_pass_ipa_increase_alignment): ...new function.
11064 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
11065 gimple_opt_pass along with...
11066 (pass_data_vrp): ...new pass_data instance and...
11067 (make_pass_vrp): ...new function.
11068 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
11069 subclass of simple_ipa_opt_pass along with...
11070 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
11071 (make_pass_ipa_free_lang_data): ...new function.
11072 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
11073 gimple_opt_pass along with...
11074 (pass_data_tsan): ...new pass_data instance and...
11075 (make_pass_tsan): ...new function.
11076 (pass_tsan_O0): Convert from a global struct to a subclass of
11077 gimple_opt_pass along with...
11078 (pass_data_tsan_O0): ...new pass_data instance and...
11079 (make_pass_tsan_O0): ...new function.
11080 * var-tracking.c (pass_variable_tracking): Convert from a global
11081 struct to a subclass of rtl_opt_pass along with...
11082 (pass_data_variable_tracking): ...new pass_data instance and...
11083 (make_pass_variable_tracking): ...new function.
11084 * web.c (pass_web): Convert from a global struct to a subclass of
11085 rtl_opt_pass along with...
11086 (pass_data_web): ...new pass_data instance and...
11087 (make_pass_web): ...new function.
11088 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
11089 declaration with that of...
11090 (make_pass_mode_switch_use): ...new function.
11091 (pass_resolve_sw_modes): Replace declaration with that of...
11092 (make_pass_resolve_sw_modes): ...new function.
11093 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
11094 from a global struct to a subclass of rtl_opt_pass along with...
11095 (pass_data_mode_switch_use): ...new pass_data instance and...
11096 (make_pass_mode_switch_use): ...new function.
11097 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
11098 from a global struct to a subclass of rtl_opt_pass along with...
11099 (pass_data_resolve_sw_modes): ...new pass_data instance and...
11100 (make_pass_resolve_sw_modes): ...new function.
11101 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
11102 struct to a subclass of rtl_opt_pass along with...
11103 (pass_data_insert_vzeroupper): ...new pass_data instance and...
11104 (make_pass_insert_vzeroupper): ...new function.
11105 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
11106 global struct to a subclass of rtl_opt_pass along with...
11107 (pass_data_work_around_errata): ...new pass_data instance and...
11108 (make_pass_work_around_errata): ...new function.
11109 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
11110 struct to a subclass of rtl_opt_pass along with...
11111 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
11112 (make_pass_mips_machine_reorg2): ...new function.
11114 2013-08-05 David Malcolm <dmalcolm@redhat.com>
11116 * passes.c (pass_manager::operator new): New.
11118 2013-08-05 David Malcolm <dmalcolm@redhat.com>
11120 Handwritten part of conversion of passes to C++ classes.
11122 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
11123 (toplev.o): Add dep on PASS_MANAGER_H.
11124 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
11125 of early local pases to reflect this moving from a global to a
11126 member of gcc::pass_manager.
11127 (cgraph_add_new_function): Likewise.
11128 * lto-cgraph.c (lto_output_node): Update for conversion of
11129 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
11130 * passes.c (opt_pass::clone): New.
11131 (opt_pass::gate): New.
11132 (opt_pass::execute): New.
11133 (opt_pass::opt_pass): New.
11134 (pass_manager::execute_early_local_passes): New.
11135 (pass_manager::execute_pass_mode_switching): new.
11136 (finish_optimization_passes): Convert to...
11137 (pass_manager::finish_optimization_passes): ...this.
11138 (finish_optimization_passes): Update for conversion of passes to
11140 (register_dump_files_1): Use has_gate since we cannot portably
11141 check a vtable entry against NULL.
11142 (dump_one_pass): Likewise.
11143 (ipa_write_summaries_2): Likewise.
11144 (ipa_write_optimization_summaries_1): Likewise.
11145 (ipa_read_summaries_1): Likewise.
11146 (ipa_read_optimization_summaries_1): Likewise.
11147 (execute_ipa_stmt_fixups): Likewise.
11148 (pass_manager::pass_manager): Rewrite pass-creation, invoking
11149 pass-creation functions rather than wiring up globals, and
11150 storing the results in fields of pass_manager generated using
11151 pass-instances.def.
11152 (pass_manager::dump_profile_report): Update for conversion of
11153 passes to C++ classes.
11154 (pass_manager::execute_ipa_summary_passes): Likewise.
11155 (execute_one_ipa_transform_pass): Likewise.
11156 (execute_one_pass): Use has_gate and has_execute since we cannot
11157 portably check a vtable entry against NULL.
11158 * pass_manager.h (pass_manager::finish_optimization_passes): New.
11159 (pass_manager): Use pass-instances.def to add fields for the
11160 various pass instances.
11161 * toplev.c (finalize): Update for move of
11162 finish_optimization_passes to a method of gcc::pass_manager.
11163 * toplev.h (finish_optimization_passes): Move to method of class
11165 * tree-pass.h (struct pass_data): New.
11166 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
11167 (opt_pass::gate): Convert to virtual function.
11168 (opt_pass::~opt_pass): New.
11169 (opt_pass::clone): New.
11170 (opt_pass::execute): Convert to virtual function.
11171 (opt_pass::opt_pass): New.
11172 (opt_pass::ctxt_): new.
11173 (gimple_opt_pass): Convert to subclass of opt_pass.
11174 (gimple_opt_pass::gimple_opt_pass): New.
11175 (rtl_opt_pass): Convert to subclass of opt_pass.
11176 (rtl_opt_pass::rtl_opt_pass): New.
11177 (ipa_opt_pass_d): Convert to subclass of opt_pass.
11178 (ipa_opt_pass_d::ipa_opt_pass_d): New.
11179 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
11180 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
11181 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
11182 invocation of pass_mode_switching to reflect this moving from a
11183 global to a member of gcc::pass_manager.
11184 (ix86_option_override): Rework how pass_insert_vzeroupper is
11185 added to the pass_manager to reflect autogenerated changes.
11186 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
11188 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
11190 PR rtl-optimization/57708
11191 * recog.c (peep2_find_free_register): Validate all regs in a
11194 2013-08-05 Jan Hubicka <jh@suse.cz>
11197 * cgraph.c (verify_cgraph_node): Accept local flags from other
11199 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
11200 (function_and_variable_visibility): Likewise.
11201 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
11203 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11205 * graph.c (init_graph_slim_pretty_print): Remove.
11206 (print_graph_cfg): Do not call it. Use local pretty printer.
11207 (start_graph_dump): Likewise.
11209 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11211 * gimple-pretty-print.c (buffer): Remove.
11212 (initialized): Likewise.
11213 (maybe_init_pretty_print): Likewise.
11214 (print_gimple_stmt): Do not call it. Use non-static local
11215 pretty_printer variable.
11216 (print_gimple_expr): Likewise.
11217 (print_gimple_seq): Likewise.
11218 (gimple_dump_bb): Likewise.
11220 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11222 * asan.c (asan_pp): Remove.
11223 (asan_pp_initialized): Likewise.
11224 (asan_pp_initialize): Likewise.
11225 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
11226 (asan_emit_stack_protection): Tidy. Use local pretty printer.
11227 (asan_add_global): Likewise.
11229 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
11231 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
11232 * diagnostic.h (diagnostic_flush_buffer): Adjust.
11233 * pretty-print.c (pp_formatted_text_data): Likewise.
11234 (pp_indent): Rename from pp_base_indent.
11235 (pp_format): Rename from pp_base_format.
11236 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
11237 (pp_format_verbatim): Rename from pp_base_format_verbatim.
11238 (pp_flush): Rename from pp_base_flush.
11239 (pp_set_line_maximum_length): Rename from
11240 pp_base_set_line_maximum_length.
11241 (pp_clear_output_area): Rename from pp_base_clear_output_area.
11242 (pp_set_prefix): Rename from pp_base_set_prefix.
11243 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
11244 (pp_emit_prefix): Rename from pp_base_emit_prefix.
11245 (pp_append_text): Rename from pp_base_append_text.
11246 (pp_formatted_text): Rename from pp_base_formatted_text.
11247 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
11248 (pp_remaining_character_count_for_line): Rename from
11249 pp_base_remaining_character_count_for_line.
11250 (pp_newline): Rename from pp_base_newline.
11251 (pp_character): Rename from pp_base_character.
11252 (pp_string): Rename from pp_base_string.
11253 (pp_maybe_space): Rename from pp_base_maybe_space.
11254 * asan.c (asan_pp_string): Adjust.
11255 (asan_emit_stack_protection): Likewise.
11256 (asan_add_global): Likewise.
11257 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
11258 * tree-mudflap.c (mf_varname_tree): Likewise.
11259 * tree-pretty-print.c (pp_tree_identifier): Rename from
11260 pp_base_tree_identifier.
11261 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
11262 Declare as function.
11264 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11266 * pretty-print.h (pp_bar_bar): New.
11267 (pp_ampersand_ampersand): Likewise.
11268 (pp_less_equal): Likewise.
11269 (pp_greater_equal): Likewise.
11270 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
11271 printer functions instead of pp_string or operators and punctuators.
11272 (dump_gimple_call): Likewise.
11273 (dump_gimple_omp_for): Likewise.
11274 (dump_gimple_transaction): Likewise.
11275 (dump_gimple_phi): Likewise.
11276 (pp_gimple_stmt_1): Likewise.
11277 * sched-vis.c (print_insn): Likewise.
11278 * tree-mudflap.c (mf_varname_tree): Likewise.
11279 * tree-pretty-print.c (dump_block_node): Likewise.
11280 (dump_generic_node): Likewise.
11282 2013-08-02 Jan Hubicka <jh@suse.cz>
11284 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
11286 * lto-streamer-out.c (tree_is_indexable): Results decls and
11287 parm decls are not indexable.
11288 (DFS_write_tree_body): Do not follow args and results.
11289 (hash_tree): Likewise.
11290 (output_functions): Rearrange so struct function is needed
11291 only when real body is output; be able to also ouptut abstract
11292 functions; output DECL_ARGUMENTS and DECL_RESULT.
11293 (lto_output): When not in WPA, ale store abstract functions.
11294 (write_symbol): Do not care about RESULT_DECL.
11295 (output_symbol_p): Handle correctly sbtract decls.
11296 * lto-streamer-in.c (input_function): Rearrange so struct
11297 function can be NULL at entry; allow streaming of
11298 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
11299 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
11300 sanity check during LTO.
11301 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
11302 RESULT_DECl and DECL_ARGUMENTS.
11303 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
11306 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11308 * pretty-print.h (pp_underscore): New.
11310 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
11311 printer functions instead of pp_character.
11312 (dump_binary_rhs): Likewise.
11313 (dump_ternary_rhs): Likewise.
11314 (dump_gimple_call_args): Likewise.
11315 (pp_points_to_solution): Likewise.
11316 (dump_gimple_call): Likewise.
11317 (dump_gimple_switch): Likewise.
11318 (dump_gimple_cond): Likewise.
11319 (dump_gimple_bind): Likewise.
11320 (dump_gimple_try): Likewise.
11321 (dump_gimple_omp_for): Likewise.
11322 (dump_gimple_omp_continue): Likewise.
11323 (dump_gimple_omp_single): Likewise.
11324 (dump_gimple_omp_sections): Likewise.
11325 (dump_gimple_omp_block): Likewise.
11326 (dump_gimple_omp_critical): Likewise.
11327 (dump_gimple_transaction): Likewise.
11328 (dump_gimple_asm): Likewise.
11329 (dump_gimple_phi): Likewise.
11330 (dump_gimple_omp_parallel): Likewise.
11331 (dump_gimple_omp_task): Likewise.
11332 (dump_gimple_omp_atomic_load): Likewise.
11333 (dump_gimple_omp_atomic_store): Likewise.
11334 (dump_gimple_mem_ops): Likewise.
11335 (pp_gimple_stmt_1): Likewise.
11336 (pp_cfg_jump): Likewise.
11337 (dump_implicit_edges): Likewise.
11338 (gimple_dump_bb_for_graph): Likewise.
11339 * graph.c (draw_cfg_node): Likewise.
11340 * langhooks.c (lhd_print_error_function): Likewise.
11341 * sched-vis.c (print_exp): Likewise.
11342 (print_value): Likewise.
11343 (print_pattern): Likewise.
11344 (print_insn): Likewise.
11345 (rtl_dump_bb_for_graph): Likewise.
11346 * tree-pretty-print.c (dump_function_declaration): Likewise.
11347 (dump_array_domain): Likewise.
11348 (dump_omp_clause): Likewise.
11349 (dump_location): Likewise.
11350 (dump_generic_node): Likewise.
11351 (print_struct_decl): Likewise.
11352 * diagnostic.c (diagnostic_show_locus): Use pp_space.
11354 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11356 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
11357 candidate table when replacing a candidate statement.
11358 (replace_rhs_if_not_dup): Likewise.
11359 (replace_one_candidate): Likewise.
11361 2013-08-02 Jan Hubicka <jh@suse.cz>
11362 Martin Liska <marxin.liska@gmail.com>
11364 * cgraphunit.c (add_new_function): Fix logic when adding from
11366 (assemble_thunk): Rename to ...
11367 (expand_thunk); .. this one; export; get it working with
11368 general functions; make produced gimple valid.
11369 * cgraph.h (expand_thunk): Declare.
11371 2013-08-02 Jan Hubicka <jh@suse.cz>
11373 * ipa-cp.c (gather_context_independent_values): Use
11374 ipa_get_param_move_cost.
11375 (get_replacement_map): Remove PARAM; move parameter folding
11377 (create_specialized_node): Update.
11378 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
11379 assert that we have gimple body; update move_cost.
11380 (count_formal_params): Assert that we have gimple body.
11381 (ipa_dump_param): New function.
11382 (ipa_alloc_node_params): Break out from ...
11383 (ipa_initialize_node_params): ... here.
11384 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
11385 (ipa_write_node_info): Stream move costs.
11386 (ipa_read_node_info): Read move costs.
11387 (ipa_update_after_lto_read): Do not recompute node params.
11388 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
11389 (ipa_get_param): Check we are not in WPA.
11390 (ipa_get_param_move_cost): New.
11391 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
11392 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
11393 parm numbers to be present.
11395 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11397 PR rtl-optimization/58048
11398 * lra-constraints.c (process_alt_operands): Don't check asm
11399 operand on register.
11401 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
11403 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
11404 the implied StoreLoad barrier for atomic operations if before.
11406 2013-08-02 Jan Hubicka <jh@suse.cz>
11407 Martin Liska <marxin.liska@gmail.com>
11409 * cgraph.c (cgraph_function_body_availability): Do not check
11411 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
11412 symtab_node_availability): Declare.
11413 * ipa.c (can_replace_by_local_alias): New.
11414 (function_and_variable_visibility): Use it.
11415 * symtab.c (symtab_for_node_and_aliases,
11416 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
11418 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11420 PR rtl-optimization/57963
11421 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
11422 (lra_constraints): Use them.
11424 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
11426 * config/arm/types.md (define_attr "type"): Add "load_acq"
11428 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
11430 (cortex_a53_store1): Likewise.
11432 2013-08-01 Jan Hubicka <jh@suse.cz>
11434 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
11435 partitions are not needed.
11437 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
11439 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
11440 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
11441 MAYBE_NON_Q_CLASS_P where appropriate.
11443 2013-08-01 Jan Hubicka <jh@suse.cz>
11445 * cgraph.h (release_function_body): Declare.
11446 * tree.c (free_lang_data_in_decl): Free, parameters and return values
11447 of unused delcarations.
11449 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11451 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
11452 RTL form when subtracting a constant.
11454 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11456 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
11457 Generate canonical plus rtx with negated immediate instead of minus
11459 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
11461 2013-08-01 Jan Hubicka <jh@suse.cz>
11463 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
11464 (cgraph_release_function_body): Likewise.
11465 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
11466 * cgraph.h (cgrpah_node): Rename abstract_and_needed
11467 to used_as_abstract_origin.
11468 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
11469 symbols used as abstract origins.
11470 * cgraphunit.c (analyze_functions): Update.
11471 * ipa.c (symtab_remove_unreachable_nodes): Recompute
11472 used_as_abstract_origin.
11473 * tree-inline.c (tree_function_versioning): Update
11474 used_as_abstract_origin; be ready for DECL_RESULT and
11475 DECL_ARGUMENTS to be NULL.
11477 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
11478 for abstract functions.
11479 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
11482 2013-08-01 Jan Hubicka <jh@suse.cz>
11484 * profile.c (compute_value_histograms): Fix thinko.
11486 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11488 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
11489 aarch-common-protos.h to extra_headers.
11490 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11491 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
11492 * config/aarch64/t-aarch64 (aarch-common.o): Define.
11494 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11496 * config/aarch64/aarch64.md (define_attr "type"): Delete.
11497 Include "../arm/types.md". Define "type" attribute for all patterns.
11498 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
11501 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
11503 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
11504 to support power8 load fusion.
11505 (fusion_gpr_mem_load): Likewise.
11507 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
11509 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
11510 declarations for power8 load fusion.
11511 (emit_fusion_gpr_load): Likewise.
11513 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
11514 tuning for power8, turn on fusion mode by default. Turn on sign
11515 extending fusion mode if normal fusion mode is on, and we are at
11517 (fusion_gpr_load_p): New function, return true if we can fuse an
11518 addis instruction with a dependent load to a GPR.
11519 (emit_fusion_gpr_load): Emit the instructions for power8 load
11522 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
11523 (VSX load fusion peepholes): New peepholes to fuse together an
11524 addi instruction with a VSX load instruction.
11526 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
11527 peepholes to fuse an addis instruction with a load to a GPR base
11528 register. If we are supporting sign extending fusions, convert
11529 sign extending loads to zero extending loads and add an explicit
11532 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11534 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
11535 aarch-common-protos.h to extra_headers.
11536 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11537 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
11538 (arm_early_store_addr_dep): Likewise.
11539 (arm_no_early_alu_shift_dep): Likewise.
11540 (arm_no_early_alu_shift_value_dep): Likewise.
11541 (arm_no_early_mul_dep): Likewise.
11542 (arm_no_early_store_addr_dep): Likewise.
11543 (arm_mac_accumulator_is_mul_result): Likewise.
11544 (arm_mac_accumulator_is_result): Likewise.
11545 * config/arm/aarch-common.c: ... here. New file.
11546 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
11548 (arm_early_store_addr_dep): Likewise.
11549 (arm_no_early_alu_shift_dep): Likewise.
11550 (arm_no_early_alu_shift_value_dep): Likewise.
11551 (arm_no_early_mul_dep): Likewise.
11552 (arm_no_early_store_addr_dep): Likewise.
11553 (arm_mac_accumulator_is_mul_result): Likewise.
11554 (arm_mac_accumulator_is_result): Likewise.
11555 * config/arm/aarch-common-protos.h: ... here. New file.
11556 * config/arm/t-arm (aarch-common.o): Define.
11558 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11560 * config/arm/arm.md: Include new file "types.md".
11561 (define_attr "type"): Move from here to ...
11562 (define_attr "mul32"): Likewise.
11563 (define_attr "mul64"): Likewise.
11564 * config/arm/types.md: ... here. New file.
11566 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
11568 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
11569 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
11571 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11573 * gen-pass-instances.awk: Fix offset of substr().
11575 2013-07-31 David Malcolm <dmalcolm@redhat.com>
11577 * Makefile.in (pass-instances.def): New.
11578 (passes.o): Replace dependency on passes.def with one on
11581 * gen-pass-instances.awk: New.
11583 * passes.c (pass_manager::pass_manager): Use pass-instances.def
11584 rather than passes.def, updating local definition of NEXT_PASS
11585 macro to add an extra NUM parameter (currently unused).
11587 2013-07-30 David Malcolm <dmalcolm@redhat.com>
11589 * Makefile.in (PASS_MANAGER_H): New.
11590 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
11591 (passes.o): Likewise.
11592 (statistics.o): Likewise.
11593 (cgraphunit.o): Likewise.
11594 (context.o): Depend on PASS_MANAGER_H.
11596 * pass_manager.h: New.
11598 * cgraphunit.c (cgraph_add_new_function): Update for moves
11599 of globals to fields of pass_manager.
11600 (analyze_function): Likewise.
11601 (expand_function): Likewise.
11602 (ipa_passes): Likewise.
11603 (compile): Likewise.
11605 * context.c (context::context): New.
11606 * context.h (context::context): New.
11607 (context::get_passes): New.
11608 (context::passes_): New.
11610 * lto-cgraph.c (input_node): Update for moves of globals to
11611 fields of pass_manager.
11613 * passes.c (all_passes): Remove, in favor of a field of the
11614 same name within the new class pass_manager.
11615 (all_small_ipa_passes): Likewise.
11616 (all_lowering_passes): Likewise.
11617 (all_regular_ipa_passes): Likewise.
11618 (all_late_ipa_passes): Likewise.
11619 (all_lto_gen_passes): Likewise.
11620 (passes_by_id): Likewise.
11621 (passes_by_id_size): Likewise.
11622 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11623 the new class pass_manager.
11624 (set_pass_for_id): Convert to...
11625 (pass_manager::set_pass_for_id): ...method.
11626 (get_pass_for_id): Convert to...
11627 (pass_manager::get_pass_for_id): ...method.
11628 (register_one_dump_file): Move body of implementation into...
11629 (pass_manager::register_one_dump_file): ...here.
11630 (register_dump_files_1): Convert to...
11631 (pass_manager::register_dump_files_1): ...method.
11632 (register_dump_files): Convert to...
11633 (pass_manager::register_dump_files): ...method.
11634 (create_pass_tab): Update for moves of globals to fields of
11636 (dump_passes): Move body of implementation into...
11637 (pass_manager::dump_passes): ...here.
11638 (register_pass): Move body of implementation into...
11639 (pass_manager::register_pass): ...here.
11640 (init_optimization_passes): Convert into...
11641 (pass_manager::pass_manager): ...constructor for new
11642 pass_manager class, and initialize the pass_lists array.
11643 (check_profile_consistency): Update for moves of globals to
11644 fields of pass_manager.
11645 (dump_profile_report): Move body of implementation into...
11646 (pass_manager::dump_profile_report): ...here.
11647 (ipa_write_summaries_1): Update for moves of pass lists from
11648 being globals to fields of pass_manager.
11649 (ipa_write_optimization_summaries): Likewise.
11650 (ipa_read_summaries): Likewise.
11651 (ipa_read_optimization_summaries): Likewise.
11652 (execute_all_ipa_stmt_fixups): Likewise.
11654 * statistics.c (statistics_fini): Update for moves of globals to
11655 fields of pass_manager.
11657 * toplev.c (general_init): Replace call to
11658 init_optimization_passes with construction of the pass_manager
11661 * tree-pass.h (all_passes): Remove, in favor of a field of the
11662 same name within the new class pass_manager.
11663 (all_small_ipa_passes): Likewise.
11664 (all_lowering_passes): Likewise.
11665 (all_regular_ipa_passes): Likewise.
11666 (all_lto_gen_passes): Likewise.
11667 (all_late_ipa_passes): Likewise.
11668 (passes_by_id): Likewise.
11669 (passes_by_id_size): Likewise.
11670 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11671 the new class pass_manager.
11672 (get_pass_for_id): Remove.
11674 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11676 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
11679 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11681 * arm.md (mulhi3): New expand pattern.
11683 2013-07-30 Jan Hubicka <jh@suse.cz>
11684 Martin Liska <marxin.liska@gmail.com>
11686 * profile.c (compute_value_histograms): Do not ICE when
11687 there is mismatch only on some counters.
11689 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11691 PR rtl-optimization/57637
11692 * function.c (move_insn_for_shrink_wrap): Also check the
11693 GEN set of the LIVE problem for the liveness analysis
11694 if it exists, otherwise give up.
11696 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11698 PR tree-optimization/57993
11699 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
11700 replaced statement in the candidate table.
11701 (phi_add_costs): Return infinite cost when the hidden basis does
11702 not dominate all phis on which the candidate is dependent.
11703 (replace_one_candidate): Record replaced statement in the
11706 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
11708 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
11709 (ashlv2si3): New expander.
11710 (*ashlv2si3_i): New define_insn_and_split.
11711 * predicates.md (float_operation): Allow patterns with three
11712 basic sub-patterns.
11714 PR rtl-optimization/58021
11715 * mode-switching.c (create_pre_exit): Always split off preceding
11716 insns if we are not at the basic block head.
11718 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
11720 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
11721 (UCLIBC_DYNAMIC_LINKER): New macro.
11722 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
11724 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
11725 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
11727 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
11728 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
11729 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
11730 for SF and DF modes. Use ieee_quad_format for TF mode.
11731 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
11732 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
11733 (mips_option_override): Handle `-mnan=legacy'.
11734 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
11735 `-mabs=2008' and `-mnan=2008'.
11736 (OPTION_DEFAULT_SPECS): Add "nan" default.
11737 (ASM_SPEC): Handle `-mnan='.
11738 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
11739 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
11740 comment accordingly.
11741 (neg<mode>2): Likewise.
11742 * config/mips/mips.opt (mabs, mnan): New options.
11743 * doc/install.texi (Configuration): Document `--with-nan=' option.
11744 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
11746 (MIPS Options): Document them.
11747 * config.gcc <mips*-*-*>: Handle `--with-nan='.
11748 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
11749 * configure: Regenerate.
11750 * config.in: Regenerate.
11752 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11754 * config/i386/i386.md (float post-reload splitters): Do not check
11755 for subregs of SSE registers.
11757 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11758 H.J. Lu <hongjiu.lu@intel.com>
11762 * config/i386/i386.md (post-reload splitter
11763 to avoid partial SSE reg dependency stalls): New pattern.
11765 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
11767 * config/s390/s390.md ("movcc"): Swap load and store instructions.
11769 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
11771 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11772 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
11774 2013-07-26 Cary Coutant <ccoutant@google.com>
11776 * dwarf2out.c (die_checksum_ordered): Don't include template
11777 instantiations in signature.
11778 (is_template_parameter): New function.
11779 (is_template_instantiation): New function.
11780 (generate_skeleton_bottom_up): Don't include template instantiations
11782 (generate_skeleton): Likewise.
11783 (break_out_comdat_types): Move recursive call to break out nested
11785 (prune_unused_types_mark_generic_parms_dies): Call
11786 is_template_parameter.
11788 2013-07-26 Ian Bolton <ian.bolton@arm.com>
11790 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
11791 uses vector registers.
11792 * config/aarch64/iterators.md: Add attributes rtn and vas.
11794 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11795 Richard Earnshaw <richard.earnshaw@arm.com>
11797 * combine.c (simplify_comparison): Re-canonicalize operands
11799 * config/arm/arm.md (movcond_addsi): New splitter.
11801 2013-07-25 Sterling Augustine <saugustine@google.com>
11803 * dwarf2out.c (size_of_pubnames): Move code to...
11804 (include_pubname_in_output): ...here. New.
11805 (want_pubnames): Rearrange.
11806 (output_pubnames): Call include_pubname_in_output. Move assertion.
11808 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11810 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
11812 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11815 * doc/extend.texi: Remove obsolete builtins. Fix
11816 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
11818 2013-07-25 Jan Hubicka <jh@suse.cz>
11820 * cgraph.c (release_function_body): Break out from ...
11821 (cgraph_release_function_body): ... this one; also release DECL_RESULT
11822 and DECL_ARGUMENTS.
11823 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
11824 old_tree in the map.
11825 (create_specialized_node): Update.
11826 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
11828 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
11829 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
11830 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
11831 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
11834 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11836 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
11837 addsi3_carryin_alt2_<optab>): Correct output template.
11839 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11841 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
11842 Adjust for arm_restrict_it.
11843 Remove trailing whitespace.
11845 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
11847 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
11848 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
11850 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
11852 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
11854 PR rtl-optimization/57960
11855 * lra-constraints.c (process_alt_operands): Use the right mode
11856 when checking strict_low.
11858 2013-07-25 Jan Hubicka <jh@suse.cz>
11860 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
11861 * cgraph.c (cgraph_remove_node): Do not release function body
11862 when in cgraph streaming.
11863 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
11864 in other partitions are not considered reachable; fix handling of
11867 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11869 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
11871 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11876 * config/arm/arm.md ("*sibcall_insn): Replace use of
11877 Ss with US. Adjust output for v5 and v4t.
11878 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
11880 * config/arm/constraints.md ("Ss"): Rename to US.
11882 2013-07-25 Terry Guo <terry.guo@arm.com>
11884 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
11885 shift_add/shift_sub0/shift_sub1 RTXs.
11887 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
11888 Anton Blanchard <anton@au1.ibm.com>
11890 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
11891 (altivec_vpks<VI_char>ss): Likewise.
11892 (altivec_vpks<VI_char>us): Likewise.
11893 (altivec_vpku<VI_char>us): Likewise.
11894 (altivec_vpku<VI_char>um): Likewise.
11896 2013-07-24 David Malcolm <dmalcolm@redhat.com>
11898 Introduce context class.
11900 * Makefile.in (CONTEXT_H): New.
11901 (OBJS): Add context.o.
11902 (toplev.o): Add CONTEXT_H to dependencies.
11905 * toplev.c (general_init): Create the singleton gcc::context instance.
11911 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
11913 PR rtl-optimization/57968
11914 * mode-switching.c (create_pre_exit): Allow instructions that
11915 don't set a return register to need a non-exit mode.
11917 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11918 Anton Blanchard <anton@au1.ibm.com>
11920 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
11921 operands to vperm for little endian.
11922 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
11923 of lvsl to create the control mask for a vperm for little endian.
11925 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11926 Anton Blanchard <anton@au1.ibm.com>
11928 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
11929 two operands for little-endian.
11931 2013-07-23 Steve Ellcey <sellcey@mips.com>
11933 * config/mips/mips.c (mips_case_values_threshold): New.
11934 (TARGET_CASE_VALUES_THRESHOLD): Define.
11936 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11937 Anton Blanchard <anton@au1.ibm.com>
11939 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
11940 selection of field for vector splat in little endian mode.
11942 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
11944 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
11945 expanders to rs6000.md.
11946 (ior<mode>3): Likewise.
11947 (and<mode>3): Likewise.
11948 (one_cmpl<mode>2): Likewise.
11949 (nor<mode>3): Likewise.
11950 (andc<mode>3): Likewise.
11951 (eqv<mode>3): Likewise.
11952 (nand<mode>3): Likewise.
11953 (orc<mode>3): Likewise.
11955 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
11958 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
11959 to split multi-word logical operations.
11960 (rs6000_split_logical_di): Likewise.
11961 (rs6000_split_logical): Likewise.
11963 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
11964 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
11965 and allow TImode operations in 32-bit.
11966 (vsx_and<mode>3_64bit): Likewise.
11967 (vsx_ior<mode>3_32bit): Likewise.
11968 (vsx_ior<mode>3_64bit): Likewise.
11969 (vsx_xor<mode>3_32bit): Likewise.
11970 (vsx_xor<mode>3_64bit): Likewise.
11971 (vsx_one_cmpl<mode>2_32bit): Likewise.
11972 (vsx_one_cmpl<mode>2_64bit): Likewise.
11973 (vsx_nor<mode>3_32bit): Likewise.
11974 (vsx_nor<mode>3_64bit): Likewise.
11975 (vsx_andc<mode>3_32bit): Likewise.
11976 (vsx_andc<mode>3_64bit): Likewise.
11977 (vsx_eqv<mode>3_32bit): Likewise.
11978 (vsx_eqv<mode>3_64bit): Likewise.
11979 (vsx_nand<mode>3_32bit): Likewise.
11980 (vsx_nand<mode>3_64bit): Likewise.
11981 (vsx_orc<mode>3_32bit): Likewise.
11982 (vsx_orc<mode>3_64bit): Likewise.
11984 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
11985 logical types in GPRs.
11987 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
11988 logical insns to rs6000.md, and allow TImode operations in
11990 (altivec_ior<mode>3): Likewise.
11991 (altivec_xor<mode>3): Likewise.
11992 (altivec_one_cmpl<mode>2): Likewise.
11993 (altivec_nor<mode>3): Likewise.
11994 (altivec_andc<mode>3): Likewise.
11996 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
11997 attributes for moving the 128-bit logical operations into
11999 (BOOL_REGS_OUTPUT): Likewise.
12000 (BOOL_REGS_OP1): Likewise.
12001 (BOOL_REGS_OP2): Likewise.
12002 (BOOL_REGS_UNARY): Likewise.
12003 (BOOL_REGS_AND_CR0): Likewise.
12004 (one_cmpl<mode>2): Add support for DI logical operations on
12005 32-bit, splitting the operations to 32-bit.
12006 (anddi3): Likewise.
12007 (iordi3): Likewise.
12008 (xordi3): Likewise.
12009 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
12010 changes to combine the 32/64-bit code, allow logical operations on
12011 TI mode in 32-bit, and to use similar match_operator patterns like
12012 scalar mode uses. Combine the Altivec and VSX code for logical
12013 operations, and move it here.
12014 (ior<mode>3, 128-bit types): Likewise.
12015 (xor<mode>3, 128-bit types): Likewise.
12016 (one_cmpl<mode>3, 128-bit types): Likewise.
12017 (nor<mode>3, 128-bit types): Likewise.
12018 (andc<mode>3, 128-bit types): Likewise.
12019 (eqv<mode>3, 128-bit types): Likewise.
12020 (nand<mode>3, 128-bit types): Likewise.
12021 (orc<mode>3, 128-bit types): Likewise.
12022 (and<mode>3_internal): Likewise.
12023 (bool<mode>3_internal): Likewise.
12024 (boolc<mode>3_internal1): Likewise.
12025 (boolc<mode>3_internal2): Likewise.
12026 (boolcc<mode>3_internal1): Likewise.
12027 (boolcc<mode>3_internal2): Likewise.
12028 (eqv<mode>3_internal1): Likewise.
12029 (eqv<mode>3_internal2): Likewise.
12030 (one_cmpl1<mode>3_internal): Likewise.
12032 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
12034 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12035 Rename flag_stack_usage to flag_stack_usage_info.
12037 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
12039 * config/microblaze/sync.md: New file.
12040 * config/microblaze/microblaze.md: Include sync.md
12041 * config/microblaze/microblaze.c: Add print_operand 'y'.
12042 * config/microblaze/constraints.md: Add memory_contraint
12043 'Q' which is a single register.
12045 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
12047 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
12049 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
12051 * reload.c (find_reloads): Exit loop once we find this operand
12052 cannot be reloaded somehow for this alternative.
12054 * reload.c (find_reloads): Exit loop once we find a hard register.
12056 * rtlanal.c (computed_jump_p): Exit loop once we find label
12059 * i386.c (ix86_pad_returns): Exit loop after setting replace.
12061 * cfgloopmanip.c (remove_path): Exit loop after setting
12064 * gensupport.c (subst_dup): Avoid loop if code is not
12065 MATCH_DUP nor MATCH_OP_DUP.
12067 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
12069 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
12071 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12073 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
12074 true for SP_REGNUM if mode == ptr_mode.
12075 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
12076 with value R0_REGNUM + 31.
12078 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12080 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
12081 pad pointer-typed argument downward.
12083 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12085 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
12086 and __ILP32__ when the ILP32 model is in use.
12088 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12090 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
12091 (aarch64_load_symref_appropriately): In the case of
12092 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
12093 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
12094 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
12095 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
12096 if 'mode' doesn't equal to 'ptr_mode'.
12097 (aarch64_output_mi_thunk): Add an assertion on the alignment of
12098 'vcall_offset'; change to call aarch64_emit_move differently depending
12099 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
12100 to calculate the upper bound of 'vcall_offset'.
12101 (aarch64_cannot_force_const_mem): Change to also return true if
12103 (aarch64_legitimize_reload_address): In the case of large
12104 displacements, add new local variable 'xmode' and an assertion
12105 based on it; change to use 'xmode' to generate the new rtx and
12107 (aarch64_asm_trampoline_template): Change to generate the template
12108 differently depending on TARGET_ILP32 or not; change to use
12109 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
12110 (aarch64_trampoline_size): Removed.
12111 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
12112 and replace immediate literals with it. Change to use 'ptr_mode'
12113 instead of 'DImode' and call convert_memory_address if the mode
12114 of 'fnaddr' doesn't equal to 'ptr_mode'.
12115 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
12117 (aarch64_elf_asm_destructor): Likewise.
12118 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
12119 on TARGET_ILP32 instead of aarch64_trampoline_size.
12120 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
12121 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
12122 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
12123 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
12124 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
12125 (storewb_pair<GPI:mode>_<P:mode>): ... this.
12126 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
12127 depending on the value of 'mode'.
12128 (add_losym_<mode>): New.
12129 (ldr_got_small_<mode>): New, based on ldr_got_small.
12130 (ldr_got_small): Remove.
12131 (ldr_got_small_sidi): New.
12132 * config/aarch64/iterators.md (P): New.
12133 (PTR): Change to 'ptr_mode' in the condition.
12135 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12137 * config.gcc (aarch64*-*-*): Support --with-abi.
12138 (aarch64*-*-elf): Support --with-multilib-list.
12139 (aarch64*-*-linux*): Likewise.
12140 (supported_defaults): Add abi to aarch64*-*-*.
12141 * configure.ac: Mention AArch64 for --with-multilib-list.
12142 * configure: Re-generated.
12143 * config/aarch64/biarchilp32.h: New file.
12144 * config/aarch64/biarchlp64.h: New file.
12145 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
12147 (MULTILIB_DEFAULTS): Ditto.
12148 (DRIVER_SELF_SPECS): Ditto.
12149 (ASM_SPEC): Update to also substitute -mabi.
12150 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
12151 file whose name depends on -mabi= and -mbig-endian.
12152 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
12154 (POINTER_SIZE): New define.
12155 (POINTERS_EXTEND_UNSIGNED): Ditto.
12156 (enum aarch64_abi_type): New enumeration tag.
12157 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
12158 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
12159 (TARGET_ILP32): New define.
12160 * config/aarch64/aarch64.opt (mabi): New.
12161 (aarch64_abi): New.
12162 (ilp32, lp64): New values for -mabi.
12163 * config/aarch64/t-aarch64 (comma): New define.
12164 (MULTILIB_OPTIONS): Ditto.
12165 (MULTILIB_DIRNAMES): Ditto.
12166 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
12167 * doc/invoke.texi: Document -mabi for AArch64.
12169 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
12171 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
12173 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12174 Anton Blanchard <anton@au1.ibm.com>
12176 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
12177 endianness when selecting field to splat.
12179 2013-07-22 Eric Christopher <echristo@gmail.com>
12181 * dwarf2out.c (die_odr_checksum): New function to use
12182 CHECKSUM_ macros and ULEB128 for DIE tag.
12183 (generate_type_signature): Use.
12185 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
12187 * config.gcc (sparc*-*-*): Accept leon3 processor.
12188 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
12189 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
12190 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
12191 * config/sparc/sparc.opt (enum processor_type): Add leon3.
12192 (mfix-ut699): Adjust comment.
12193 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
12194 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
12195 (CPP_CPU_SPEC): Likewise.
12196 (ASM_CPU_SPEC): Likewise.
12197 * config/sparc/sparc.c (leon3_cost): New constant.
12198 (sparc_option_override): Add leon3 support.
12199 (mem_ref): New function.
12200 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
12201 (sparc_do_work_around_errata): Look into the instruction in the delay
12202 slot and adjust accordingly. Add fix for the data cache nullify issues
12203 of the UT699. Change insertion position for the NOP.
12204 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
12205 (leon3_load): New reservation.
12206 (leon_store): Bump latency to 2.
12207 (grfpu): New automaton.
12208 (grfpu_alu): New unit.
12209 (grfpu_ds): Likewise.
12210 (leon_fp_alu): Adjust.
12211 (leon_fp_mult): Delete.
12212 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
12213 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
12214 * config/sparc/sparc.md (cpu): Add leon3.
12215 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
12216 (swapsi): Likewise.
12217 (atomic_test_and_set): Likewise.
12218 (ldstub): Likewise.
12220 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
12222 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
12223 default for R5900 targets.
12224 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
12225 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
12226 * config/mips/mips.c (mips_option_override): Report an error for
12227 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
12228 for -march=r5900 -mhard-float.
12230 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
12232 * df-problems.c (can_move_insns_across): Exit loop once we
12233 find a non-fixed, non-global register.
12235 * ipa-pure-const.c (propagate_nothrow): Exit loop after
12238 * omega.c (omega_eliminate_red): Break after setting red_found.
12239 (omega_problem_has_red_equations): Similarly after setting found.
12240 (omega_query_variable): Similarly after setting coupled.
12242 2013-07-22 Marek Polacek <polacek@redhat.com>
12244 * gimplify.c: Don't include gimple.h twice.
12246 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12248 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
12249 instead of TARGET_THUMB1.
12250 (Pz): New constraint.
12251 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
12253 (compare_negsi_si): Likewise.
12254 (compare_addsi2_op0): Likewise.
12255 (compare_addsi2_op1): Likewise.
12256 (addsi3_carryin_<optab>): Likewise.
12257 (addsi3_carryin_alt2_<optab>): Likewise.
12258 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
12259 for arm_restrict_it.
12260 (subsi3_carryin): Likewise.
12261 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
12262 (minmax_arithsi): Disable for arm_restrict_it.
12263 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
12264 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
12265 (satsi_<SAT:code>_shift): Likewise.
12266 (arm_shiftsi3): Add alternative for 16-bit encoding.
12267 (arm32_movhf): Disable for arm_restrict_it.
12268 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
12269 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
12271 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
12273 * config/arm/arm.md (attribute "insn"): Delete.
12274 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
12275 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
12276 (not_shiftsi): Update for attribute change.
12277 (not_shiftsi_compare0): Likewise.
12278 (not_shiftsi_compare0_scratch): Likewise.
12279 (arm_one_cmplsi2): Likewise.
12280 (thumb1_one_cmplsi2): Likewise.
12281 (notsi_compare0): Likewise.
12282 (notsi_compare0_scratch): Likewise.
12283 (thumb1_movdi_insn): Likewise.
12284 (arm_movsi_insn): Likewise.
12285 (movhi_insn_arch4): Likewise.
12286 (movhi_bytes): Likewise.
12287 (arm_movqi_insn): Likewise.
12288 (thumb1_movqi_insn): Likewise.
12289 (arm32_movhf): Likewise.
12290 (thumb1_movhf): Likewise.
12291 (arm_movsf_soft_insn): Likewise.
12292 (thumb1_movsf_insn): Likewise.
12293 (thumb_movdf_insn): Likewise.
12294 (movsicc_insn): Likewise.
12295 (movsfcc_soft_insn): Likewise.
12296 (and_scc): Likewise.
12297 (cond_move): Likewise.
12298 (if_move_not): Likewise.
12299 (if_not_move): Likewise.
12300 (if_shift_move): Likewise.
12301 (if_move_shift): Likewise.
12302 (if_shift_shift): Likewise.
12303 (if_not_arith): Likewise.
12304 (if_arith_not): Likewise.
12305 (cond_move_not): Likewise.
12306 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
12307 (neon_mov<mode>): Likewise.
12308 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
12309 (thumb2_movsi_vfp): Likewise.
12310 (movsf_vfp): Likewise.
12311 (thumb2_movsf_vfp): Likewise.
12312 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12314 (cortexa7_older_only): Likewise.
12315 (cortexa7_younger): Likewise.
12316 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
12317 (1020alu_shift_op): Likewise.
12318 (1020alu_shift_reg_op): Likewise.
12319 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
12320 (alu_shift_op): Likewise.
12321 (alu_shift_reg_op): Likewise.
12322 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
12323 (11_alu_shift_op): Likewise.
12324 (11_alu_shift_reg_op): Likewise.
12325 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
12326 (9_alu_shift_reg_op): Likewise.
12327 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12329 (cortex_a15_alu_shift): Likewise.
12330 (cortex_a15_alu_shift_reg): Likewise.
12331 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
12332 (cortex_a5_alu_shift): Likewise.
12333 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
12335 (cortex_a53_alu_shift): Likewise.
12336 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12338 (cortex_a7_alu_reg): Likewise.
12339 (cortex_a7_alu_shift): Likewise.
12340 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12341 (cortex_a8_alu_shift): Likewise.
12342 (cortex_a8_alu_shift_reg): Likewise.
12343 (cortex_a8_mov): Likewise.
12344 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
12345 (cortex_a9_dp_shift): Likewise.
12346 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
12347 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
12348 (cortex_r4_mov): Likewise.
12349 (cortex_r4_alu_shift): Likewise.
12350 (cortex_r4_alu_shift_reg): Likewise.
12351 * config/arm/fa526.md (526_alu_op): Update for attribute change.
12352 (526_alu_shift_op): Likewise.
12353 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
12354 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
12355 (626te_alu_shift_op): Likewise.
12356 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
12357 (726te_alu_op): Likewise.
12358 (726te_alu_shift_op): Likewise.
12359 (726te_alu_shift_reg_op): Likewise.
12360 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
12361 (mp626_alu_shift_op): Likewise.
12362 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
12363 (pj4_alu_e1_conds): Likewise.
12364 (pj4_alu): Likewise.
12365 (pj4_alu_conds): Likewise.
12366 (pj4_shift): Likewise.
12367 (pj4_shift_conds): Likewise.
12368 (pj4_alu_shift): Likewise.
12369 (pj4_alu_shift_conds): Likewise.
12371 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12373 * config/arm/predicates.md (shiftable_operator_strict_it):
12375 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
12376 Disable cond_exec version for arm_restrict_it.
12377 (thumb2_smaxsi3): Convert to generate cond_exec.
12378 (thumb2_sminsi3): Likewise.
12379 (thumb32_umaxsi3): Likewise.
12380 (thumb2_uminsi3): Likewise.
12381 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
12382 (thumb2_neg_abssi2): Likewise.
12383 (thumb2_mov_scc): Add alternative for 16-bit encoding.
12384 (thumb2_movsicc_insn): Adjust alternatives.
12385 (thumb2_mov_negscc): Disable for arm_restrict_it.
12386 (thumb2_mov_negscc_strict_it): New pattern.
12387 (thumb2_mov_notscc_strict_it): New pattern.
12388 (thumb2_mov_notscc): Disable for arm_restrict_it.
12389 (thumb2_ior_scc): Likewise.
12390 (thumb2_ior_scc_strict_it): New pattern.
12391 (thumb2_cond_move): Adjust for arm_restrict_it.
12392 (thumb2_cond_arith): Disable for arm_restrict_it.
12393 (thumb2_cond_arith_strict_it): New pattern.
12394 (thumb2_cond_sub): Adjust for arm_restrict_it.
12395 (thumb2_movcond): Likewise.
12396 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
12397 (thumb2_zero_extendhisi2_v6): Likewise.
12398 (thumb2_zero_extendqisi2_v6): Likewise.
12399 (orsi_notsi_si): Likewise.
12400 (orsi_not_shiftsi_si): Likewise.
12402 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
12404 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
12405 instruction sequence is 1 byte shorter.
12407 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
12409 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
12410 it is not needed after split.
12412 2013-07-20 Iain Sandoe <iain@codesourcery.com>
12415 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
12416 second label for nonlocal goto receivers. Don't output pic base labels
12417 unless we're producing PIC; mark that action unreachable().
12418 (ix86_save_reg): If the function contains a nonlocal label, save the
12420 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
12421 * config/darwin.c (emitted_pic_label_num): New GTY.
12422 (update_pic_label_number_if_needed): New.
12423 (machopic_output_function_base_name): Adjust for nonlocal receiver
12425 (machopic_should_output_picbase_label): New.
12426 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
12427 (nonlocal_goto_receiver): New insn and split.
12429 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
12431 * config/aarch64/aarch64-builtins.c
12432 (aarch64_fold_builtin): Fold abs in all modes.
12433 * config/aarch64/aarch64-simd-builtins.def
12434 (abs): Enable for all modes.
12435 * config/aarch64/arm_neon.h
12436 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
12437 (vabs_f64): Add missing intrinsic.
12439 2013-07-19 Ian Bolton <ian.bolton@arm.com>
12441 * config/aarch64/arm_neon.h (vabs_s64): New function
12443 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
12446 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
12447 * config/avr/avr.md (adjust_len): Add `round'.
12448 * config/avr/avr-protos.h (avr_out_round): New prototype.
12449 (avr_out_plus): Add `out_label' argument.
12450 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
12451 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
12452 Handle the case where `insn' is just a pattern.
12453 (avr_out_bitop): Handle the case where `insn' is just a pattern.
12454 (avr_out_round): New function.
12455 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
12457 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
12459 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12460 Add check for flag_stack_usage to handle -fstack-usage support
12462 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
12464 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
12465 interaction for new Power8 flags and VSX.
12467 2013-07-18 Sriraman Tallam <tmsriram@google.com>
12469 PR middle-end/57698
12470 * tree-inline.c (expand_call_inline): Emit errors during
12471 early_inlining only if optimization is not turned on.
12473 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12477 * passes.c (init_optimization_passes): Move the construction of
12478 the pass hierarchy into a new passes.def file.
12480 * Makefile.in (passes.o): Add dependency on passes.def.
12482 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12484 * passes.c (init_optimization_passes): Introduce macros for
12485 constructing the tree of passes (INSERT_PASSES_AFTER,
12486 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
12487 TERMINATE_PASS_LIST).
12489 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
12490 Wei Mi <wmi@google.com>
12492 PR rtl-optimization/57878
12493 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
12495 (reload_pseudo_compare_func): Check nregs first for reload
12498 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12500 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
12502 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
12504 * read-rtl.c (validate_const_int): Once an invalid character is
12505 seen, quit the loop.
12507 * gengtype.c (write_roots): Similarly once we find the "deletable"
12508 or "if_marked" option.
12510 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12512 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
12513 "xtab" and "sat". Move value "clz" from here to ...
12514 (attriubte "type"): ... here.
12515 (satsi_<SAT:code>): Delete "insn" attribute.
12516 (satsi_<SAT:code>_shift): Likewise.
12517 (arm_zero_extendqisi2addsi): Likewise.
12518 (arm_extendqisi2addsi): Likewise.
12519 (clzsi2): Update for attribute changes.
12520 (rbitsi2): Likewise.
12521 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
12523 (arm_usatsihi): Likewise.
12524 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12526 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12528 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
12529 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
12530 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
12531 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
12532 in alphabetical order.
12533 (attribute "core_cycles"): Update for attribute changes.
12534 (arm_addsi3): Likewise.
12535 (addsi3_compare0): Likewise.
12536 (addsi3_compare0_scratch): Likewise.
12537 (addsi3_compare_op1): Likewise.
12538 (addsi3_compare_op2): Likewise.
12539 (compare_addsi2_op0): Likewise.
12540 (compare_addsi2_op1): Likewise.
12541 (addsi3_carryin_shift_<optab>): Likewise.
12542 (subsi3_carryin_shift): Likewise.
12543 (rsbsi3_carryin_shift): Likewise.
12544 (arm_subsi3_insn): Likewise.
12545 (subsi3_compare0): Likewise.
12546 (subsi3_compare): Likewise.
12547 (arm_andsi3_insn): Likewise.
12548 (thumb1_andsi3_insn): Likewise.
12549 (andsi3_compare0): Likewise.
12550 (andsi3_compare0_scratch): Likewise.
12551 (zeroextractsi_compare0_scratch
12552 (andsi_not_shiftsi_si): Likewise.
12553 (iorsi3_insn): Likewise.
12554 (iorsi3_compare0): Likewise.
12555 (iorsi3_compare0_scratch): Likewise.
12556 (arm_xorsi3): Likewise.
12557 (thumb1_xorsi3_insn): Likewise.
12558 (xorsi3_compare0): Likewise.
12559 (xorsi3_compare0_scratch): Likewise.
12560 (satsi_<SAT:code>_shift): Likewise.
12562 (arm_shiftsi3): Likewise.
12563 (shiftsi3_compare0): Likewise.
12564 (not_shiftsi): Likewise.
12565 (not_shiftsi_compare0): Likewise.
12566 (not_shiftsi_compare0_scratch): Likewise.
12567 (arm_one_cmplsi2): Likewise.
12568 (thumb_one_complsi2): Likewise.
12569 (notsi_compare0): Likewise.
12570 (notsi_compare0_scratch): Likewise.
12571 (thumb1_zero_extendhisi2): Likewise.
12572 (arm_zero_extendhisi2): Likewise.
12573 (arm_zero_extendhisi2_v6): Likewise.
12574 (arm_zero_extendhisi2addsi): Likewise.
12575 (thumb1_zero_extendqisi2): Likewise.
12576 (thumb1_zero_extendqisi2_v6): Likewise.
12577 (arm_zero_extendqisi2): Likewise.
12578 (arm_zero_extendqisi2_v6): Likewise.
12579 (arm_zero_extendqisi2addsi): Likewise.
12580 (thumb1_extendhisi2): Likewise.
12581 (arm_extendhisi2): Likewise.
12582 (arm_extendhisi2_v6): Likewise.
12583 (arm_extendqisi): Likewise.
12584 (arm_extendqisi_v6): Likewise.
12585 (arm_extendqisi2addsi): Likewise.
12586 (thumb1_extendqisi2): Likewise.
12587 (thumb1_movdi_insn): Likewise.
12588 (arm_movsi_insn): Likewise.
12589 (movsi_compare0): Likewise.
12590 (movhi_insn_arch4): Likewise.
12591 (movhi_bytes): Likewise.
12592 (arm_movqi_insn): Likewise.
12593 (thumb1_movqi_insn): Likewise.
12594 (arm32_movhf): Likewise.
12595 (thumb1_movhf): Likewise.
12596 (arm_movsf_soft_insn): Likewise.
12597 (thumb1_movsf_insn): Likewise.
12598 (movdf_soft_insn): Likewise.
12599 (thumb_movdf_insn): Likewise.
12600 (arm_cmpsi_insn): Likewise.
12601 (cmpsi_shiftsi): Likewise.
12602 (cmpsi_shiftsi_swp): Likewise.
12603 (arm_cmpsi_negshiftsi_si): Likewise.
12604 (movsicc_insn): Likewise.
12605 (movsfcc_soft_insn): Likewise.
12606 (arith_shiftsi): Likewise.
12607 (arith_shiftsi_compare0
12608 (arith_shiftsi_compare0_scratch
12609 (sub_shiftsi): Likewise.
12610 (sub_shiftsi_compare0
12611 (sub_shiftsi_compare0_scratch
12612 (and_scc): Likewise.
12613 (cond_move): Likewise.
12614 (if_plus_move): Likewise.
12615 (if_move_plus): Likewise.
12616 (if_move_not): Likewise.
12617 (if_not_move): Likewise.
12618 (if_shift_move): Likewise.
12619 (if_move_shift): Likewise.
12620 (if_shift_shift): Likewise.
12621 (if_not_arith): Likewise.
12622 (if_arith_not): Likewise.
12623 (cond_move_not): Likewise.
12624 (thumb1_ashlsi3): Set type attribute.
12625 (thumb1_ashrsi3): Likewise.
12626 (thumb1_lshrsi3): Likewise.
12627 (thumb1_rotrsi3): Likewise.
12628 (shiftsi3_compare0_scratch): Likewise.
12629 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
12630 (neon_mov<mode>): Likewise.
12631 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
12633 (thumb2_movsi_insn): Likewise.
12634 (thumb2_cmpsi_neg_shiftsi): Likewise.
12635 (thumb2_extendqisi_v6): Likewise.
12636 (thumb2_zero_extendhisi2_v6): Likewise.
12637 (thumb2_zero_extendqisi2_v6): Likewise.
12638 (thumb2_shiftsi3_short): Likewise.
12639 (thumb2_addsi3_compare0_scratch): Likewise.
12640 (orsi_not_shiftsi_si): Likewise.
12641 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
12642 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
12644 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
12645 (1020alu_shift_op): Likewise.
12646 (1020alu_shift_reg_op): Likewise.
12647 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
12648 (alu_shift_op): Likewise.
12649 (alu_shift_reg_op): Likewise.
12650 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
12651 (11_alu_shift_op): Likewise.
12652 (11_alu_shift_reg_op): Likewise.
12653 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
12654 (9_alu_shift_reg_op): Likewise.
12655 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12657 (cortex_a15_alu_shift): Likewise.
12658 (cortex_a15_alu_shift_reg): Likewise.
12659 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
12661 (cortex_a5_alu_shift): Likewise.
12662 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
12664 (cortex_a53_alu_shift): Likewise.
12665 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12667 (cortex_a7_alu_reg): Likewise.
12668 (cortex_a7_alu_shift): Likewise.
12669 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
12671 (cortex_a8_alu_shift): Likewise.
12672 (cortex_a8_alu_shift_reg): Likewise.
12673 (cortex_a8_mov): Likewise.
12674 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
12675 (cortex_a9_dp_shift): Likewise.
12676 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
12678 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
12680 (cortex_r4_mov): Likewise.
12681 (cortex_r4_alu_shift): Likewise.
12682 (cortex_r4_alu_shift_reg): Likewise.
12683 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
12684 (526_alu_shift_op): Likewise.
12685 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
12686 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
12687 (626te_alu_shift_op): Likewise.
12688 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
12689 (726te_alu_op): Likewise.
12690 (726te_alu_shift_op): Likewise.
12691 (726te_alu_shift_reg_op): Likewise.
12692 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
12693 (mp626_alu_shift_op): Likewise.
12694 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
12695 (pj4_alu_e1_conds): Likewise.
12696 (pj4_alu): Likewise.
12697 (pj4_alu_conds): Likewise.
12698 (pj4_shift): Likewise.
12699 (pj4_shift_conds): Likewise.
12700 (pj4_alu_shift): Likewise.
12701 (pj4_alu_shift_conds): Likewise.
12702 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12704 (cortexa7_older_only): Likewise.
12705 (cortexa7_younger): Likewise.
12707 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12709 * ipa-pure-const.c (generate_summary): Rename to...
12710 (pure_const_generate_summary): ... this.
12712 2013-07-17 Iain Sandoe <iain@codesourcery.com>
12714 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
12716 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
12719 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
12722 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12724 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
12725 enabled without -march=zEC12.
12726 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
12729 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12731 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
12732 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
12733 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
12734 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
12735 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
12736 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
12737 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
12739 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
12740 and ISA_HAS_NMADD3_NMSUB3.
12741 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
12742 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
12743 (nmsub4<mode>, nmsub3<mode>): Likewise.
12744 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
12746 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12748 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
12749 TARGET_MIPS5400 checking.
12751 2013-07-16 Jakub Jelinek <jakub@redhat.com>
12752 Peter Bergner <bergner@vnet.ibm.com>
12754 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
12755 registers in the comment.
12756 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
12757 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
12758 rather than FIRST_PSEUDO_REGISTERS.
12760 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
12762 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
12763 enable extra ISA flags with TARGET_HTM.
12765 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12767 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
12770 2013-07-15 Cong Hou <congh@google.com>
12772 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
12773 in compare function for sorting.
12775 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
12777 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
12778 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
12779 * config/rs6000/rs6000.opt: Add -mhtm option.
12780 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
12781 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
12782 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12783 __HTM__ if the HTM instructions are available.
12784 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
12785 htm_spr_reg_operand): New define_predicates.
12786 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
12787 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
12789 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
12790 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
12791 HTM builtin functions.
12792 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
12793 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
12794 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
12795 (rs6000_builtin_mask_calculate): Likewise.
12796 (rs6000_option_override_internal): Likewise.
12797 (bdesc_htm): Add new HTM builtin support.
12798 (htm_spr_num): New function.
12799 (htm_spr_regno): Likewise.
12800 (rs6000_htm_spr_icode): Likewise.
12801 (htm_expand_builtin): Likewise.
12802 (htm_init_builtins): Likewise.
12803 (rs6000_expand_builtin): Add support for HTM builtin functions.
12804 (rs6000_init_builtins): Likewise.
12805 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
12807 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
12808 (TARGET_HTM, MASK_HTM): Define macros.
12809 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
12810 (FIXED_REGISTERS): Likewise.
12811 (CALL_USED_REGISTERS): Likewise.
12812 (CALL_REALLY_USED_REGISTERS): Likewise.
12813 (REG_ALLOC_ORDER): Likewise.
12814 (enum reg_class): Likewise.
12815 (REG_CLASS_NAMES): Likewise.
12816 (REG_CLASS_CONTENTS): Likewise.
12817 (REGISTER_NAMES): Likewise.
12818 (ADDITIONAL_REGISTER_NAMES): Likewise.
12819 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
12820 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
12821 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
12822 * config/rs6000/htm.md: New file.
12823 * config/rs6000/htmintrin.h: New file.
12824 * config/rs6000/htmxlintrin.h: New file.
12826 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
12828 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
12829 Define SYMBOL_TINY_GOT, update comment.
12830 * config/aarch64/aarch64.c
12831 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
12832 (aarch64_expand_mov_immediate): Likewise.
12833 (aarch64_print_operand): Likewise.
12834 (aarch64_classify_symbol): Likewise.
12835 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
12836 (ldr_got_tiny): Define.
12838 2013-07-13 Tobias Grosser <tobias@grosser.es>
12840 PR tree-optimization/54094
12841 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
12842 scheduling dimension for the parallelism check from the polyhedral
12843 information in the AST.
12844 * graphite-dependences.c (carries_deps): Do not assume the schedule is
12847 2013-07-13 Jason Merrill <jason@redhat.com>
12849 * print-tree.c (debug_vec_tree): Use debug_raw.
12850 (debug_raw (vec<tree, va_gc> &)): New.
12851 (debug_raw (vec<tree, va_gc> *)): New.
12852 * tree.h: Declare them.
12854 2013-07-13 Bin Cheng <bin.cheng@arm.com>
12856 * ifcvt.c (ifcvt_after_combine): New static variable.
12857 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
12859 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
12860 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
12861 rest_of_handle_if_after_reload): Pass new argument for if_convert.
12863 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
12865 * config/mips/mips.c (mips_expand_call): Remove empty statement.
12867 2013-07-12 Michael Matz <matz@suse.de>
12869 PR middle-end/55771
12870 * convert.c (convert_to_real): Reject non-float inner types.
12872 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
12874 * config/aarch64/aarch64-protos.h
12875 (aarch64_simd_immediate_valid_for_move): Remove.
12876 * config/aarch64/aarch64.c (simd_immediate_info): New member.
12877 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
12879 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
12881 2013-07-11 Steve Ellcey <sellcey@mips.com>
12883 * config/mips/mips.c (mips_conditional_register_usage): Do not
12884 use t[0-7] registers in MIPS16 mode when optimizing for size.
12886 2013-07-11 Sriraman Tallam <tmsriram@google.com>
12888 * config/i386/i386.c (dispatch_function_versions): Fix array
12889 indexing of function_version_info to match actual_versions.
12891 2013-07-11 Teresa Johnson <tejohnson@google.com>
12893 * vec.h (struct va_gc): Move release out-of-line.
12894 (va_gc::release): Call ggc_free on released vec.
12896 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12898 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
12899 Require GOT register as additional operand in UNSPEC.
12900 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
12901 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12902 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12903 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
12904 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
12905 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
12906 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
12908 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
12911 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
12912 name seen by assembler/linker rather if available.
12914 2013-07-11 Andreas Schwab <schwab@suse.de>
12916 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
12918 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
12920 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
12922 2013-07-10 Joseph Myers <joseph@codesourcery.com>
12924 * doc/tm.texi.in: Move hook documentation to ....
12925 * target.def: ... here.
12927 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
12928 text on @hook line.
12929 * doc/tm.texi: Regenerate.
12931 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
12934 * doc/invoke.texi: Document -Wconditionally-supported.
12936 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12939 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
12942 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12945 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
12946 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
12947 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
12948 Remove duplicate devices.
12949 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
12950 * config/avr/t-multilib: Regenerate.
12951 * config/avr/avr-tables.opt: Regenerate.
12952 * doc/avr-mmcu.texi: Regenerate.
12954 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12957 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
12959 2013-07-10 Graham Stott <graham.stott@btinternet.com>
12961 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
12962 the cost of MULT when optimizing for size.
12964 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12966 * config/cr16/cr16-protos.h: Don't include target.h.
12968 2013-07-09 Joseph Myers <joseph@codesourcery.com>
12970 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
12971 adjust register size for TDmode and TFmode for VSX registers.
12973 2013-07-08 Kai Tietz <ktietz@redhat.com>
12976 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
12977 hook_bool_const_tree_true.
12979 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12981 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
12982 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
12983 * config/s390/s390.md: Define FPR*_REGNUM constants.
12984 Fix FPR2_REGNUM constant (18 -> 17).
12985 ("*trunc<BFP:mode><DFP_ALL:mode>2")
12986 ("*trunc<DFP_ALL:mode><BFP:mode>2")
12987 ("trunc<BFP:mode><DFP_ALL:mode>2")
12988 ("trunc<DFP_ALL:mode><BFP:mode>2")
12989 ("*extend<BFP:mode><DFP_ALL:mode>2")
12990 ("*extend<DFP_ALL:mode><BFP:mode>2")
12991 ("extend<BFP:mode><DFP_ALL:mode>2")
12992 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
12995 2013-07-08 Graham Stott <graham.stott@btinternet.com>
12997 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
12999 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13001 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
13002 and cfun_fpr_bit_p to cfun_fpr_save_p.
13003 (s390_frame_area, s390_register_info, s390_frame_info)
13004 (s390_emit_prologue, s390_emit_epilogue)
13005 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
13007 * config/s390/s390.h: Define *_REGNUM macros for floating point
13010 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
13012 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
13014 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
13016 PR rtl-optimization/57786
13017 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
13018 and break out of the loop when it is set to false.
13020 2013-07-08 Jakub Jelinek <jakub@redhat.com>
13023 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
13024 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
13025 (const_int 63)) 0)).
13026 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
13027 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
13028 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
13030 PR rtl-optimization/57829
13031 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
13032 mask bits outside of mode are just sign-extension from mode to HWI.
13034 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
13036 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
13037 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
13038 adjust_address instead of change_address to keep info about alignment.
13039 (emit_strmov): Remove.
13040 (emit_memmov): New function.
13041 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
13042 (expand_movmem_epilogue): Likewise and return updated rtx for
13044 (expand_constant_movmem_prologue): Likewise and return updated rtx for
13045 destination and source.
13046 (decide_alignment): Refactor, handle vector_loop.
13047 (ix86_expand_movmem): Likewise.
13048 (ix86_expand_setmem): Likewise.
13049 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
13051 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
13053 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
13054 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
13056 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
13058 * config/i386/sse.md (sse_movlhps): Change alternative 3
13059 of operand 2 to "m".
13061 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
13064 * config/i386/sse.md (iptr): New mode attribute.
13065 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
13066 (<sse>_vm<plusminus_insn><mode>3): Ditto.
13067 (<sse>_vmmul<mode>3): Ditto.
13068 (<sse>_vmdiv<mode>3): Ditto.
13069 (sse_vmrcpv4sf2): Ditto.
13070 (<sse>_vmsqrt<mode>2): Ditto.
13071 (sse_vmrsqrtv4sf2): Ditto.
13072 (<sse>_vm<code><mode>3): Ditto.
13073 (avx_vmcmp<mode>3): Ditto.
13074 (<sse>_vmmaskcmp<mode>3): Ditto.
13075 (<sse>_comi): Ditto.
13076 (<sse>_ucomi): Ditto.
13077 (*xop_vmfrcz_<mode>): Ditto.
13078 (*fmai_fmadd_<mode>): Ditto.
13079 (*fmai_fmsub_<mode>): Ditto.
13080 (*fmai_fnmadd_<mode>): Ditto.
13081 (*fmai_fnmsub_<mode>): Ditto.
13082 (*fma4i_vmfmadd_<mode>): Ditto.
13083 (*fma4i_vmfmsub_<mode>): Ditto.
13084 (*fma4i_vmfnmadd_<mode>): Ditto.
13085 (*fma4i_vmfnmsub_<mode>): Ditto.
13086 (*xop_vmfrcz_<mode>): Ditto.
13087 (sse_cvtps2pi): Ditto.
13088 (sse_cvttps2pi): Ditto.
13089 (sse_cvtss2si): Ditto.
13090 (sse_cvtss2si_2): Ditto.
13091 (sse_cvtss2siq_2): Ditto.
13092 (sse_cvttss2si): Ditto.
13093 (sse_cvttss2siq): Ditto.
13094 (sse_cvtsd2si): Ditto.
13095 (sse_cvtsd2si_2): Ditto.
13096 (sse_cvtsd2siq_2): Ditto.
13097 (sse_cvttsd2si): Ditto.
13098 (sse_cvttsd2siq): Ditto.
13099 (sse_cvtsd2ss): Ditto.
13100 (sse_cvtss2sd): Ditto.
13101 (avx2_pbroadcast<mode>): Ditto.
13102 (avx2_pbroadcast<mode>_1): Ditto.
13103 (*avx_vperm_broadcast_v4sf): Ditto.
13105 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
13106 (sse_movlhps): Ditto.
13107 (sse_storehps): Ditto.
13108 (sse_loadhps): Ditto.
13109 (sse_storelps): Ditto.
13110 (sse_loadlps): Ditto.
13111 (*vec_concatv4sf): Ditto.
13112 (*vec_interleave_highv2df): Ditto.
13113 (*vec_interleave_lowv2df): Ditto.
13114 (*vec_extractv2df_1_sse): Ditto.
13115 (*vec_extractv2df_0_sse): Ditto.
13116 (sse2_storelpd): Ditto.
13117 (sse2_loadlpd): Ditto.
13118 (sse2_movsd): Ditto.
13119 (*vec_concatv4si): Ditto.
13120 (vec_concatv2di): Ditto.
13122 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
13123 for Intel asm dialect.
13124 (mmx_punpcklwd): Ditto.
13125 (mmx_punpckldq): Ditto.
13127 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
13128 for intel assembler dialect.
13130 2013-07-06 Jakub Jelinek <jakub@redhat.com>
13133 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
13134 for BUILT_IN_C{LZ,LRSB}*.
13135 * tree.h (CASE_INT_FN): Add FN##IMAX case.
13136 * tree-vrp.c (extract_range_basic): Handle
13137 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
13138 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
13139 fall thru to code calling set_value*.
13140 * builtins.c (expand_builtin): Remove *IMAX cases.
13141 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
13142 if width is bigger than 2*HWI.
13144 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
13146 PR rtl-optimization/55342
13147 * lra-int.h (lra_subreg_reload_pseudos): New.
13148 * lra.c: Add undoing optional reloads to the block diagram.
13149 (lra_subreg_reload_pseudos): New.
13150 (lra_optional_reload_pseudos): Change comments.
13151 (lra): Init and clear lra_subreg_reload_pseudos. Clear
13152 lra_optional_reload_pseudos after undo transformations.
13153 * lra-assigns.c (pseudo_prefix_title): New.
13154 (lra_setup_reg_renumber): Use it.
13155 (spill_for): Ditto. Check subreg reload pseudos too.
13156 (assign_by_spills): Consider subreg reload pseudos too.
13157 * lra-constraints.c (simplify_operand_subreg): Use
13158 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
13159 (curr_insn_transform): Recognize and do optional reloads.
13160 (undo_optional_reloads): New.
13161 (lra_undo_inheritance): Call undo_optional_reloads.
13163 2013-07-05 Thomas Quinot <quinot@adacore.com>
13165 * tree-complex.c (expand_complex_operations_1): Fix typo.
13167 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
13169 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
13170 (tune_params): New member 'const vec_costs'.
13171 * config/aarch64/aarch64.c (generic_vector_cost): New.
13172 (generic_tunings): New member 'generic_vector_cost'.
13173 (aarch64_builtin_vectorization_cost): New.
13174 (aarch64_add_stmt_cost): New.
13175 (TARGET_VECTORIZE_ADD_STMT_COST): New.
13176 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
13178 2013-07-03 Jakub Jelinek <jakub@redhat.com>
13181 * config/i386/predicates.md (vsib_address_operand): Disallow
13182 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
13184 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
13186 PR middle-end/55030
13187 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
13188 expand_builtin_setjmp_receiver.
13189 (expand_label): Adjust, call expand_builtin_setjmp_receiver
13190 with NULL for the label parameter.
13191 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
13192 the frame-pointer. Adjust comments.
13193 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
13194 only if LABEL is non-NULL.
13196 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
13198 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
13199 (ARM_ABI_AAPCS64): Ditto.
13201 (ARM_DEFAULT_ABI): Ditto.
13203 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
13205 * config/aarch64/aarch64-builtins.c
13206 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
13207 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
13209 * config/aarch64/aarch64-simd.md
13210 (aarch64_ld1<VALL:mode>): New.
13211 (aarch64_st1<VALL:mode>): Likewise.
13212 * config/aarch64/arm_neon.h
13213 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
13215 2013-07-02 Sriraman Tallam <tmsriram@google.com>
13217 * config/i386/i386.c (gate_insert_vzeroupper): Check if
13219 (ix86_option_override_internal):Turn on all -mavx target flags by
13220 default as they are dependent on AVX anyway.
13222 2013-07-02 Cary Coutant <ccoutant@google.com>
13224 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
13225 deterministic hash.
13226 (loc_checksum_ordered): Likewise.
13227 (hash_loc_operands): Remove inline keyword.
13229 2013-07-02 Jakub Jelinek <jakub@redhat.com>
13231 PR tree-optimization/57741
13232 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
13233 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
13234 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
13235 Allow REAL_CST step_exprs if flag_associative_math.
13236 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
13238 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13240 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
13242 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13244 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
13246 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13248 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
13250 (iorsi3_insn): Likewise.
13251 (arm_xorsi3): Likewise.
13253 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
13255 * arm.md (attribute "wtype"): Delete. Move attribute values from here
13257 (attribute "type"): ... here, and prefix with "wmmx_".
13258 (attribute "core_cycles"): Update for attribute changes.
13259 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
13260 (tbcstv4hi): Likewise.
13261 (tbcstv2si): Likewise.
13262 (iwmmxt_iordi3): Likewise.
13263 (iwmmxt_xordi3): Likewise.
13264 (iwmmxt_anddi3): Likewise.
13265 (iwmmxt_nanddi3): Likewise.
13266 (iwmmxt_arm_movdi): Likewise.
13267 (iwmmxt_movsi_insn): Likewise.
13268 (mov<mode>_internal): Likewise.
13269 (and<mode>3_iwmmxt): Likewise.
13270 (ior<mode>3_iwmmxt): Likewise.
13271 (xor<mode>3_iwmmxt): Likewise.
13272 (add<mode>3_iwmmxt): Likewise.
13273 (ssaddv8qi3): Likewise.
13274 (ssaddv4hi3): Likewise.
13275 (ssaddv2si3): Likewise.
13276 (usaddv8qi3): Likewise.
13277 (usaddv4hi3): Likewise.
13278 (usaddv2si3): Likewise.
13279 (sub<mode>3_iwmmxt): Likewise.
13280 (sssubv8qi3): Likewise.
13281 (sssubv4hi3): Likewise.
13282 (sssubv2si3): Likewise.
13283 (ussubv8qi3): Likewise.
13284 (ussubv4hi3): Likewise.
13285 (ussubv2si3): Likewise.
13286 (mulv4hi3_iwmmxt): Likewise.
13287 (smulv4hi3_highpart): Likewise.
13288 (umulv4hi3_highpart): Likewise.
13289 (iwmmxt_wmacs): Likewise.
13290 (iwmmxt_wmacsz): Likewise.
13291 (iwmmxt_wmacu): Likewise.
13292 (iwmmxt_wmacuz): Likewise.
13293 (iwmmxt_clrdi): Likewise.
13294 (iwmmxt_clrv8qi): Likewise.
13295 (iwmmxt_clr4hi): Likewise.
13296 (iwmmxt_clr2si): Likewise.
13297 (iwmmxt_uavgrndv8qi3): Likewise.
13298 (iwmmxt_uavgrndv4hi3): Likewise.
13299 (iwmmxt_uavgv8qi3): Likewise.
13300 (iwmmxt_uavgv4hi3): Likewise.
13301 (iwmmxt_tinsrb): Likewise.
13302 (iwmmxt_tinsrh): Likewise.
13303 (iwmmxt_tinsrw): Likewise.
13304 (iwmmxt_textrmub): Likewise.
13305 (iwmmxt_textrmsb): Likewise.
13306 (iwmmxt_textrmuh): Likewise.
13307 (iwmmxt_textrmsh): Likewise.
13308 (iwmmxt_textrmw): Likewise.
13309 (iwmxxt_wshufh): Likewise.
13310 (eqv8qi3): Likewise.
13311 (eqv4hi3): Likewise.
13312 (eqv2si3): Likewise.
13313 (gtuv8qi3): Likewise.
13314 (gtuv4hi3): Likewise.
13315 (gtuv2si3): Likewise.
13316 (gtv8qi3): Likewise.
13317 (gtv4hi3): Likewise.
13318 (gtv2si3): Likewise.
13319 (smax<mode>3_iwmmxt): Likewise.
13320 (umax<mode>3_iwmmxt): Likewise.
13321 (smin<mode>3_iwmmxt): Likewise.
13322 (umin<mode>3_iwmmxt): Likewise.
13323 (iwmmxt_wpackhss): Likewise.
13324 (iwmmxt_wpackwss): Likewise.
13325 (iwmmxt_wpackdss): Likewise.
13326 (iwmmxt_wpackhus): Likewise.
13327 (iwmmxt_wpackwus): Likewise.
13328 (iwmmxt_wpackdus): Likewise.
13329 (iwmmxt_wunpckihb): Likewise.
13330 (iwmmxt_wunpckihh): Likewise.
13331 (iwmmxt_wunpckihw): Likewise.
13332 (iwmmxt_wunpckilb): Likewise.
13333 (iwmmxt_wunpckilh): Likewise.
13334 (iwmmxt_wunpckilw): Likewise.
13335 (iwmmxt_wunpckehub): Likewise.
13336 (iwmmxt_wunpckehuh): Likewise.
13337 (iwmmxt_wunpckehuw): Likewise.
13338 (iwmmxt_wunpckehsb): Likewise.
13339 (iwmmxt_wunpckehsh): Likewise.
13340 (iwmmxt_wunpckehsw): Likewise.
13341 (iwmmxt_wunpckelub): Likewise.
13342 (iwmmxt_wunpckeluh): Likewise.
13343 (iwmmxt_wunpckeluw): Likewise.
13344 (iwmmxt_wunpckelsb): Likewise.
13345 (iwmmxt_wunpckelsh): Likewise.
13346 (iwmmxt_wunpckelsw): Likewise.
13347 (ror<mode>3): Likewise.
13348 (ashr<mode>3_iwmmxt): Likewise.
13349 (lshr<mode>3_iwmmxt): Likewise.
13350 (ashl<mode>3_iwmmxt): Likewise.
13351 (ror<mode>3_di): Likewise.
13352 (ashr<mode>3_di): Likewise.
13353 (lshr<mode>3_di): Likewise.
13354 (ashl<mode>3_di): Likewise.
13355 (iwmmxt_wmadds): Likewise.
13356 (iwmmxt_wmaddu): Likewise.
13357 (iwmmxt_tmia): Likewise.
13358 (iwmmxt_tmiaph): Likewise.
13359 (iwmmxt_tmiabb): Likewise.
13360 (iwmmxt_tmiatb): Likewise.
13361 (iwmmxt_tmiabt): Likewise.
13362 (iwmmxt_tmiatt): Likewise.
13363 (iwmmxt_tmovmskb): Likewise.
13364 (iwmmxt_tmovmskh): Likewise.
13365 (iwmmxt_tmovmskw): Likewise.
13366 (iwmmxt_waccb): Likewise.
13367 (iwmmxt_wacch): Likewise.
13368 (iwmmxt_waccw): Likewise.
13369 (iwmmxt_waligni): Likewise.
13370 (iwmmxt_walignr): Likewise.
13371 (iwmmxt_walignr0): Likewise.
13372 (iwmmxt_walignr1): Likewise.
13373 (iwmmxt_walignr2): Likewise.
13374 (iwmmxt_walignr3): Likewise.
13375 (iwmmxt_wsadb): Likewise.
13376 (iwmmxt_wsadh): Likewise.
13377 (iwmmxt_wsadbz): Likewise.
13378 (iwmmxt_wsadhz): Likewise.
13379 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
13380 (iwmmxt_wabsdiffb): Likewise.
13381 (iwmmxt_wabsdiffh): Likewise.
13382 (iwmmxt_wabsdiffw): Likewise.
13383 (iwmmxt_waddsubhx): Likewise
13384 (iwmmxt_wsubaddhx): Likewise.
13385 (addc<mode>3): Likewise.
13386 (iwmmxt_avg4): Likewise.
13387 (iwmmxt_avg4r): Likewise.
13388 (iwmmxt_wmaddsx): Likewise.
13389 (iwmmxt_wmaddux): Likewise.
13390 (iwmmxt_wmaddsn): Likewise.
13391 (iwmmxt_wmaddun): Likewise.
13392 (iwmmxt_wmulwsm): Likewise.
13393 (iwmmxt_wmulwum): Likewise.
13394 (iwmmxt_wmulsmr): Likewise.
13395 (iwmmxt_wmulumr): Likewise.
13396 (iwmmxt_wmulwsmr): Likewise.
13397 (iwmmxt_wmulwumr): Likewise.
13398 (iwmmxt_wmulwl): Likewise.
13399 (iwmmxt_wqmulm): Likewise.
13400 (iwmmxt_wqmulwm): Likewise.
13401 (iwmmxt_wqmulmr): Likewise.
13402 (iwmmxt_wqmulwmr): Likewise.
13403 (iwmmxt_waddbhusm): Likewise.
13404 (iwmmxt_waddbhusl): Likewise.
13405 (iwmmxt_wqmiabb): Likewise.
13406 (iwmmxt_wqmiabt): Likewise.
13407 (iwmmxt_wqmiatb): Likewise.
13408 (iwmmxt_wqmiatt): Likewise.
13409 (iwmmxt_wqmiabbn): Likewise.
13410 (iwmmxt_wqmiabtn): Likewise.
13411 (iwmmxt_wqmiatbn): Likewise.
13412 (iwmmxt_wqmiattn): Likewise.
13413 (iwmmxt_wmiabb): Likewise.
13414 (iwmmxt_wmiabt): Likewise.
13415 (iwmmxt_wmiatb): Likewise.
13416 (iwmmxt_wmiatt): Likewise.
13417 (iwmmxt_wmiabbn): Likewise.
13418 (iwmmxt_wmiabtn): Likewise.
13419 (iwmmxt_wmiatbn): Likewise.
13420 (iwmmxt_wmiattn): Likewise.
13421 (iwmmxt_wmiawbb): Likewise.
13422 (iwmmxt_wmiawbt): Likewise.
13423 (iwmmxt_wmiawtb): Likewise.
13424 (iwmmxt_wmiawtt): Likewise.
13425 (iwmmxt_wmiawbbn): Likewise.
13426 (iwmmxt_wmiawbtn): Likewise.
13427 (iwmmxt_wmiawtbn): Likewise.
13428 (iwmmxt_wmiawttn): Likewise.
13429 (iwmmxt_wmerge): Likewise.
13430 (iwmmxt_tandc<mode>3): Likewise.
13431 (iwmmxt_torc<mode>3): Likewise.
13432 (iwmmxt_torvsc<mode>3): Likewise.
13433 (iwmmxt_textrc<mode>3): Likewise.
13434 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
13435 (wmmxt_pack): Likewise.
13436 (wmmxt_mult_c1): Likewise.
13437 (wmmxt_mult_c2): Likewise.
13438 (wmmxt_alu_c1): Likewise.
13439 (wmmxt_alu_c2): Likewise.
13440 (wmmxt_alu_c3): Likewise.
13441 (wmmxt_transfer_c1): Likewise.
13442 (wmmxt_transfer_c2): Likewise.
13443 (wmmxt_transfer_c3): Likewise.
13444 (marvell_f_iwmmxt_wstr): Likewise.
13445 (marvell_f_iwmmxt_wldr): Likewise.
13447 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
13449 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
13451 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13454 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13455 * lra-constraints.c (need_for_split_p): Check call used hard regs
13456 living through calls.
13458 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
13459 call used regs for call insn.
13461 2013-06-28 Jakub Jelinek <jakub@redhat.com>
13464 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
13465 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
13467 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
13469 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
13471 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13473 * lra-constraints.c (need_for_split_p): Check call used hard regs
13474 living through calls.
13476 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
13479 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
13480 to tie with any other modes. Eliminate Altivec vector mode tests,
13481 since these are a subset of ALTIVEC or VSX vector modes. Simplify
13482 code, to return 0 if testing MODE2 for a condition, if we've
13483 already tested MODE1 for the same condition.
13485 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13487 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
13490 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13492 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
13493 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
13495 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13497 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
13499 (aarch64_symbolic_constant_p): Remove.
13500 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
13501 static. Fix line length and white space.
13502 (aarch64_symbolic_constant_p): Remove.
13503 * config/aarch64/predicates.md (aarch64_valid_symref):
13504 Use aarch64_classify_symbol_expression.
13506 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13508 * config/arm/constraints.md (Ts): New constraint.
13509 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
13511 (compare_scc): Use "Ts" constraint for operand 0.
13512 (ior_scc_scc): Likewise.
13513 (and_scc_scc): Likewise.
13514 (and_scc_scc_nodom): Likewise.
13515 (ior_scc_scc_cmp): Likewise for operand 7.
13516 (and_scc_scc_cmp): Likewise.
13517 * config/arm/thumb2.md (thumb2_movsi_insn):
13518 Add alternatives for 16-bit encodings.
13519 (thumb2_movhi_insn): Likewise.
13520 (thumb2_movsicc_insn): Likewise.
13521 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
13522 (thumb2_negscc): Use "Ts" constraint.
13523 Move mvn instruction outside cond_exec block.
13524 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
13525 for 16-bit encodings.
13527 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13529 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
13531 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
13532 (mulsi3subsi): Likewise.
13533 (mulsidi3adddi): Likewise.
13534 (mulsidi3_v6): Likewise.
13535 (umulsidi3_v6): Likewise.
13536 (umulsidi3adddi_v6): Likewise.
13537 (smulsi3_highpart_v6): Likewise.
13538 (umulsi3_highpart_v6): Likewise.
13539 (mulhisi3tb): Likewise.
13540 (mulhisi3bt): Likewise.
13541 (mulhisi3tt): Likewise.
13542 (maddhisi4): Likewise.
13543 (maddhisi4tb): Likewise.
13544 (maddhisi4tt): Likewise.
13545 (maddhidi4): Likewise.
13546 (maddhidi4tb): Likewise.
13547 (maddhidi4tt): Likewise.
13548 (zeroextractsi_compare0_scratch): Likewise.
13549 (insv_zero): Likewise.
13550 (insv_t2): Likewise.
13551 (anddi_notzesidi_di): Likewise.
13552 (anddi_notsesidi_di): Likewise.
13553 (andsi_notsi_si): Likewise.
13554 (iordi_zesidi_di): Likewise.
13555 (xordi_zesidi_di): Likewise.
13556 (andsi_iorsi3_notsi): Likewise.
13557 (smax_0): Likewise.
13558 (smax_m1): Likewise.
13559 (smin_0): Likewise.
13560 (not_shiftsi): Likewise.
13561 (unaligned_loadsi): Likewise.
13562 (unaligned_loadhis): Likewise.
13563 (unaligned_loadhiu): Likewise.
13564 (unaligned_storesi): Likewise.
13565 (unaligned_storehi): Likewise.
13566 (extv_reg): Likewise.
13567 (extzv_t2): Likewise.
13568 (divsi3): Likewise.
13569 (udivsi3): Likewise.
13570 (arm_zero_extendhisi2addsi): Likewise.
13571 (arm_zero_extendqisi2addsi): Likewise.
13572 (compareqi_eq0): Likewise.
13573 (arm_extendhisi2_v6): Likewise.
13574 (arm_extendqisi2addsi): Likewise.
13575 (arm_movt): Likewise.
13576 (thumb2_ldrd): Likewise.
13577 (thumb2_ldrd_base): Likewise.
13578 (thumb2_ldrd_base_neg): Likewise.
13579 (thumb2_strd): Likewise.
13580 (thumb2_strd_base): Likewise.
13581 (thumb2_strd_base_neg): Likewise.
13582 (arm_negsi2): Add alternative for 16-bit encoding.
13583 (arm_one_cmplsi2): Likewise.
13585 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13587 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
13588 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
13589 (movdfcc): Likewise.
13590 * config/arm/vfp.md (*thumb2_movsf_vfp):
13591 Disable predication for arm_restrict_it.
13592 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
13593 (*thumb2_movdfcc_vfp): Likewise.
13594 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
13595 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
13596 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
13597 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
13598 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
13599 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
13600 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
13601 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
13602 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
13603 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
13604 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
13605 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
13606 Disable predication for arm_restrict_it.
13608 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
13610 * config/i386/bmiintrin.h (_bextr_u32): New.
13611 (_bextr_u64): Ditto.
13613 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
13615 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
13616 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
13617 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
13618 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
13619 * config/mips/n32-elf.h: ...this new file.
13621 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
13624 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
13625 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
13627 2013-06-27 Catherine Moore <clm@codesourcery.com>
13629 * config/mips/mips-tables.opt: Regenerate.
13630 * config/mips/mips-cpus.def: Add m14ke and m14kec.
13631 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
13632 * doc/invoke.texi: Add -m14kc.
13634 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13637 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
13638 constraints of operand 1 and 2.
13641 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
13642 to match RTL canonicalization. Swap predicates and
13643 constraints of operand 1 and 2.
13645 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
13647 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
13648 Process OP_INOUT regs for splitting too.
13650 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13652 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
13653 decl before the loop, initialize to NULL.
13654 (vectorizable_load): Initialize ptr_incr to NULL.
13656 2013-06-27 Martin Jambor <mjambor@suse.cz>
13659 * ipa-ref.h (ipa_maybe_record_reference): Declare.
13660 * ipa-ref.c (ipa_maybe_record_reference): New function.
13661 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
13662 * ipa-cp.c (create_specialized_node): Record potential references from
13664 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
13666 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13668 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
13669 parameter 'mode' of type 'enum machine_mode mode'; change to pass
13670 'mode' to force_reg.
13671 (aarch64_add_offset): Update calls to aarch64_force_temporary.
13672 (aarch64_expand_mov_immediate): Likewise.
13674 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13676 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
13677 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
13679 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13681 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
13682 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13683 (struct machine_function): Add tbegin_p.
13684 (s390_canonicalize_comparison): Fold CC mode compares to
13685 conditional jump if possible.
13686 (s390_emit_jump): Return the emitted jump.
13687 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
13688 Handle CCRAWmode compares.
13689 (s390_option_override): Default to -mhtm if available.
13690 (s390_reg_clobbered_rtx): Handle floating point regs as well.
13691 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
13692 FPRs instead of df_regs_ever_live_p.
13693 (s390_optimize_nonescaping_tx): New function.
13694 (s390_init_frame_layout): Extend clobbered_regs array to cover
13696 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
13697 (s390_expand_tbegin): New function.
13698 (enum s390_builtin): New enum definition.
13699 (code_for_builtin): New array definition.
13700 (s390_init_builtins): New function.
13701 (s390_expand_builtin): New function.
13702 (TARGET_INIT_BUILTINS): Define.
13703 (TARGET_EXPAND_BUILTIN): Define.
13704 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
13705 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
13706 (s390_alc_comparison): Likewise.
13707 * config/s390/s390-modes.def: Add CCRAWmode.
13708 * config/s390/s390.h (processor_flags): Add PF_TX.
13709 (TARGET_CPU_HTM): Define macro.
13710 (TARGET_HTM): Define macro.
13711 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
13712 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
13713 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13714 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
13715 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
13716 (TBEGIN_MASK, TBEGINC_MASK): New constants.
13717 ("*cc_to_int"): Move up.
13718 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
13719 constants other than 0.
13720 ("*ccraw_to_int"): New insn and splitter definition.
13721 ("tbegin", "tbegin_nofloat", "tbegin_retry")
13722 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
13723 ("tx_assist"): New expander.
13724 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
13725 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
13726 * config/s390/s390.opt: Add -mhtm option.
13727 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
13728 * config/s390/htmxlintrin.h: New file.
13729 * config/s390/htmintrin.h: New file.
13730 * config/s390/s390intrin.h: New file.
13731 * doc/extend.texi: Document htm builtins.
13732 * config.gcc: Add the new header files to extra_headers.
13734 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
13736 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
13737 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
13739 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
13740 Pat Haugen <pthaugen@us.ibm.com>
13741 Peter Bergner <bergner@vnet.ibm.com>
13743 * config/rs6000/power8.md: New.
13744 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
13745 setting for power8 entry.
13746 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
13747 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
13748 test for Power4/Power5 only.
13749 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
13751 (force_new_group): Adjust comment.
13752 * config/rs6000/rs6000.md: Include power8.md.
13754 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
13756 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
13757 * config/arm/arm-protos.h (arm_max_conditional_execute): New
13759 (tune_params): Update comment.
13760 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
13761 (arm_max_conditional_execute): New function.
13762 (thumb2_final_prescan_insn): Use max_insn_skipped and
13763 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
13765 2013-06-25 Jakub Jelinek <jakub@redhat.com>
13767 PR tree-optimization/57705
13768 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
13769 SSA_NAME step, provided that it is not defined inside the loop.
13770 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
13771 (get_initial_def_for_induction): Handle SSA_NAME IV step.
13773 2013-06-25 Martin Jambor <mjambor@suse.cz>
13775 PR middle-end/57670
13776 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
13777 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
13779 (ipa_note_param_call): Initialize member_ptr flag.
13780 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
13781 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
13782 (ipa_write_indirect_edge_info): Stream member_ptr flag.
13783 (ipa_read_indirect_edge_info): Likewise.
13785 2013-06-25 Richard Biener <rguenther@suse.de>
13787 PR middle-end/56977
13788 * passes.c (init_optimization_passes): Move pass_fold_builtins
13789 and pass_dce earlier with -Og.
13791 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
13793 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
13794 <BIT_FIELD_REF>: Remove trailing TAB.
13795 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
13798 2013-06-24 Martin Jambor <mjambor@suse.cz>
13800 PR tree-optimization/57358
13801 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
13802 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
13803 (ipa_analyze_params_uses): Generate pessimistic info when true.
13805 2013-06-24 Martin Jambor <mjambor@suse.cz>
13807 PR tree-optimization/57539
13808 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
13809 global.inlined_to of the new node to it. All callers changed.
13810 * ipa-inline-transform.c (clone_inlined_nodes): New variable
13811 inlining_into, pass it to cgraph_clone_node.
13812 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
13813 ipa_free_edge_args_substructures.
13814 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
13815 rdesc linked list. Do not assert rdesc edges have inlined caller.
13816 Assert we have found an rdesc in the rdesc list.
13818 2013-06-24 Richard Biener <rguenther@suse.de>
13820 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
13821 (pointer_set_lookup): Declare.
13822 (class pointer_map): New template class implementing a
13823 generic pointer to T map.
13824 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
13825 pointer_map<T>::contains, pointer_map<T>::insert,
13826 pointer_map<T>::traverse): New functions.
13827 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
13828 (pointer_set_lookup): New function.
13829 (pointer_set_contains): Use pointer_set_lookup.
13830 (pointer_set_insert): Likewise.
13831 (insert_aux): Remove.
13832 (struct pointer_map_t): Embed a pointer_set_t.
13833 (pointer_map_create): Adjust.
13834 (pointer_map_destroy): Likewise.
13835 (pointer_map_contains): Likewise.
13836 (pointer_map_insert): Likewise.
13837 (pointer_map_traverse): Likewise.
13838 * tree-streamer.h (struct streamer_tree_cache_d): Use a
13839 pointer_map<unsigned> instead of a pointer_map_t.
13840 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
13841 (streamer_tree_cache_lookup): Likewise.
13842 (streamer_tree_cache_create): Likewise.
13843 (streamer_tree_cache_delete): Likewise.
13844 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
13845 pointer_map<unsigned> instead of a pointer_map_t.
13846 (lto_init_tree_ref_encoder): Adjust.
13847 (lto_destroy_tree_ref_encoder): Likewise.
13848 * lto-section-out.c (lto_output_decl_index): Likewise.
13849 (lto_record_function_out_decl_state): Likewise.
13850 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
13852 2013-06-24 Richard Biener <rguenther@suse.de>
13854 PR tree-optimization/57488
13855 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
13857 2013-06-24 Alan Modra <amodra@gmail.com>
13859 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
13860 (gen_easy_altivec_constant): Likewise.
13861 * config/rs6000/predicates.md (easy_vector_constant_add_self,
13862 easy_vector_constant_msb): Likewise.
13864 2013-06-23 Jakub Jelinek <jakub@redhat.com>
13867 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
13868 add missing return true.
13870 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
13873 * config/sh/predicates.md (general_extend_operand): Invoke
13874 general_movsrc_operand for memory operands.
13875 (general_movsrc_operand): Allow reg+reg addressing, do not use
13876 general_operand for memory operands.
13878 2013-06-23 Sriraman Tallam <tmsriram@google.com>
13880 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
13881 when current target options does not apply.
13882 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
13883 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
13884 * config/i386/bmiintrin.h: Pass appropriate target
13885 attributes to header.
13886 * config/i386/mmintrin.h: Ditto.
13887 * config/i386/nmmintrin.h: Ditto.
13888 * config/i386/avx2intrin.h: Ditto.
13889 * config/i386/fxsrintrin.h: Ditto.
13890 * config/i386/tbmintrin.h: Ditto.
13891 * config/i386/xsaveintrin.h: Ditto.
13892 * config/i386/f16cintrin.h: Ditto.
13893 * config/i386/xtestintrin.h: Ditto.
13894 * config/i386/xsaveoptintrin.h: Ditto.
13895 * config/i386/bmi2intrin.h: Ditto.
13896 * config/i386/lzcntintrin.h: Ditto.
13897 * config/i386/smmintrin.h: Ditto.
13898 * config/i386/wmmintrin.h: Ditto.
13899 * config/i386/x86intrin.h: Remove all header include guards.
13900 * config/i386/prfchwintrin.h: Ditto.
13901 * config/i386/pmmintrin.h: Ditto.
13902 * config/i386/tmmintrin.h: Ditto.
13903 * config/i386/xmmintrin.h: Ditto.
13904 * config/i386/popcntintrin.h: Ditto.
13905 * config/i386/rdseedintrin.h: Ditto.
13906 * config/i386/ammintrin.h: Ditto.
13907 * config/i386/emmintrin.h: Ditto.
13908 * config/i386/immintrin.h: Remove all header include guards.
13909 * config/i386/fma4intrin.h: Ditto.
13910 * config/i386/lwpintrin.h: Ditto.
13911 * config/i386/xopintrin.h: Ditto.
13912 * config/i386/ia32intrin.h: Ditto.
13913 * config/i386/avxintrin.h: Ditto.
13914 * config/i386/rtmintrin.h: Ditto.
13915 * config/i386/fmaintrin.h: Ditto.
13916 * config/i386/mm3dnow.h: Ditto.
13918 2013-06-22 Sriraman Tallam <tmsriram@google.com>
13920 * common/config/i386/i386-common.c: Handle LZCNT.
13922 2013-06-22 Andi Kleen <ak@linux.intel.com>
13924 * doc/extend.texi: Use __atomic_store_n instead of
13925 __atomic_store in HLE example.
13927 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
13929 * config/sh/sh.c: Remove <cstdlib> workaround.
13931 2013-06-21 Andi Kleen <ak@linux.intel.com>
13933 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
13935 2013-06-21 Andi Kleen <ak@linux.intel.com>
13937 * doc/extend.texi: Document that __atomic_clear and
13938 __atomic_test_and_set should only be used with bool.
13940 2013-06-20 Jan Hubicka <jh@suse.cz>
13942 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
13943 types_same_for_odr.
13944 * tree.c (decls_same_for_odr): New function.
13945 (same_for_edr): New function.
13946 (types_same_for_odr): New function.
13947 (get_binfo_at_offset): Use it.
13948 * tree.h (types_same_for_odr): Declare.
13950 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
13951 Jason Merrill <jason@redhat.com>
13953 * system.h: Include <cstdlib> as well as <stdlib.h>.
13955 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
13958 * config/i386/i386.c (construct_container): Report error if
13959 long double is used with disabled x87 float returns.
13961 2013-06-20 Jan Hubicka <jh@suse.cz>
13963 * lto-cgraph.c (input_symtab): Do not set cgraph state.
13965 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
13967 PR rtl-optimization/57425
13968 PR rtl-optimization/57569
13969 * alias.c (write_dependence_p): Remove parameters mem_mode and
13970 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
13971 Changed all callers.
13972 (canon_anti_dependence): Get comments and semantics in sync.
13973 Add parameter mem_canonicalized. Changed all callers.
13974 * rtl.h (canon_anti_dependence): Update prototype.
13976 2013-06-20 Richard Biener <rguenther@suse.de>
13978 * data-streamer-in.c (streamer_read_uhwi): Optimize single
13979 byte case, inline streamer_read_uchar and defer section
13982 2013-06-20 Richard Biener <rguenther@suse.de>
13984 PR tree-optimization/57584
13985 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
13986 SSA names into the expanded expression that take part in
13987 abnormal coalescing.
13989 2013-06-19 Sharad Singhai <singhai@google.com>
13991 * gcov.c (print_usage): Handle new option.
13992 (process_args): Ditto.
13993 (get_gcov_intermediate_filename): New function.
13994 (output_intermediate_file): New function.
13995 (output_gcov_file): New function
13996 (generate_results): Handle new option.
13997 (release_function): Relase demangled name.
13998 (read_graph_file): Handle demangled name.
13999 (output_lines): Ditto.
14000 * doc/gcov.texi: Document gcov intermediate format.
14002 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
14005 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
14006 (lra_emit_add): Use the functions. Add comment about Y as an
14009 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
14012 * collect2.c (collect_atexit): New.
14013 (collect_exit): Delete.
14014 (main): Register collect_atexit with atexit.
14015 (collect_wait): Change collect_exit to exit.
14017 * collect2.h (collect_exit): Delete.
14018 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
14020 2013-06-19 Wei Mi <wmi@google.com>
14022 PR rtl-optimization/57518
14023 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
14024 if regno is used in paradoxical subreg.
14025 (update_equiv_regs): Check pdx_subregs[regno] before
14026 set a reg to be equivalent with a mem.
14028 2013-06-19 Matthias Klose <doko@ubuntu.com>
14031 * file-find.h (find_a_file): Add a mode parameter.
14032 * file-find.c (find_a_file): Likewise.
14033 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
14034 with X_OK for the executables.
14035 * collect2.c (main): Call find_a_file with X_OK.
14037 2013-06-19 Steve Ellcey <sellcey@mips.com>
14040 * config/mips/mips.md (casesi_internal_mips16_<mode>):
14041 Use NEXT_INSN instead of next_real_insn.
14043 2013-06-19 Jan Hubicka <jh@suse.cz>
14045 * cgraph.h (const_value_known_p): Replace by ...
14046 (ctor_for_folding): .. this one.
14047 * cgraphunit.c (process_function_and_variable_attributes): Use it.
14048 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
14049 * expr.c (expand_expr_real_1): Likewise.
14050 (string_constant): Likewise.
14051 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
14052 * ipa.c (process_references): Likewise.
14053 (symtab_remove_unreachable_nodes): Likewise.
14054 * ipa-inline-analysis.c (param_change_prob): Likewise.
14055 * gimple-fold.c (canonicalize_constructor_val): Likewise.
14056 (get_base_constructor): Likwise.
14057 * varpool.c (varpool_remove_node): Likewise.
14058 (varpool_remove_initializer): LIkewise.
14059 (dump_varpool_node): LIkwise.
14060 (const_value_known_p): Rewrite to ...
14061 (ctor_for_folding): ... this one.
14063 2013-06-19 Jakub Jelinek <jakub@redhat.com>
14066 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
14067 PERSONALITY in $PATH derived prefixes.
14069 2013-06-19 Jeff Law <law@redhat.com>
14071 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
14074 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
14075 (simplify_bitwise_binary): Use it to simpify certain binary ops on
14078 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
14080 * config/arm/vfp.md: Move VFP instruction classification documentation
14082 * config/arm/arm.md: ... here. Update instruction classification
14085 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
14087 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
14088 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
14089 pattern. Use more efficient sequences on ARMv5 and Thumb2.
14091 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
14094 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14095 with NEXT_INSN. Use tablejump_p to check for jump table data
14098 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
14101 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
14102 that now in C++ the value is correct per the C++ standards.
14104 2013-06-19 Richard Biener <rguenther@suse.de>
14106 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
14107 for global context.
14109 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14112 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14115 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14116 with next_active_insn.
14118 2013-06-18 Sriraman Tallam <tmsriram@google.com>
14120 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
14121 functions are inlined during failures to flag an error.
14122 * tree-inline.c (expand_call_inline): Allow the error to be flagged
14123 in early inline pass.
14125 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
14127 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
14130 2013-06-18 Julian Brown <julian@codesourcery.com>
14132 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
14133 Permit virtual register pre-reload if !strict.
14134 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
14136 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
14138 * config/arm/neon.md (movmisalign<mode>): Use
14139 neon_perm_struct_or_reg_operand instead of
14140 neon_struct_or_register_operand.
14141 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
14142 neon_permissive_struct_operand instead of neon_struct_operand.
14143 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
14144 neon_vector_mem_operand.
14145 * config/arm/predicates.md (neon_struct_operand): Adjust call to
14146 neon_vector_mem_operand.
14147 (neon_permissive_struct_operand): New.
14148 (neon_struct_or_register_operand): Rename to...
14149 (neon_perm_struct_or_reg_operand): This. Adjust call to
14150 neon_vector_mem_operand.
14152 2013-06-18 Richard Biener <rguenther@suse.de>
14154 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
14155 * lto-streamer.h: Include pointer-set.h.
14156 (struct lto_decl_slot): Remove.
14157 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
14158 Remove next_index entry.
14159 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14160 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14161 (lto_init_tree_ref_encoder): Adjust.
14162 (lto_destroy_tree_ref_encoder): Likewise.
14163 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14164 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14165 (lto_output_decl_index): Adjust.
14166 (lto_new_out_decl_state): Likewise.
14167 (lto_record_function_out_decl_state): Likewise.
14168 * lto-streamer-out.c (copy_function): Likewise.
14170 2013-06-18 Richard Biener <rguenther@suse.de>
14172 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
14173 * cgraphunit.c: Include cfgloop.h.
14174 (init_lowered_empty_function): Initialize the loop tree.
14175 (assemble_thunk): Insert new BBs into loops.
14177 2013-06-18 Richard Biener <rguenther@suse.de>
14179 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
14180 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
14181 the map from cache entry to cache index optional.
14182 (streamer_tree_cache_replace_tree): Adjust accordingly.
14183 (streamer_tree_cache_append): Likewise.
14184 (streamer_tree_cache_delete): Likewise.
14185 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
14186 streamer cache map from cache entry to cache index.
14187 * lto-streamer-out.c (create_output_block): Adjust.
14189 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
14191 * config/arm/arm.md (attribute "insn"): Move multiplication and
14192 division attributes to...
14193 (attribute "type"): ... here. Remove mult.
14194 (attribute "mul32"): New attribute.
14195 (attribute "mul64"): Add umaal.
14196 (*arm_mulsi3): Update attributes.
14197 (*arm_mulsi3_v6): Likewise.
14198 (*thumb_mulsi3): Likewise.
14199 (*thumb_mulsi3_v6): Likewise.
14200 (*mulsi3_compare0): Likewise.
14201 (*mulsi3_compare0_v6): Likewise.
14202 (*mulsi_compare0_scratch): Likewise.
14203 (*mulsi_compare0_scratch_v6): Likewise.
14204 (*mulsi3addsi): Likewise.
14205 (*mulsi3addsi_v6): Likewise.
14206 (*mulsi3addsi_compare0): Likewise.
14207 (*mulsi3addsi_compare0_v6): Likewise.
14208 (*mulsi3addsi_compare0_scratch): Likewise.
14209 (*mulsi3addsi_compare0_scratch_v6): Likewise.
14210 (*mulsi3subsi): Likewise.
14211 (*mulsidi3adddi): Likewise.
14212 (*mulsi3addsi_v6): Likewise.
14213 (*mulsidi3adddi_v6): Likewise.
14214 (*mulsidi3_nov6): Likewise.
14215 (*mulsidi3_v6): Likewise.
14216 (*umulsidi3_nov6): Likewise.
14217 (*umulsidi3_v6): Likewise.
14218 (*umulsidi3adddi): Likewise.
14219 (*umulsidi3adddi_v6): Likewise.
14220 (*smulsi3_highpart_nov6): Likewise.
14221 (*smulsi3_highpart_v6): Likewise.
14222 (*umulsi3_highpart_nov6): Likewise.
14223 (*umulsi3_highpart_v6): Likewise.
14224 (mulhisi3): Likewise.
14225 (*mulhisi3tb): Likewise.
14226 (*mulhisi3bt): Likewise.
14227 (*mulhisi3tt): Likewise.
14228 (maddhisi4): Likewise.
14229 (*maddhisi4tb): Likewise.
14230 (*maddhisi4tt): Likewise.
14231 (maddhidi4): Likewise.
14232 (*maddhidi4tb): Likewise.
14233 (*maddhidi4tt): Likewise.
14234 (divsi3): Likewise.
14235 (udivsi3): Likewise.
14236 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
14237 (thumb2_mulsi_short_compare0): Likewise.
14238 (thumb2_mulsi_short_compare0_scratch): Likewise.
14239 * config/arm/arm1020e.md (1020mult1): Update attribute change.
14240 (1020mult2): Likewise.
14241 (1020mult3): Likewise.
14242 (1020mult4): Likewise.
14243 (1020mult5): Likewise.
14244 (1020mult6): Likewise.
14245 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
14247 (cortex_a15_mult64): Likewise.
14248 (cortex_a15_sdiv): Likewise.
14249 (cortex_a15_udiv): Likewise.
14250 * config/arm/arm1026ejs.md (mult1): Update attribute change.
14256 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
14257 (pj4_ir_div): Likewise.
14258 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
14259 (11_mult2): Likewise.
14260 (11_mult3): Likewise.
14261 (11_mult4): Likewise.
14262 (11_mult5): Likewise.
14263 (11_mult6): Likewise.
14264 (11_mult7): Likewise.
14265 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
14266 (cortex_a8_mla): Likewise.
14267 (cortex_a8_mull): Likewise.
14268 (cortex_a8_smulwy): Likewise.
14269 (cortex_a8_smlald): Likewise.
14270 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
14271 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
14272 (cortex_r4_mul_3): Likewise.
14273 (cortex_r4_mla_4): Likewise.
14274 (cortex_r4_mla_3): Likewise.
14275 (cortex_r4_smlald): Likewise.
14276 (cortex_r4_mull): Likewise.
14277 (cortex_r4_sdiv): Likewise.
14278 (cortex_r4_udiv): Likewise.
14279 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
14280 (cortex_a7_idiv): Likewise.
14281 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
14282 (9_mult2): Likewise.
14283 (9_mult3): Likewise.
14284 (9_mult4): Likewise.
14285 (9_mult5): Likewise.
14286 (9_mult6): Likewise.
14287 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
14288 (cortex_a53_sdiv): Likewise.
14289 (cortex_a53_udiv): Likewise.
14290 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
14291 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
14292 (mp626_mult2): Likewise.
14293 (mp626_mult3): Likewise.
14294 (mp626_mult4): Likewise.
14295 * config/arm/fa526.md (526_mult1): Update attribute change.
14296 (526_mult2): Likewise.
14297 * config/arm/arm-generic.md (mult): Update attribute change.
14298 (mult_ldsched_strongarm): Likewise.
14299 (mult_ldsched): Likewise.
14300 (multi_cycle): Likewise.
14301 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
14302 * config/arm/fa606te.md (606te_mult1): Update attribute change.
14303 (606te_mult2): Likewise.
14304 (606te_mult3): Likewise.
14305 (606te_mult4): Likewise.
14306 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
14307 (cortex_a9_mac16): Likewise.
14308 (cortex_a9_multiply): Likewise.
14309 (cortex_a9_mac): Likewise.
14310 (cortex_a9_multiply_long): Likewise.
14311 * config/arm/fa626te.md (626te_mult1): Update attribute change.
14312 (626te_mult2): Likewise.
14313 (626te_mult3): Likewise.
14314 (626te_mult4): Likewise.
14316 2013-06-18 Richard Biener <rguenther@suse.de>
14319 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
14321 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14324 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14325 with next_active_insn.
14327 2013-06-18 Alan Modra <amodra@gmail.com>
14329 * config/rs6000/rs6000.h (enum data_align): New.
14330 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
14331 (DATA_ABI_ALIGNMENT): Define.
14332 (CONSTANT_ALIGNMENT): Correct comment.
14333 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
14334 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
14336 2013-06-17 David Malcolm <dmalcolm@redhat.com>
14338 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
14339 ATTRIBUTE_UNUSED marking.
14341 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
14343 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
14344 alternative and update.
14345 (aarch64_dup_lanedi): Delete.
14346 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
14347 * config/aarch64/aarch64-simd-builtins.def: Update.
14349 2013-06-17 Richard Biener <rguenther@suse.de>
14351 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
14352 (lto_input_scc): Declare.
14353 (lto_input_tree_1): Likewise.
14354 (struct lto_stats_d): Add num_tree_bodies_output and
14355 num_pickle_refs_output.
14356 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
14357 (lto_read_tree_1): Split out from ...
14358 (lto_read_tree): ... this.
14359 (lto_input_scc): New function.
14360 (lto_input_tree_1): Split out from ...
14361 (lto_input_tree): ... this. Handle LTO_tree_scc.
14362 (lto_data_in_create): Create the streamer cache without hashes.
14363 * lto-streamer-out.c (create_output_block): Create the streamer
14364 cache with hashes when not doing WPA.
14365 (lto_write_tree_1): Split out from ...
14366 (lto_write_tree): ... this.
14367 (get_symbol_initial_value): New function.
14368 (lto_output_tree_1): Split out from ...
14369 (lto_output_tree): ... this. Write trees as series of SCCs
14370 using a DFS walk via DFS_write_tree.
14371 (struct sccs, struct scc_entry): New types.
14372 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
14373 (DFS_write_tree_body): New function.
14374 (DFS_write_tree): Likewise.
14375 (hash_tree): Likewise.
14376 (scc_entry_compare): Likewise.
14377 (hash_scc): Likewise.
14378 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
14379 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
14380 TREE_CHAIN as regular reference.
14381 (streamer_read_integer_cst): Remove.
14382 (streamer_get_pickled_tree): Adjust.
14383 * tree-streamer-out.c (streamer_write_chain): Disable streaming
14384 of DECL_EXTERNALs in BLOCK_VARS for now.
14385 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
14387 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
14388 Add hash value argument and record that if hashes are recorded
14390 (streamer_tree_cache_insert_1): Adjust.
14391 (streamer_tree_cache_insert): Likewise.
14392 (streamer_tree_cache_insert_at): Rename to ...
14393 (streamer_tree_cache_replace_tree): ... this and adjust.
14394 (streamer_tree_cache_append): Adjust.
14395 (record_common_node): Likewise.
14396 (streamer_tree_cache_create): Add argument whether to
14397 record hash values together with trees.
14398 (streamer_tree_cache_delete): Adjust.
14399 * tree-streamer.h (struct streamer_tree_cache_d): Add
14401 (streamer_read_integer_cst): Remove.
14402 (streamer_tree_cache_insert): Adjust.
14403 (streamer_tree_cache_append): Likewise.
14404 (streamer_tree_cache_insert_at): Rename to ...
14405 (streamer_tree_cache_replace_tree): ... this and adjust.
14406 (streamer_tree_cache_create): Add argument whether to record hashes.
14407 (streamer_tree_cache_get): Rename to ...
14408 (streamer_tree_cache_get_tree): ... this.
14409 (streamer_tree_cache_get_hash): New function.
14410 * tree.c (cache_integer_cst): New function.
14411 * tree.h (cache_integer_cst): Declare.
14412 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
14413 * lto-symtab.c (lto_varpool_replace_node): Only release
14414 DECL_INITIAL of non-prevailing decls.
14415 * varpool.c (varpool_remove_initializer): Do not release
14416 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
14418 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14420 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
14421 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
14422 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
14423 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
14424 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
14425 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
14426 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
14427 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
14428 instead of TARGET_64BIT.
14429 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
14430 Require ISA_HAS_<D>DIV.
14432 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
14434 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
14435 (mips*-*-linux*): Move default with_llsc setting to where other
14437 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
14439 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
14440 Likewise. Remove default with_tune setting. Move default float
14441 setting to its own block. Handle with_llsc in the same block as above.
14443 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
14445 PR rtl-optimization/57425
14446 PR rtl-optimization/57569
14447 * alias.c (write_dependence_p): Add new parameters mem_mode,
14448 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
14449 Changed all callers.
14450 (canon_anti_dependence): New function.
14451 * cse.c (check_dependence): Use canon_anti_dependence.
14452 * cselib.c (cselib_invalidate_mem): Likewise.
14453 * rtl.h (canon_anti_dependence): Declare.
14455 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14457 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
14458 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
14459 ".set mips3" for 64-bit targets.
14461 2013-06-15 Dehao Chen <dehao@google.com>
14463 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
14464 * gimple-low.c (gimple_check_call_matching_types): Likewise.
14465 (gimple_check_call_args): Likewise.
14466 * value-prof.c (check_ic_target): Likewise.
14467 * ipa-inline.c (early_inliner): Likewise.
14468 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
14469 * cgraph.c (cgraph_create_edge_1): Likewise.
14470 (cgraph_make_edge_direct): Likewise.
14472 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
14475 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
14476 rs6000_output_move_128bit to handle emitting quad memory
14477 operations. Set attribute length to 8 bytes.
14479 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
14481 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
14483 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
14484 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
14485 (aarch64_<su>mlsl<mode>): Likewise.
14487 2013-06-14 Mike Stump <mikestump@comcast.net>
14489 * Makefile.in (TARGET_H): Add insn-codes.h.
14491 2013-06-14 Alan Modra <amodra@gmail.com>
14493 PR middle-end/57134
14494 PR middle-end/57586
14495 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
14496 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
14497 bitfield expansion when EXPAND_MEMORY.
14498 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
14500 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
14502 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
14503 test for clearing quad memory on 32-bit later.
14505 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14507 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
14508 (fold_negate_expr): Likewise.
14509 (fold_real_zero_addition_p): Handle vectors.
14510 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
14512 2013-06-14 Alan Modra <amodra@gmail.com>
14514 * varasm.c (force_const_mem): Revert 2013-06-07 change.
14516 2013-06-13 Jan Hubicka <jh@suse.cz>
14518 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
14519 Local comdats are not externally visible.
14520 * symtab.c (dump_symtab_base): Dump externally visible.
14521 (verify_symtab_base): Verify back links in the symtab hash.
14523 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14525 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
14526 CONVERT_EXPR as equal nodes.
14528 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14530 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
14532 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14534 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
14535 Generalize to complex and vector.
14536 * tree.c (build_all_ones_cst): New function.
14537 * tree.h (build_all_ones_cst): Declare it.
14539 2013-06-13 Alan Modra <amodra@gmail.com>
14541 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
14542 * config/rs6000/rs6000.md (signbittf2): New insn.
14543 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
14544 (abstf2_internal, cmptf_internal2): Likewise.
14545 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
14547 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
14548 Pat Haugen <pthaugen@us.ibm.com>
14549 Peter Bergner <bergner@vnet.ibm.com>
14551 * config/rs6000/rs6000.c (emit_load_locked): Add support for
14552 power8 byte, half-word, and quad-word atomic instructions.
14553 (emit_store_conditional): Likewise.
14554 (rs6000_expand_atomic_compare_and_swap): Likewise.
14555 (rs6000_expand_atomic_op): Likewise.
14557 * config/rs6000/sync.md (larx): Add new modes for power8.
14559 (AINT): New mode iterator to include TImode as well as normal
14560 integer modes on power8.
14561 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
14562 that VSX registers are not considered. Use AINT mode iterator
14563 instead of INT1 to allow inclusion of quad word atomic operations
14565 (load_locked<mode>): Likewise.
14566 (store_conditional<mode>): Likewise.
14567 (atomic_compare_and_swap<mode>): Likewise.
14568 (atomic_exchange<mode>): Likewise.
14569 (atomic_nand<mode>): Likewise.
14570 (atomic_fetch_<fetchop_name><mode>): Likewise.
14571 (atomic_nand_fetch<mode>): Likewise.
14572 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
14574 (ATOMIC): On power8, add QImode, HImode modes.
14575 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
14576 modes that promote to SImode.
14577 (load_lockedti): Convert TImode arguments to PTImode, so that we
14578 get a guaranteed even/odd register pair.
14579 (load_lockedpti): Likewise.
14580 (store_conditionalti): Likewise.
14581 (store_conditionalpti): Likewise.
14583 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
14584 atomic load/store instructions.
14587 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
14589 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
14591 (insn_count): New attribute, with most cases extracted from...
14592 (length): ...here. Redefine most cases in terms of insn_count.
14593 (single_insn): Delete.
14594 (can_delay): Use insn_count to check for single instructions.
14595 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
14596 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
14597 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
14598 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
14599 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
14600 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
14601 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
14602 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
14603 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
14604 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
14605 rather than "length".
14606 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
14607 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
14608 Use "insn_count" rather than "length".
14609 * config/mips/mips-dsp.md
14610 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
14611 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
14614 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
14616 PR tree-optimization/57361
14617 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
14619 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
14621 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
14623 (aarch64_simd_combine<mode>): New instruction expansion.
14624 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
14625 function prototype.
14626 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
14627 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
14629 2013-06-12 Jan Hubicka <jh@suse.cz>
14631 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
14632 decl has when in streaming stage.
14633 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
14634 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
14636 2013-06-12 Roland Stigge <stigge@antcom.de>
14639 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
14641 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14643 PR tree-optimization/57537
14644 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
14645 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
14647 2013-06-12 Richard Biener <rguenther@suse.de>
14649 * data-streamer.h (streamer_write_char_stream): CSE
14650 obs->current_pointer.
14651 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
14652 streamer_write_char_stream manually and optimize the resulting loop.
14653 (streamer_write_hwi_stream): Likewise.
14655 2013-06-12 Jan Hubicka <jh@suse.cz>
14657 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
14658 * cgraph.h (varpool_create_empty_node): Declare.
14659 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
14661 * symtab.c (symtab_unregister_node): Be lax about missin entries
14663 (symtab_get_node): Update comment.
14664 * varpool.c (varpool_create_empty_node): Break out from ...
14665 (varpool_node_for_decl): ... here.
14666 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
14668 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
14670 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
14671 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
14672 part. Use straight-line flow at the end.
14673 <COMPONENT_REF>: Remove superfluous else.
14674 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
14676 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14679 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
14680 target hook even for !TREE_PUBLIC decls. If no resolution info
14681 is available, return false for common and external decls.
14683 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14685 * config/rl78/constraints.md (U): New constraint.
14686 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
14689 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
14692 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
14693 to allow returning address to AT_PLATFORM name.
14695 2013-06-11 Jan Hubicka <jh@suse.cz>
14697 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
14698 * cgraph.h (symtab_node_base): Add weakref flag.
14699 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
14700 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
14701 (output_weakrefs): Use weakref flag.
14702 * fold-const.c (simple_operand_p): Handle WEAK.
14703 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
14704 * ipa.c (varpool_externally_visible_p): Drop weakref.
14705 (function_and_variable_visibility): Update comment; fix weakref
14706 sanity checks; do not clear DECL_WEAK on them.
14707 * lto-cgraph.c (lto_output_node): update.
14708 (lto_output_varpool_node): Update.
14709 (input_overwrite_node): Update.
14710 (input_node): Update.
14711 (input_varpool_node): Update.
14712 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
14713 (lto_symtab_merge_symbols): Add sanity check.
14714 (lto_symtab_prevailing_decl): Do not special case weakrefs.
14715 * passes.c (rest_of_decl_compilation): Set static flag, too.
14716 * symtab.c (dump_symtab_base): Dump weakref.
14717 (verify_symtab_base): Sanity check weakrefs.
14718 (symtab_make_decl_local): Remove duplicated code.
14719 (symtab_alias_ultimate_target): Simplify.
14720 * varpool.c (varpool_create_variable_alias): Set weakref flag.
14722 2013-06-11 Tom de Vries <tom@codesourcery.com>
14724 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
14725 sequence_vect == NULL.
14727 2013-06-11 DJ Delorie <dj@redhat.com>
14729 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
14730 (rl78_unwind_word_mode): New.
14732 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14734 * final.c (debug_prefix_maps): Make static.
14736 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14738 * function.c (initial_trampoline): Remove stray copy.
14740 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
14742 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
14744 2013-06-11 Martin Jambor <mjambor@suse.cz>
14746 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
14747 within bounds at the beginning of the function.
14749 2013-06-11 Alan Modra <amodra@gmail.com>
14751 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
14753 (get_named_section): Don't NULL !DECL_P decl.
14755 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
14757 * doc/invoke.texi (core-avx2): Document.
14759 (atom): Updated with MOVBE.
14761 2013-06-11 Richard Biener <rguenther@suse.de>
14763 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
14765 2013-06-11 Anton Blanchard <anton@samba.org>
14767 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
14768 correct shift value in little-endian mode.
14770 2013-06-11 Jakub Jelinek <jakub@redhat.com>
14773 * varasm.c (get_variable_align): Move #endif to the right place.
14775 2013-06-10 Cary Coutant <ccoutant@google.com>
14777 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
14778 for hash so that hash table traversal order is deterministic.
14780 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
14781 Pat Haugen <pthaugen@us.ibm.com>
14782 Peter Bergner <bergner@vnet.ibm.com>
14784 * config/rs6000/vector.md (GPR move splitter): Do not split moves
14785 of vectors in GPRS if they are direct moves or quad word load or
14788 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
14790 (direct_move_p): Likewise.
14791 (quad_load_store_p): Likewise.
14793 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
14794 classes into bins based on the physical register type.
14795 (reg_class_to_reg_type): Likewise.
14796 (IS_STD_REG_TYPE): Likewise.
14797 (IS_FP_VECT_REG_TYPE): Likewise.
14798 (reload_fpr_gpr): Arrays to determine what insn to use if we can
14799 use direct move instructions.
14800 (reload_gpr_vsx): Likewise.
14801 (reload_vsx_gpr): Likewise.
14802 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
14803 information that is a simplification of register classes. Also
14804 precalculate direct move reload helpers.
14805 (direct_move_p): New function to return true if the operation can
14806 be done as a direct move instruciton.
14807 (quad_load_store_p): New function to return true if the operation
14808 is a quad memory operation.
14809 (rs6000_legitimize_address): If quad memory, only allow register
14810 indirect for TImode addresses.
14811 (rs6000_legitimate_address_p): Likewise.
14812 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
14813 (rs6000_reload_register_type): Likewise.
14814 (register_to_reg_type): Return register type.
14815 (rs6000_secondary_reload_simple_move): New helper function for
14816 secondary reload and secondary memory needed to identify anything
14817 that is a simple move, and does not need reloading.
14818 (rs6000_secondary_reload_direct_move): New helper function for
14819 secondary reload to identify cases that can be done with several
14820 instructions via the direct move instructions.
14821 (rs6000_secondary_reload_move): New helper function for secondary
14822 reload to identify moves between register types that can be done.
14823 (rs6000_secondary_reload): Add support for quad memory operations
14824 and for direct move.
14825 (rs6000_secondary_memory_needed): Likewise.
14826 (rs6000_debug_secondary_memory_needed): Change argument names.
14827 (rs6000_output_move_128bit): New function to return the move to
14828 use for 128-bit moves, including knowing about the various
14829 limitations of quad memory operations.
14831 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
14832 memory operations. call rs6000_output_move_128bit for the actual
14833 instruciton(s) to generate.
14834 (vsx_movti_64bit): Likewise.
14836 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
14837 (UNSPEC_P8V_MTVSRWZ): Likewise.
14838 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
14839 (UNSPEC_P8V_MTVSRD): Likewise.
14840 (UNSPEC_P8V_XXPERMDI): Likewise.
14841 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
14842 (UNSPEC_FUSION_GPR): Likewise.
14843 (FMOVE128_GPR): New iterator for direct move.
14844 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
14845 (f32_sv): Likewise.
14846 (f32_dm): Likewise.
14847 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
14848 loads and direct move instructions.
14849 (zero_extendsidi2_lfiwzx): Likewise.
14850 (extendsidi2_lfiwax): Likewise.
14851 (extendsidi2_nocell): Likewise.
14852 (floatsi<mode>2_lfiwax): Likewise.
14853 (lfiwax): Likewise.
14854 (floatunssi<mode>2_lfiwzx): Likewise.
14855 (lfiwzx): Likewise.
14856 (fix_trunc<mode>_stfiwx): Likewise.
14857 (fixuns_trunc<mode>_stfiwx): Likewise.
14858 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
14859 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
14860 (parity<mode>2_cmpb): Set length/type attr.
14861 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
14862 for 'mr.' to fast_compare.
14863 (bpermd_<mode>): Change type attr to popcnt.
14864 (p8_fmrgow_<mode>): New insns for power8 direct move support.
14865 (p8_mtvsrwz_1): Likewise.
14866 (p8_mtvsrwz_2): Likewise.
14867 (reload_fpr_from_gpr<mode>): Likewise.
14868 (p8_mtvsrd_1): Likewise.
14869 (p8_mtvsrd_2): Likewise.
14870 (p8_xxpermdi_<mode>): Likewise.
14871 (reload_vsx_from_gpr<mode>): Likewise.
14872 (reload_vsx_from_gprsf): Likewise.
14873 (p8_mfvsrd_3_<mode>): LIkewise.
14874 (reload_gpr_from_vsx<mode>): Likewise.
14875 (reload_gpr_from_vsxsf): Likewise.
14876 (p8_mfvsrd_4_disf): Likewise.
14877 (multi-word GPR splits): Do not split direct moves or quad memory
14880 2013-06-10 David Malcolm <dmalcolm@redhat.com>
14882 * tree-into-ssa.c (interesting_blocks): Make static.
14884 2013-06-10 Jakub Jelinek <jakub@redhat.com>
14887 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
14888 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
14889 Use DATA_ABI_ALIGNMENT for that case instead if defined.
14890 (get_variable_align): New function.
14891 (get_variable_section, emit_bss, emit_common,
14892 assemble_variable_contents, place_block_symbol): Use
14893 get_variable_align instead of DECL_ALIGN.
14894 (assemble_noswitch_variable): Add align argument, use it
14895 instead of DECL_ALIGN.
14896 (assemble_variable): Adjust caller. Use get_variable_align
14897 instead of DECL_ALIGN.
14898 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
14900 (DATA_ABI_ALIGNMENT): Define.
14901 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
14902 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
14903 opt is false, only return the psABI mandated alignment increase.
14904 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
14905 (DATA_ABI_ALIGNMENT): ... this.
14906 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
14907 (DATA_ABI_ALIGNMENT): ... this.
14908 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
14909 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
14910 (DATA_ABI_ALIGNMENT): ... this.
14911 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
14912 * doc/tm.texi: Regenerated.
14914 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
14916 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
14917 cmp_code to construct REG_EQUAL note.
14919 2013-06-09 Jakub Jelinek <jakub@redhat.com>
14922 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
14923 that operands[2] doesn't overlap with operands[0].
14925 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
14926 Jan Hubicka <jh@suse.cz>
14928 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
14929 hack to mark symbols as used.
14931 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
14933 PR rtl-optimization/57559
14934 * lra-constraints.c (process_alt_operands): Don't discourage
14935 memory with known offset for offsetable memory constraint.
14936 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
14938 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
14940 * varasm.c (struct oc_local_state): Reorder fields.
14941 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
14942 and adjust accordingly.
14943 (output_constructor): Reorder initialization code and adjust call to
14944 output_constructor_bitfield.
14946 2013-06-07 Jan Hubicka <jh@suse.cz>
14948 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
14950 2013-06-07 David Malcolm <dmalcolm@redhat.com>
14952 * tree-object-size.c (unknown): Make const.
14954 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14956 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
14957 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
14958 for last alternative in the cpu_facility attribute.
14960 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14963 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
14964 (xordi3): Change operand 2 constraint to arm_xordi_operand.
14965 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
14966 * config/arm/constraints.md (Dg): New constraint.
14967 * config/arm/neon.md (xordi3_neon): Remove.
14968 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
14969 * config/arm/predicates.md (arm_xordi_operand): New predicate.
14971 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14973 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
14974 Clean up alternatives.
14976 2013-06-07 Alan Modra <amodra@gmail.com>
14978 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
14981 2013-06-07 Alan Modra <amodra@gmail.com>
14983 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
14985 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14987 * config/arm/constraints.md (Df): New constraint.
14988 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
14989 Correct length attribute for last two alternatives.
14991 2013-06-07 Alan Modra <amodra@gmail.com>
14993 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
14994 override user -mfp-in-toc.
14995 (offsettable_ok_by_alignment): Consider just the current access
14996 rather than the whole object, unless BLKmode. Handle
14997 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
14998 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
14999 for -mcmodel=medium.
15000 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
15001 override user -mfp-in-toc or -msum-in-toc. Default to
15002 -mno-fp-in-toc for -mcmodel=medium.
15004 2013-06-06 DJ Delorie <dj@redhat.com>
15006 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
15007 TARGET_VALID_POINTER_MODE.
15009 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
15010 Pat Haugen <pthaugen@us.ibm.com>
15011 Peter Bergner <bergner@vnet.ibm.com>
15013 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
15014 Document new power8 builtins.
15016 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
15017 condition code register, to allow 128-bit logical operations to be
15018 done in the VSX or GPR registers.
15019 (nor<mode>3): Use the canonical form for nor.
15020 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
15021 vclz*, and vpopcnt* vector instructions.
15022 (nand<mode>3): Likewise.
15023 (orc<mode>3): Likewise.
15024 (clz<mode>2): LIkewise.
15025 (popcount<mode>2): Likewise.
15027 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
15028 that only the GPRs are recognized.
15030 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15031 support for new power8 builtins.
15033 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
15035 (xscvdpspn): Likewise.
15041 (vpopcnt): Likewise.
15042 (vpopcntb): Likewise.
15043 (vpopcnth): Likewise.
15044 (vpopcntw): Likewise.
15045 (vpopcntd): Likewise.
15047 (vmrgew): Likewise.
15048 (vmrgow): Likewise.
15050 (eqv_v16qi3): Likewise.
15051 (eqv_v8hi3): Likewise.
15052 (eqv_v4si3): Likewise.
15053 (eqv_v2di3): Likewise.
15054 (eqv_v4sf3): Likewise.
15055 (eqv_v2df3): Likewise.
15057 (nand_v16qi3): Likewise.
15058 (nand_v8hi3): Likewise.
15059 (nand_v4si3): Likewise.
15060 (nand_v2di3): Likewise.
15061 (nand_v4sf3): Likewise.
15062 (nand_v2df3): Likewise.
15064 (orc_v16qi3): Likewise.
15065 (orc_v8hi3): Likewise.
15066 (orc_v4si3): Likewise.
15067 (orc_v2di3): Likewise.
15068 (orc_v4sf3): Likewise.
15069 (orc_v2df3): Likewise.
15071 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
15072 allow power8 quad mode in 64-bit.
15073 (rs6000_builtin_vectorized_function): Add support to vectorize
15074 ISA 2.07 count leading zeros, population count builtins.
15075 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
15076 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
15077 (builtin_function_type): Add vgbbd builtin function which takes an
15079 (altivec_expand_vec_perm_const): Add support for new power8 merge
15082 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
15083 that does not include TImdoe for use with 32-bit.
15084 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
15086 (UNSPEC_VSX_CVDPSPN): Likewise.
15087 (vsx_xscvdpspn): Likewise.
15088 (vsx_xscvspdpn): Likewise.
15089 (vsx_xscvdpspn_scalar): Likewise.
15090 (vsx_xscvspdpn_directmove): Likewise.
15091 (vsx_and<mode>3): Split logical operations into 32-bit and
15092 64-bit. Add support to do logical operations on TImode as well as
15093 VSX vector types. Allow logical operations to be done in either
15094 VSX registers or in general purpose registers in 64-bit mode. Add
15095 splitters if GPRs were used. For AND, add clobber of CCmode to
15096 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
15098 (vsx_and<mode>3_32bit): Likewise.
15099 (vsx_and<mode>3_64bit): Likewise.
15100 (vsx_ior<mode>3): Likewise.
15101 (vsx_ior<mode>3_32bit): Likewise.
15102 (vsx_ior<mode>3_64bit): Likewise.
15103 (vsx_xor<mode>3): Likewise.
15104 (vsx_xor<mode>3_32bit): Likewise.
15105 (vsx_xor<mode>3_64bit): Likewise.
15106 (vsx_one_cmpl<mode>2): Likewise.
15107 (vsx_one_cmpl<mode>2_32bit): Likewise.
15108 (vsx_one_cmpl<mode>2_64bit): Likewise.
15109 (vsx_nor<mode>3): Likewise.
15110 (vsx_nor<mode>3_32bit): Likewise.
15111 (vsx_nor<mode>3_64bit): Likewise.
15112 (vsx_andc<mode>3): Likewise.
15113 (vsx_andc<mode>3_32bit): Likewise.
15114 (vsx_andc<mode>3_64bit): Likewise.
15115 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
15116 and xxlorc instructions.
15117 (vsx_eqv<mode>3_64bit): Likewise.
15118 (vsx_nand<mode>3_32bit): Likewise.
15119 (vsx_nand<mode>3_64bit): Likewise.
15120 (vsx_orc<mode>3_32bit): Likewise.
15121 (vsx_orc<mode>3_64bit): Likewise.
15123 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
15125 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
15127 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
15128 (p8_vmrgow): Likewise.
15129 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
15130 GPRs to be split under VSX.
15131 (p8v_clz<mode>2): Add power8 count leading zero support.
15132 (p8v_popcount<mode>2): Add power8 population count support.
15133 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
15136 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
15139 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
15141 (vec_nand): Likewise.
15142 (vec_vclz): Likewise.
15143 (vec_vclzb): Likewise.
15144 (vec_vclzd): Likewise.
15145 (vec_vclzh): Likewise.
15146 (vec_vclzw): Likewise.
15147 (vec_vgbbd): Likewise.
15148 (vec_vmrgew): Likewise.
15149 (vec_vmrgow): Likewise.
15150 (vec_vpopcnt): Likewise.
15151 (vec_vpopcntb): Likewise.
15152 (vec_vpopcntd): Likewise.
15153 (vec_vpopcnth): Likewise.
15154 (vec_vpopcntw): Likewise.
15156 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15158 PR rtl-optimization/57468
15159 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
15162 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15164 PR rtl-optimization/57459
15165 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
15166 type when setting live regs.
15168 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15170 * config/s390/s390.opt (mlra): New option.
15171 * config/s390/s390.c (s390_decompose_address): Check displacement
15172 for all registers for LRA.
15173 (s390_secondary_reload): Don't used secondary reloads for LRA.
15174 (s390_lra_p): New function.
15175 (TARGET_LRA_P): Define.
15176 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
15177 of attribute cpu_facility to zarch for the last alternative.
15178 (*cmpmem_short): Ditto.
15180 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
15182 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
15183 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
15184 (arm_expand_prologue): Likewise.
15186 2013-06-06 Teresa Johnson <tejohnson@google.com>
15189 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
15190 as this is now done by redirect_edge_and_branch_force.
15191 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
15192 barriers, and fix interaction with splitting.
15193 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
15194 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
15195 reflect changes made in the routine.
15196 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
15197 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
15198 since this is called in cfglayout mode, and replace partition fixup
15199 with assert as that is now done by force_nonfallthru_and_redirect.
15200 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
15201 already be marked with region crossing note.
15202 (insert_section_boundary_note): Make non-static, gate on flag
15203 has_bb_partition, rewrite to also check for multiple partitions.
15204 (rest_of_handle_reorder_blocks): Remove call to
15205 insert_section_boundary_note, now done later during free_cfg.
15206 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
15207 * bb-reorder.h (insert_section_boundary_note): Declare.
15208 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
15209 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
15210 invoke insert_section_boundary_note.
15211 (try_redirect_by_replacing_jump): Remove unnecessary
15212 check for region crossing note.
15213 (fixup_partition_crossing): New function.
15214 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
15215 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
15216 in non-cfglayout mode.
15217 (force_nonfallthru_and_redirect): Fixup partition boundaries,
15218 remove old code that tried to do this. Emit barrier correctly
15219 when we are in cfglayout mode.
15220 (last_bb_in_partition): New function.
15221 (rtl_split_edge): Correctly fixup partition boundaries.
15222 (commit_one_edge_insertion): Remove old code that tried to
15223 fixup region crossing edge since this is now handled in
15224 split_block, and set up insertion point correctly since
15225 block may now end in a jump.
15226 (verify_hot_cold_block_grouping): Guard against checking when not in
15227 linearized RTL mode.
15228 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
15230 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
15231 rtl_verify_flow_info, so not called in cfglayout mode.
15232 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
15233 (fixup_reorder_chain): Remove old code that attempted to fixup region
15234 crossing note as this is now handled in force_nonfallthru_and_redirect.
15235 (duplicate_insn_chain): Don't duplicate switch section notes.
15236 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
15238 * basic-block.h (emit_barrier_after_bb): Declare.
15240 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15242 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
15243 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
15244 arm_usatsihi): Adjust alternatives for arm_restrict_it.
15246 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15248 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
15250 * config/arm/ldmstm.md: Regenerate.
15252 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15254 * config/arm/sync.md (atomic_loaddi_1):
15255 Disable predication for arm_restrict_it.
15256 (arm_load_exclusive<mode>): Likewise.
15257 (arm_load_exclusivesi): Likewise.
15258 (arm_load_exclusivedi): Likewise.
15259 (arm_load_acquire_exclusive<mode>): Likewise.
15260 (arm_load_acquire_exclusivesi): Likewise.
15261 (arm_load_acquire_exclusivedi): Likewise.
15262 (arm_store_exclusive<mode>): Likewise.
15263 (arm_store_exclusive<mode>): Likewise.
15264 (arm_store_release_exclusivedi): Likewise.
15265 (arm_store_release_exclusive<mode>): Likewise.
15267 2013-06-06 Richard Biener <rguenther@suse.de>
15269 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
15271 (lto_tag_is_tree_code_p): Adjust.
15272 (lto_tag_is_gimple_code_p): Likewise.
15273 (lto_gimple_code_to_tag): Likewise.
15274 (lto_tag_to_gimple_code): Likewise.
15275 (lto_tree_code_to_tag): Likewise.
15276 (lto_tag_to_tree_code): Likewise.
15277 * data-streamer.h (streamer_write_hwi_in_range): Use
15278 uhwi streaming to stream the normalized range.
15279 (streamer_read_hwi_in_range): Likewise.
15281 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15283 * config/arm/arm.md (enabled_for_depr_it): New attribute.
15284 (predicable_short_it): Likewise.
15285 (predicated): Likewise.
15286 (enabled): Handle above.
15287 (define_cond_exec): Set predicated attribute to yes.
15289 2013-06-05 Mike Stump <mikestump@comcast.net>
15291 * gdbinit.in (__FUNCTION__): Add.
15293 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
15295 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
15296 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
15298 2013-06-05 Jan Hubicka <jh@suse.cz>
15300 * varasm.c (mark_decl_referenced): Revert the removal until targets
15303 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
15305 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
15306 instead of mark_decl_referenced.
15308 2013-06-05 Jan Hubicka <jh@suse.cz>
15310 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
15311 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
15312 and symtab_used_from_object_file_p.
15313 (cgraph_make_node_local_1): Clear forced_by_abi.
15314 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
15315 * cgraph.h (symtab_node_base): Add forced_by_abi.
15316 (decide_is_variable_needed): Remove.
15317 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
15318 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
15319 (decide_is_symbol_needed): ... this one; handle symbols in general;
15320 always analyze virtuals; honnor forced_by_abi.
15321 (cgraph_finalize_function): Update.
15322 (varpool_finalize_decl): Update.
15323 (symbol_defined_and_needed): Remove.
15324 (analyze_functions): Update.
15325 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
15326 output_refs, input_overwrite_node): Handle forced_by_abi.
15327 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
15328 (address_taken_from_non_vtable_p): ... this one.
15329 (comdat_can_be_unshared_p_1): New function.
15330 (cgraph_comdat_can_be_unshared_p): Rename to ...
15331 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
15332 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
15333 (function_and_variable_visibility): Clear forced_by_abi as needed.
15334 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
15335 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
15336 * symtab.c (dump_symtab_base): Dump forced_by_abi.
15337 * varpool.c (decide_is_variable_needed): Remove.
15339 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15341 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
15342 (arm_option_override): Override arm_restrict_it where appropriate.
15343 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
15344 * config/arm/arm.opt (mrestrict-it): New command-line option.
15345 * doc/invoke.texi: Document -mrestrict-it.
15347 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15349 * tsan.c (tsan_atomic_table): Make const.
15351 2013-06-05 Richard Biener <rguenther@suse.de>
15353 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
15354 index associated with the tree we are supposed to replace.
15355 * tree-streamer-out.c (pack_ts_base_value_fields): Output
15356 TREE_ASM_WRITTEN as zero for everything but SSA names.
15358 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15360 * tree-ssa-structalias.c (call_stmt_vars): Make static.
15362 2013-06-04 Jan Hubicka <jh@suse.cz>
15364 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
15365 (input_node, input_varpool_node): Handle correctly external same
15367 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
15368 nodes at ltrans stage.
15370 2013-06-04 Jan Hubicka <jh@suse.cz>
15372 * ipa-inline.c (update_caller_keys): Fix availability test.
15373 (update_callee_keys): Likewise.
15374 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
15375 to follow ELF standard.
15377 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
15379 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
15380 (mips64r5900el-*-elf*): New configurations.
15381 * config/mips/mips-cpus.def (r5900): New processor.
15382 * config/mips/mips-tables.opt: Regenerate.
15383 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
15384 (mips_issue_rate): Handle PROCESSOR_R5900.
15385 (mips_reorg_process_insns): Force reorder mode for the R5900.
15386 * config/mips/mips.h (TARGET_MIPS5900): Define.
15387 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
15389 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
15391 * config/mips/mips.md (processor): Add r5900.
15392 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
15394 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15396 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
15397 into function to generate MOVI instruction.
15398 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
15399 (aarch64_preferred_simd_mode): Turn into wrapper.
15400 (aarch64_output_scalar_simd_mov_immediate): New function.
15401 * config/aarch64/aarch64-protos.h: Add prototype for above.
15403 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15405 * config/aarch64/aarch64.c (simd_immediate_info): Remove
15406 element_char member.
15407 (sizetochar): Return signed char.
15408 (aarch64_simd_valid_immediate): Remove elchar and other
15409 unnecessary variables.
15410 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
15411 Calculate element_char as required.
15412 * config/aarch64/aarch64-protos.h: Update and move prototype
15413 for aarch64_output_simd_mov_immediate.
15414 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
15417 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15419 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
15420 information completed by aarch64_simd_valid_immediate.
15421 (aarch64_legitimate_constant_p): Update arguments.
15422 (aarch64_simd_valid_immediate): Work with struct rather than many
15424 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
15425 (aarch64_simd_make_constant): Update arguments.
15426 (aarch64_output_simd_mov_immediate): Work with struct rather than
15427 many pointers. Output immediate directly rather than as operand.
15428 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15430 * config/aarch64/constraints.md (Dn): Update arguments.
15432 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15434 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
15436 (aarch64_simd_immediate_valid_for_move): Remove.
15437 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
15438 (aarch64_simd_make_constant): Update call.
15439 (aarch64_output_simd_mov_immediate): Update call.
15440 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15442 * config/aarch64/constraints.md (Dn): Update call.
15444 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15446 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
15447 return type to bool for prototype.
15448 (aarch64_legitimate_constant_p): Check for true instead of not -1.
15449 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
15450 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
15452 2013-06-04 Catherine Moore <clm@codesourcery.com>
15454 * config/mips/mips.opt (meva): New.
15455 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
15456 (ASM_SPEC): Handle -meva.
15457 * doc/invoke.texi (meva): Document.
15459 2013-06-04 Alan Modra <amodra@gmail.com>
15461 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
15464 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15466 * rtl.def: Add extra fourth optional field to define_cond_exec.
15467 * gensupport.c (process_one_cond_exec): Process attributes from
15469 * doc/md.texi: Document fourth field in define_cond_exec.
15471 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
15473 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
15474 out the processing order as in store_bit_field_1.
15476 2013-06-04 Jan Hubicka <jh@suse.cz>
15478 PR middle-end/57500
15479 * cgraphunit.c (cgraph_process_same_body_aliases): Create
15480 non-VAR_DECL node if it does not exist yet.
15482 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
15484 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
15485 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
15486 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
15487 target_cpu_default setting.
15489 2013-06-03 Teresa Johnson <tejohnson@google.com>
15491 * dumpfile.c (opt_info_switch_p): Change -fopt-info
15492 default to -fopt-info=optimized instead of all.
15493 * doc/invoke.texi: Ditto.
15494 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
15495 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
15496 (execute_vect_slp): Emit BB vectorization success under
15497 MSG_OPTIMIZED_LOCATIONS.
15498 * tree-vect-slp.c (vect_slp_transform_bb): Change
15499 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
15500 * tree-vect-loop.c (vect_transform_loop): Ditto.
15502 2013-06-03 Jason Merrill <jason@redhat.com>
15505 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
15506 Use TARGET_EXPR for C++.
15508 2013-06-03 Jakub Jelinek <jakub@redhat.com>
15510 PR rtl-optimization/57268
15511 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
15512 if DEBUG_INSN_P (insn).
15515 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15517 PR rtl-optimization/57268
15518 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15519 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15521 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15523 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
15524 (ix86_avoid_lea_for_addr): Likewise.
15525 (exact_dependency_1): Likewise.
15526 (ix86_adjust_cost): Likewise.
15527 (swap_top_of_ready_list): Fix formatting and !reload_completed check
15529 (do_reorder_for_imul): Fix typo, formatting and
15530 !reload_completed check removed.
15531 (ix86_sched_reorder): Fix typo and formatting.
15532 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
15535 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
15537 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
15539 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
15541 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
15542 <STRING_CST>: Likewise.
15543 <VECTOR_CST>: Likewise.
15545 2013-06-01 Janus Weil <janus@gcc.gnu.org>
15546 Mikael Morin <mikael@gcc.gnu.org>
15548 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
15549 * config.in: Regenerated.
15550 * configure: Regenerated.
15552 2013-06-01 Jan Hubicka <jh@suse.cz>
15554 PR middle-end/57366
15555 * cgraphunit.c (compile): When weakref is not supported,
15556 set up transparent aliases before final output pass.
15557 * varasm.c (assemble_alias): Do not try to do it here.
15559 2013-06-01 Jan Hubicka <jh@suse.cz>
15561 PR middle-end/57467
15562 * passes.c (for_per_function): Skip unanalyzed functions.
15564 2013-06-01 Jan Hubicka <jh@suse.cz>
15566 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
15567 (lto_symtab_merge_symbols_1): ... this one.
15568 (lto_symtab_merge_cgraph_nodes): Rename to ...
15569 (lto_symtab_merge_symbols): ... this one; simplify.
15570 * cgraph.c (same_body_aliases_done): Rename to ...
15571 (cpp_implicit_aliases_done): ... this one.
15572 (cgraph_create_function_alias): Update.
15573 (cgraph_same_body_alias): Update.
15574 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
15575 (verify_edge_corresponds_to_fndecl): Simplify.
15576 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
15577 (cgraph_node): Remove same_body_alias.
15578 (varpool_node): Remove alias_of and extra_name_alias.
15579 (same_body_aliases_done): Rename to ..
15580 (cpp_implicit_aliases_done): ... this one.
15581 (symtab_alias_ultimate_target): Add default parameter.
15582 (symtab_resolve_alias): New function.
15583 (fixup_same_cpp_alias_visibility): Declare.
15584 (cgraph_function_node): Add default parameter.
15585 (cgraph_node_asm_name): Likewise.
15586 (cgraph_function_or_thunk_node): Add default parameter; do
15587 not ICE when it is NULL.
15588 (varpool_variable_node): Likewise.
15589 * tree-emutls.c (create_emultls_var): Update.
15590 (ipa_lower_emutls): Update.
15591 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
15592 (cgraph_reset_node): Reset alias info.
15593 (cgraph_finalize_function): Update.
15594 (fixup_same_cpp_alias_visibility): Move to symtab.c.
15595 (analyze_function): Simplify.
15596 (cgraph_process_same_body_aliases): Simplify.
15597 (analyze_functions): Fixup same body aliases.
15598 (handle_alias_pairs): Simplify.
15599 (assemble_thunk): Update.
15600 (assemble_thunks_and_aliases): Update.
15601 (output_weakrefs): Rewrite.
15602 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
15603 (lto_output_varpool_node): Likewise.
15604 (compute_ltrans_boundary): Remve assert.
15605 (get_alias_symbol): New functoin.
15606 (input_node): Rewrite alias handling.
15607 (input_varpool_node): Likewise.
15608 * ipa-pure-const.c (propagate_pure_const): Fix formating.
15609 * ipa.c (process_references): Handle weakrefs correctly.
15610 (symtab_remove_unreachable_nodes): Likewise.
15611 * trans-mem.c (get_cg_data): Update.
15612 (ipa_tm_create_version_alias): Update.
15613 (ipa_tm_execute): Update.
15614 * symtab.c (dump_symtab_base): Dump aliases.
15615 (verify_symtab_base): Verify aliases.
15616 (symtab_node_availability): New function.
15617 (symtab_alias_ultimate_target): Simplify.
15618 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
15619 handle all the fixup cases.
15620 (symtab_resolve_alias): New function.
15621 * passes.c (ipa_write_summaries): Handle weakrefs.
15622 * varpool.c (varpool_analyze_node): Simplify.
15623 (assemble_aliases): Update.
15624 (varpool_create_variable_alias): Simplify.
15625 (varpool_extra_name_alias): Simplify.
15626 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
15627 (lto_symtab_merge_symbols): ... this one.
15629 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
15632 PR rtl-optimization/57268
15633 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15634 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15636 2013-06-01 Tobias Burnus <burnus@net-b.de>
15638 Partially reverted:
15639 2013-05-31 Tobias Burnus <burnus@net-b.de>
15641 PR middle-end/57073
15642 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
15645 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15647 PR rtl-optimization/57268
15648 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15649 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15651 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15653 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
15654 unordered comparison operators when -fno-trapping-math is in effect
15656 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
15657 and implement unordered comparison operators properly on the e500.
15659 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15661 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
15662 for constant scalar integers.
15663 (simplify_relational_operation_1): Likewise.
15665 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
15667 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
15668 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
15671 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15672 Igor Zamyatin <igor.zamyatin@intel.com>
15674 Silvermont (SLM) architecture performance tuning.
15675 * config/i386/i386.h (enum ix86_tune_indices): Add
15676 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
15677 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
15679 * config/i386/i386.c (initial_ix86_tune_features)
15680 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
15681 (ix86_lea_outperforms): Handle Silvermont tuning.
15682 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
15684 (ix86_use_lea_for_mov): Likewise.
15685 (ix86_avoid_lea_for_addr): Likewise.
15686 (ix86_lea_for_add_ok): Likewise.
15687 (exact_dependency_1): New function.
15688 (exact_store_load_dependency): Likewise.
15689 (ix86_adjust_cost): Handle Silvermont tuning.
15690 (do_reoder_for_imul): Likewise.
15691 (swap_top_of_ready_list): New function.
15692 (ix86_sched_reorder): Changed to handle Silvermont tuning.
15694 * config/i386/i386.md (peepholes that split memory operand in fp
15697 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15699 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
15700 Remove un-necessary braces.
15702 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15704 * config/aarch64/aarch64.c (aarch64_classify_symbol):
15705 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
15707 2013-05-31 Tobias Burnus <burnus@net-b.de>
15709 PR middle-end/57073
15710 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
15712 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15715 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
15716 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
15717 * config/arm/neon.md (iordi3_neon): Remove.
15718 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
15719 * config/arm/predicates.md (imm_for_neon_logic_operand):
15720 Move to earlier in the file.
15721 (neon_logic_op2): Likewise.
15722 (arm_iordi_operand_neon): New predicate.
15724 2013-05-31 Richard Biener <rguenther@suse.de>
15726 PR tree-optimization/57478
15727 PR tree-optimization/57453
15728 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
15731 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15733 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
15734 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
15736 2013-05-30 Tobias Burnus <burnus@net-b.de>
15737 Thomas Koenig <tkoenig@gcc.gnu.org>
15739 PR middle-end/57073
15740 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
15741 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
15743 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
15745 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
15747 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
15749 * target.def (register_usage_leveling_p): New hook.
15750 * targhooks.c (default_register_usage_leveling_p): New.
15751 * targhooks.h (default_register_usage_leveling_p): New prototype.
15752 * lra-assigns.c (register_usage_leveling_p): Use the hook.
15753 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
15754 * doc/tm.texi: Update.
15755 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
15757 2013-05-30 Ian Bolton <ian.bolton@arm.com>
15759 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
15760 (*insv_reg<mode>): New define_insn.
15762 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
15764 PR rtl-optimization/57439
15765 * postreload.c (move2add_valid_value_p): Check that we have
15766 a zero subreg_regno_offset when accessing the register in
15767 the requested mode.
15769 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15770 Igor Zamyatin <igor.zamyatin@intel.com>
15772 Silvermont (SLM) architecture pipeline model, tuning and
15774 * config.gcc: Add slm config options and target.
15776 * config/i386/slm.md: New.
15778 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
15780 * config/i386/i386-c.c (ix86_target_macros_internal): New case
15782 (ix86_target_macros_internal): Likewise.
15784 * config/i386/i386.c (slm_cost): New cost.
15785 (m_SLM): New macro flag.
15786 (initial_ix86_tune_features): Set m_SLM.
15787 (x86_accumulate_outgoing_args): Likewise.
15788 (x86_arch_always_fancy_math_387): Likewise.
15789 (processor_target_table): Add slm cost.
15790 (cpu_names): Add slm cpu name.
15791 (x86_option_override_internal): Set SLM ISA.
15792 (ix86_issue_rate): New case PROCESSOR_SLM.
15793 (ia32_multipass_dfa_lookahead): Likewise.
15794 (fold_builtin_cpu): Add slm.
15796 * config/i386/i386.h (TARGET_SLM): New target macro.
15797 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
15798 (processor_type): Add PROCESSOR_SLM.
15800 * config/i386/i386.md (cpu): Add new value "slm".
15801 (slm.md): Include slm.md.
15803 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
15804 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15806 * config/arm/arm-protos.h: Add and update function protos.
15807 * config/arm/arm.c (use_simple_return_p): New added.
15808 (thumb2_expand_return): Check simple_return flag.
15809 * config/arm/arm.md: Add simple_return and conditional simple_return.
15810 * config/arm/iterators.md: Add iterator for return and simple_return.
15812 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15814 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
15815 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
15816 (arm_emit_vfp_multi_reg_pop): Likewise.
15817 (thumb2_emit_ldrd_pop): Likewise.
15818 (arm_expand_epilogue): Add misc REG_CFA notes.
15819 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
15821 2013-05-29 Lawrence Crowl <crowl@google.com>
15823 * config/arm/t-arm: Update for below.
15825 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
15826 Change type to hash_table. Update dependent calls and types.
15828 * config/i386/t-cygming: Update for below.
15830 * config/i386/t-interix: Update for below.
15832 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
15833 Change type to hash_table. Update dependent calls and types.
15834 (i386_find_on_wrapper_list::wrappers): Likewise.
15836 * config/ia64/t-ia64: Update for below.
15838 * config/ia64/ia64.c (bundle_state_table):
15839 Change type to hash_table. Update dependent calls and types.
15841 * config/mips/mips.c (mips_reorg_process_insns::htab):
15842 Change type to hash_table. Update dependent calls and types.
15844 * config/sol2.c (solaris_comdat_htab):
15845 Change type to hash_table. Update dependent calls and types.
15847 * config/t-sol2: Update for above.
15849 2013-05-29 Teresa Johnson <tejohnson@google.com>
15851 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
15852 functions are not yet marked as defined.
15854 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
15855 Pat Haugen <pthaugen@us.ibm.com>
15856 Peter Bergner <bergner@vnet.ibm.com>
15858 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
15862 (vrotl<mode>3): Likewise.
15863 (vashl<mode>3): Likewise.
15864 (vlshr<mode>3): Likewise.
15865 (vashr<mode>3): Likewise.
15867 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15868 support for power8 V2DI builtins.
15870 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
15871 power8 V2DI builtins.
15872 (vupkhsw): Likewise.
15873 (vupklsw): Likewise.
15874 (vaddudm): Likewise.
15875 (vminsd): Likewise.
15876 (vmaxsd): Likewise.
15877 (vminud): Likewise.
15878 (vmaxud): Likewise.
15879 (vpkudum): Likewise.
15880 (vpksdss): Likewise.
15881 (vpkudus): Likewise.
15882 (vpksdus): Likewise.
15887 (vsubudm): Likewise.
15888 (vcmpequd): Likewise.
15889 (vcmpgtsd): Likewise.
15890 (vcmpgtud): Likewise.
15891 (vcmpequd_p): Likewise.
15892 (vcmpgtsd_p): Likewise.
15893 (vcmpgtud_p): Likewise.
15894 (vupkhsw): Likewise.
15895 (vupklsw): Likewise.
15896 (vaddudm): Likewise.
15897 (vmaxsd): Likewise.
15898 (vmaxud): Likewise.
15899 (vminsd): Likewise.
15900 (vminud): Likewise.
15901 (vpksdss): Likewise.
15902 (vpksdus): Likewise.
15903 (vpkudum): Likewise.
15904 (vpkudus): Likewise.
15909 (vsubudm): Likewise.
15911 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
15912 support for power8 V2DI instructions.
15914 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
15915 power8 V2DI instructions. Combine pack and unpack insns to use an
15916 iterator for each mode. Check whether a particular mode supports
15917 Altivec instructions instead of just checking TARGET_ALTIVEC.
15918 (UNSPEC_VPKUWUM): Likewise.
15919 (UNSPEC_VPKSHSS): Likewise.
15920 (UNSPEC_VPKSWSS): Likewise.
15921 (UNSPEC_VPKUHUS): Likewise.
15922 (UNSPEC_VPKSHUS): Likewise.
15923 (UNSPEC_VPKUWUS): Likewise.
15924 (UNSPEC_VPKSWUS): Likewise.
15925 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
15926 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
15927 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
15928 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
15929 (UNSPEC_VUPKHSB): Likewise.
15930 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
15931 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
15932 (UNSPEC_VUPKHSH): Likewise.
15933 (UNSPEC_VUPKLSB): Likewise.
15934 (UNSPEC_VUPKLSH): Likewise.
15936 (VI_char): Likewise.
15937 (VI_scalar): Likewise.
15938 (VI_unit): Likewise.
15940 (VP_small): Likewise.
15941 (VP_small_lc): Likewise.
15942 (VU_char): Likewise.
15943 (add<mode>3): Likewise.
15944 (altivec_vaddcuw): Likewise.
15945 (altivec_vaddu<VI_char>s): Likewise.
15946 (altivec_vadds<VI_char>s): Likewise.
15947 (sub<mode>3): Likewise.
15948 (altivec_vsubcuw): Likewise.
15949 (altivec_vsubu<VI_char>s): Likewise.
15950 (altivec_vsubs<VI_char>s): Likewise.
15951 (altivec_vavgs<VI_char>): Likewise.
15952 (altivec_vcmpbfp): Likewise.
15953 (altivec_eq<mode>): Likewise.
15954 (altivec_gt<mode>): Likewise.
15955 (altivec_gtu<mode>): Likewise.
15956 (umax<mode>3): Likewise.
15957 (smax<mode>3): Likewise.
15958 (umin<mode>3): Likewise.
15959 (smin<mode>3): Likewise.
15960 (altivec_vpkuhum): Likewise.
15961 (altivec_vpkuwum): Likewise.
15962 (altivec_vpkshss): Likewise.
15963 (altivec_vpkswss): Likewise.
15964 (altivec_vpkuhus): Likewise.
15965 (altivec_vpkshus): Likewise.
15966 (altivec_vpkuwus): Likewise.
15967 (altivec_vpkswus): Likewise.
15968 (altivec_vpks<VI_char>ss): Likewise.
15969 (altivec_vpks<VI_char>us): Likewise.
15970 (altivec_vpku<VI_char>us): Likewise.
15971 (altivec_vpku<VI_char>um): Likewise.
15972 (altivec_vrl<VI_char>): Likewise.
15973 (altivec_vsl<VI_char>): Likewise.
15974 (altivec_vsr<VI_char>): Likewise.
15975 (altivec_vsra<VI_char>): Likewise.
15976 (altivec_vsldoi_<mode>): Likewise.
15977 (altivec_vupkhsb): Likewise.
15978 (altivec_vupkhs<VU_char>): Likewise.
15979 (altivec_vupkls<VU_char>): Likewise.
15980 (altivec_vupkhsh): Likewise.
15981 (altivec_vupklsb): Likewise.
15982 (altivec_vupklsh): Likewise.
15983 (altivec_vcmpequ<VI_char>_p): Likewise.
15984 (altivec_vcmpgts<VI_char>_p): Likewise.
15985 (altivec_vcmpgtu<VI_char>_p): Likewise.
15986 (abs<mode>2): Likewise.
15987 (vec_unpacks_hi_v16qi): Likewise.
15988 (vec_unpacks_hi_v8hi): Likewise.
15989 (vec_unpacks_lo_v16qi): Likewise.
15990 (vec_unpacks_hi_<VP_small_lc>): Likewise.
15991 (vec_unpacks_lo_v8hi): Likewise.
15992 (vec_unpacks_lo_<VP_small_lc>): Likewise.
15993 (vec_pack_trunc_v8h): Likewise.
15994 (vec_pack_trunc_v4si): Likewise.
15995 (vec_pack_trunc_<mode>): Likewise.
15997 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
15999 (vec_vmaxsd): Likewise.
16000 (vec_vmaxud): Likewise.
16001 (vec_vminsd): Likewise.
16002 (vec_vminud): Likewise.
16003 (vec_vpksdss): Likewise.
16004 (vec_vpksdus): Likewise.
16005 (vec_vpkudum): Likewise.
16006 (vec_vpkudus): Likewise.
16007 (vec_vrld): Likewise.
16008 (vec_vsld): Likewise.
16009 (vec_vsrad): Likewise.
16010 (vec_vsrd): Likewise.
16011 (vec_vsubudm): Likewise.
16012 (vec_vupkhsw): Likewise.
16013 (vec_vupklsw): Likewise.
16015 2013-05-29 Jan Hubicka <jh@suse.cz>
16017 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
16018 flags; reorder rest of fields in more consistent way.
16019 (varpool_node): Remove analyzed, finalized and alias.
16020 (cgraph_ndoe): Likewise.
16021 (symtab_alias_ultimate_target): New function.
16022 (cgraph_function_node): Move offline.
16023 (cgraph_reset_node): Declare.
16024 (cgraph_comdat_can_be_unshared_p): Remove.
16025 (varpool_remove_initializer): Declare.
16026 (varpool_first_defined_variable, varpool_next_defined_variable
16027 cgraph_first_defined_function, cgraph_next_defined_function): Update.
16028 (cgraph_function_with_gimple_body_p): Update.
16029 (varpool_all_refs_explicit_p): Update.
16030 (symtab_alias_target): New function.
16031 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
16032 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
16033 (cgraph_function_or_thunk_node): Simplify using
16034 symtab_alias_ultimate_target.
16035 (varpool_variable_node): Likewise.
16036 * cgraph.c (cgraph_create_function_alias): Update.
16037 (cgraph_add_thunk): Update.
16038 (cgraph_remove_node): Update.
16039 (dump_cgraph_node): Do not dump removed flags.
16040 (cgraph_function_body_availability): Update.
16041 (cgraph_propagate_frequency): Update.
16042 (verify_cgraph_node): Check sanity of local flag.
16043 (cgraph_function_node): Move here from cgraph.h; revamp for
16044 cgraph_function_or_thunk_node.
16045 * lto-symtab.c (lto_varpool_replace_node): Update.
16046 (lto_symtab_resolve_can_prevail_p): Update.
16047 (lto_symtab_merge_cgraph_nodes): Update.
16048 * ipa-cp.c (determine_versionability, initialize_node_lattices,
16049 propagate_constants_accross_call, devirtualization_time_bonus,
16050 ipcp_propagate_stage): Update.
16051 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
16052 * ipa-inline-transform.c (clone_inlined_nodes,
16053 preserve_function_body_p): Update.
16054 * ipa-reference.c (propagate): Update.
16055 (write_node_summary_p): Update.
16056 * toplev.c (wrapup_global_declaration_2): Update.
16057 * cgraphunit.c (cgraph_analyze_function): Rename to ...
16058 (analyze_function) ... this one.
16059 (cgraph_process_new_functions): Update.
16060 (cgraph_reset_node): Export.
16061 (cgraph_finalize_function): Update.
16062 (cgraph_add_new_function): Update.
16063 (process_function_and_variable_attributes): Update.
16064 (varpool_finalize_decl): Update.
16065 (symbol_finalized): Remove.
16066 (symbol_finalized_and_needed): Rename to ...
16067 (symbol_defined_and_needed): ... update.
16068 (cgraph_analyze_functions): Update.
16069 (handle_alias_pairs): Update.
16070 (mark_functions_to_output): Update.
16071 (assemble_thunk): Update.
16072 (output_in_order): Update.
16073 (output_weakrefs): Update.
16074 (finalize_compilation_unit): Update.
16075 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
16076 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
16077 input_node, input_varpool_node): Update.
16078 * dbxout.c (dbxout_expand_expr): Update.
16079 * cgraphclones.c (cgraph_clone_node): Update.
16080 (cgraph_copy_node_for_versioning): Update.
16081 (cgraph_materialize_clone): Update.
16082 (cgraph_materialize_all_clones): Update.
16083 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
16084 propagate_pure_const, propagate_nothrow): Update.
16085 * lto-streamer-out.c (lto_output, write_symbol): Update.
16086 * ipa-utils.c (ipa_reverse_postorder): Update.
16087 * ipa-inline.c (can_inline_edge_p): Update.
16088 (update_caller_keys, ipa_inline): Update.
16089 * dwarf2out.c (reference_to_unused,
16090 premark_types_used_by_global_vars_helper): Update.
16091 * tree-eh.c (tree_could_trap_p): Update.
16092 * ipa-split.c (consider_split, execute_split_functions): Update.
16093 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
16094 has_addr_references_p): Update; move ahead in file for better
16096 (process_references): Simplify.
16097 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
16098 bodies are removed.
16099 (cgraph_comdat_can_be_unshared_p): Make static.
16100 (cgraph_externally_visible_p): Update.
16101 (varpool_externally_visible_p): Update.
16102 (function_and_variable_visibility): Update.
16103 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
16104 ipa_tm_mark_force_output_node): Update.
16105 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
16106 estimate_edge_devirt_benefit, inline_generate_summary,
16107 inline_write_summary): Update.
16108 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
16109 * ipa-prop.c (ipa_compute_jump_functions): Update.
16110 (ipa_print_node_params, ipa_prop_read_section,
16111 ipa_update_after_lto_read, read_replacements_section): Update.
16112 * varasm.c (mark_decl_referenced): Update.
16113 (assemble_alias, dump_tm_clone_pairs): Update.
16114 * tree-inline.c (copy_bb): Update.
16115 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
16117 * symtab.c (dump_symtab_base): Print new flags.
16118 (verify_symtab_base): Verify new flags.
16119 (symtab_alias_ultimate_target): New function.
16120 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
16121 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
16123 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
16125 * i386.c (ix86_get_function_versions_dispatcher,
16126 ix86_generate_version_dispatcher_body): Update.
16127 (fold_builtin_cpu): Use varpool_add_new_variable.
16128 * varpool.c (varpool_remove_initializer): Break out from ...
16129 (varpool_remove_node): ... this one.
16130 (dump_varpool_node, varpool_node_for_asm,
16131 cgraph_variable_initializer_availability, varpool_analyze_node,
16132 varpool_assemble_decl, varpool_remove_unreferenced_decls,
16133 varpool_finalize_named_section_flags, varpool_create_variable_alias):
16136 2013-05-29 Jan Hubicka <jh@suse.cz>
16138 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
16140 2013-05-29 Easwaran Raman <eraman@google.com>
16142 PR tree-optimization/57442
16143 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
16144 when control exits the main loop.
16146 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
16148 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
16150 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
16151 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
16152 * rx/t-rx: Add rx100 under multi library matches option for nofpu
16155 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16157 PR tree-optimization/57441
16158 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
16159 Don't limit size of incr_vec to number of candidates.
16161 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
16163 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
16164 and mips16 directories.
16165 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
16166 (MULTILIB_DIRNAMES): Ditto.
16167 (MULTILIB_EXCEPTIONS): Add new exceptions.
16168 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
16169 (MULTILIB_DIRNAMES): Ditto.
16170 (MULTILIB_EXCEPTIONS): Add new exceptions.
16172 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16173 Marcus Shawcroft <marcus.shawcroft@arm.com>
16175 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
16176 SYMBOL_TINY_ABSOLUTE.
16177 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
16178 SYMBOL_TINY_ABSOLUTE.
16179 (aarch64_expand_mov_immediate): Likewise.
16180 (aarch64_classify_symbol): Likewise.
16181 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
16182 Permit SYMBOL_TINY_ABSOLUTE.
16183 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
16185 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16186 Marcus Shawcroft <marcus.shawcroft@arm.com>
16188 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
16189 Refactor if/switch. Replace gcc_assert with if.
16191 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16193 * config/i386/i386.c (initial_ix86_tune_features): Enable
16194 FP Reassociation for AMD bdver1 and bdver2.
16196 2013-05-29 Martin Jambor <mjambor@suse.cz>
16198 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
16199 and IMAGPART_EXPR do not occur within other handled_components.
16201 2013-05-29 Richard Biener <rguenther@suse.de>
16203 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
16204 access on whether the use is in the BB we currently try to
16206 (vect_bb_vectorization_profitable_p): Pass the BB we currently
16207 vectorize to vect_bb_slp_scalar_cost.
16209 2013-05-29 Richard Biener <rguenther@suse.de>
16211 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
16212 computing scalar cost offsetted by stmts that are kept live
16214 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
16215 for computation of scalar cost.
16217 2013-05-28 Steve Ellcey <sellcey@mips.com>
16219 * config/mips/mips-cpus.def (mips32r2): Change processor type.
16221 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
16223 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
16224 array notation built-in reduction functions.
16225 * doc/passes.texi (Passes): Added documentation about changes done
16227 * doc/invoke.texi (C Dialect Options): Added documentation about
16228 the -fcilkplus flag.
16229 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
16230 (BUILTINS_DEF): Depend on cilkplus.def.
16231 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
16232 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
16233 * cilkplus.def: New file.
16235 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
16237 PR rtl-optimization/57439
16238 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
16240 2013-05-28 Easwaran Raman <eraman@google.com>
16242 PR tree-optimization/57337
16243 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
16244 (find_insert_point): Correctly identify the insertion point
16245 when two statements with the same UID is compared.
16247 2013-05-28 Richard Biener <rguenther@suse.de>
16249 PR tree-optimization/56787
16250 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
16251 from the list of data references.
16252 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
16254 (vect_analyze_loop_operations): Likewise.
16255 (vect_transform_loop): Remove clobbers.
16257 2013-05-28 Martin Jambor <mjambor@suse.cz>
16259 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
16260 and REALPART_EXPRs have scalar type.
16262 2013-05-28 Richard Biener <rguenther@suse.de>
16264 PR tree-optimization/57411
16265 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
16267 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
16268 virtual operand propagation.
16270 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16272 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
16273 destination register for bmasksi_vis.
16274 (vector_init_bshuffle): Likewise.
16275 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
16277 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16279 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
16280 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
16281 mode if the instruction isn't available in the original mode.
16282 * config/sparc/sparc.opt (mfix-ut699): New option.
16283 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
16284 (divdf3): Turn into expander.
16285 (divdf3_nofix): New insn.
16286 (divdf3_fix): Likewise.
16287 (divsf3): Disable if -mfix-ut699.
16288 (sqrtdf2): Turn into expander.
16289 (sqrtdf2_nofix): New insn.
16290 (sqrtdf2_fix): Likewise.
16291 (sqrtsf2): Disable if -mfix-ut699.
16293 2013-05-27 Richard Biener <rguenther@suse.de>
16295 PR middle-end/57412
16296 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
16297 block for the new loop.
16299 2013-05-27 Richard Biener <rguenther@suse.de>
16301 PR tree-optimization/57343
16302 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
16303 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
16304 (number_of_iterations_cond): Do not build the folded tree.
16306 2013-05-27 Richard Biener <rguenther@suse.de>
16309 PR middle-end/57381
16310 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16311 OEP_CONSTANT_ADDRESS_OF retained.
16313 PR tree-optimization/57417
16314 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
16315 for unchanged base.
16316 (set_ssa_val_to): Compare addresses using
16317 get_addr_base_and_unit_offset.
16319 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
16321 PR rtl-optimization/56833
16322 * postreload.c (move2add_record_mode): New function.
16323 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
16324 (move2add_use_add2_insn): Use move2add_record_sym_value.
16325 (move2add_use_add3_insn): Likewise.
16326 (reload_cse_move2add): Use move2add_valid_value_p and
16327 move2add_record_mode. Invalidate call-clobbered and REG_INC
16328 affected regs by setting reg_mode to VOIDmode.
16329 (move2add_note_store): Don't pretend the inside of a SUBREG is
16330 the actual destination. Invalidate single/leading registers by
16331 setting reg_mode to VOIDmode.
16332 Use move2add_record_sym_value, move2add_valid_value_p and
16333 move2add_record_mode.
16335 2013-05-27 Richard Biener <rguenther@suse.de>
16337 PR tree-optimization/57396
16338 * tree-affine.c (double_int_constant_multiple_p): Properly
16339 return false for val == 0 and div != 0.
16341 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16343 * config/mips/mips.h: Use #elif in preprocessor conditions.
16345 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16348 * config/mips/constraints.md (kl): New constraint.
16349 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
16350 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
16351 constraint for operand 0. Split after CSE for MIPS16. Emit a move
16352 from LO for MIPS16.
16353 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
16355 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16358 * config/mips/mips.c (mips_can_inline_p): New function.
16359 (TARGET_CAN_INLINE_P): Define.
16361 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
16363 * sched-int.h (ds_t, dw_t): Make unsigned int.
16364 Fix documentation that describes how all the ds_t bits are used.
16365 Reserve the last bit for delayed-branch scheduling.
16366 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
16367 (BITS_PER_DEP_WEAK): Fix definition and documentation.
16368 (gen_dep_weak_1): Remove prototype.
16369 * sched-deps.c (get_dep_weak_1): Make static.
16370 * target.def (speculate_insn, needs_block_p, gen_spec_check,
16371 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
16372 * doc/tm.texi: Regenerate.
16373 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
16375 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
16378 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
16380 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
16381 Sandra Loosemore <sandra@codesourcery.com>
16383 * config.gcc (powerpc-*): Allow native for with-cpu.
16385 2013-05-24 Jeff Law <law@redhat.com>
16387 PR tree-optimization/57124
16388 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
16389 conversion feeding a condition if the range has an overflow
16390 if -fstrict-overflow. Add warnings for when we do make the
16393 2013-05-24 Dehao Chen <dehao@google.com>
16395 * tree-cfg.c (locus_discrim_map): Fix the typo.
16396 (locus_discrim_hasher): Likewise.
16397 (locus_discrim_hasher::hash): Likewise.
16398 (locus_discrim_hasher::equal): Likewise.
16400 2013-05-24 Martin Jambor <mjambor@suse.cz>
16402 PR tree-optimization/57294
16403 * cgraph.h (ipa_record_stmt_references): Declare.
16404 * cgraphbuild.c (ipa_record_stmt_references): New function.
16405 (build_cgraph_edges): Use ipa_record_stmt_references.
16406 (rebuild_cgraph_edges): Likewise.
16407 (cgraph_rebuild_references): Likewise.
16408 * ipa-prop.c (ipa_modify_call_arguments): Discard references
16409 associated with the old statement and build references from the
16410 newly built statements.
16411 * ipa-ref.c (ipa_remove_stmt_references): New function.
16412 * ipa-ref.h (ipa_remove_stmt_references): Declare.
16414 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
16416 * lra-constraints.c (emit_spill_move): Use smaller mode for
16418 (check_and_process_move): Consider mem-reg moves for secondary
16420 (curr_insn_transform): Don't lose insns emitted before for
16421 secondary memory moves.
16422 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
16423 reg set up in the current insn.
16425 2013-05-24 Dehao Chen <dehao@google.com>
16427 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
16429 (locus_descrim_hasher::equal): Likewise.
16430 (build_gimple_cfg): New discriminator assignment algorithm.
16431 (make_edges): Likewise.
16432 (next_discriminator_for_locus): Likewise.
16433 (same_line_p): Likewise.
16434 (assign_discriminators): Likewise.
16435 (make_cond_expr_edges): Likewise.
16436 (make_gimple_switch_edges): Likewise.
16437 (make_goto_expr_edges): Likewise.
16438 (make_gimple_asm_edges): Likewise.
16440 2013-05-24 Ian Bolton <ian.bolton@arm.com>
16442 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
16443 X format specifier to only display bottom 16 bits.
16444 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
16445 immediate to match for operand 2, since it will be masked.
16447 2013-05-24 Richard Biener <rguenther@suse.de>
16449 PR tree-optimization/57287
16450 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
16451 all SSA names that occur in abnormal PHIs.
16453 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
16456 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
16457 that index is not negative.
16459 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
16461 PR rtl-optimization/55177
16462 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
16463 (simplify_byte_swapping_operation): New.
16464 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
16465 (simplify_relational_operation_1): Deal with BSWAP.
16467 2013-05-23 Richard Henderson <rth@redhat.com>
16470 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
16471 (ix86_reorg): Call it.
16473 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
16476 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
16477 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
16478 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
16480 2013-05-23 Christian Bruel <christian.bruel@st.com>
16483 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
16485 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16486 Marcus Shawcroft <marcus.shawcroft@arm.com>
16488 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
16489 * config/aarch64/constraints.md (Usa): Remove.
16490 * doc/md.texi (AArch64 Usa): Remove.
16492 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16493 Marcus Shawcroft <marcus.shawcroft@arm.com>
16495 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
16496 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
16497 * config/aarch64/predicates.md (aarch64_const_address): Remove.
16498 (aarch64_mov_operand): Use aarch64_mov_operand_p.
16500 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
16502 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
16503 instruction (AdvSIMD).
16504 * config/aarch64/aarch64-builtins.c
16505 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
16506 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
16508 2013-05-23 Martin Jambor <mjambor@suse.cz>
16510 PR middle-end/57347
16511 * tree.h (contains_bitfld_component_ref_p): Declare.
16512 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
16513 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
16515 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
16516 not access a bit-field. Assert all final offsets are byte-aligned.
16518 2013-05-23 Richard Biener <rguenther@suse.de>
16520 PR tree-optimization/57380
16521 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
16522 least one invariant or re-used load.
16523 * passes.c (init_optimization_passes): Move pass_phiprop before
16526 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
16528 * config/aarch64/aarch64-simd.md
16529 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
16531 2013-05-23 Richard Biener <rguenther@suse.de>
16533 PR middle-end/57381
16534 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16535 OEP_CONSTANT_ADDRESS_OF retained.
16537 2013-05-23 Jakub Jelinek <jakub@redhat.com>
16539 PR middle-end/57344
16540 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
16541 don't lower unit. Handle unit not being always BITS_PER_WORD.
16543 2013-05-23 Richard Biener <rguenther@suse.de>
16545 PR rtl-optimization/57341
16546 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
16547 instead of true_dependence.
16549 2013-05-22 David Malcolm <dmalcolm@redhat.com>
16551 * bb-reorder.c (branch_threshold): Make const.
16552 (exec_threshold): Ditto.
16554 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16555 Pat Haugen <pthaugen@us.ibm.com>
16556 Peter Bergner <bergner@vnet.ibm.com>
16558 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
16559 documentation for the power8 crypto builtins.
16561 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
16563 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
16564 macros for defining power8 builtin functions.
16565 (BU_P8V_AV_2): Likewise.
16566 (BU_P8V_AV_P): Likewise.
16567 (BU_P8V_VSX_1): Likewise.
16568 (BU_P8V_OVERLOAD_1): Likewise.
16569 (BU_P8V_OVERLOAD_2): Likewise.
16570 (BU_CRYPTO_1): Likewise.
16571 (BU_CRYPTO_2): Likewise.
16572 (BU_CRYPTO_3): Likewise.
16573 (BU_CRYPTO_OVERLOAD_1): Likewise.
16574 (BU_CRYPTO_OVERLOAD_2): Likewise.
16575 (XSCVSPDP): Fix typo, point to the correct instruction.
16576 (VCIPHER): Add power8 crypto builtins.
16577 (VCIPHERLAST): Likewise.
16578 (VNCIPHER): Likewise.
16579 (VNCIPHERLAST): Likewise.
16580 (VPMSUMB): Likewise.
16581 (VPMSUMH): Likewise.
16582 (VPMSUMW): Likewise.
16583 (VPERMXOR_V2DI): Likewise.
16584 (VPERMXOR_V4SI: Likewise.
16585 (VPERMXOR_V8HI: Likewise.
16586 (VPERMXOR_V16QI: Likewise.
16587 (VSHASIGMAW): Likewise.
16588 (VSHASIGMAD): Likewise.
16589 (VPMSUM): Likewise.
16590 (VPERMXOR): Likewise.
16591 (VSHASIGMA): Likewise.
16593 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
16594 __CRYPTO__ if the crypto instructions are available.
16595 (altivec_overloaded_builtins): Add support for overloaded power8
16598 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
16599 support for power8 crypto builtins.
16600 (builtin_function_type): Likewise.
16601 (altivec_init_builtins): Add support for builtins that take vector
16602 long long (V2DI) arguments.
16604 * config/rs6000/crypto.md: New file, define power8 crypto
16607 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16608 Pat Haugen <pthaugen@us.ibm.com>
16609 Peter Bergner <bergner@vnet.ibm.com>
16611 * doc/invoke.texi (Option Summary): Add power8 options.
16612 (RS/6000 and PowerPC Options): Likewise.
16614 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
16615 constraints.md instead of rs6000.h. Reorder w* constraints. Add
16616 wm, wn, wr documentation.
16618 * config/rs6000/constraints.md (wm): New constraint for VSX
16619 registers if direct move instructions are enabled.
16620 (wn): New constraint for no registers.
16621 (wq): New constraint for quad word even GPR registers.
16622 (wr): New constraint if 64-bit instructions are enabled.
16623 (wv): New constraint if power8 vector instructions are enabled.
16624 (wQ): New constraint for quad word memory locations.
16626 * config/rs6000/predicates.md (const_0_to_15_operand): New
16627 constraint for 0..15 for crypto instructions.
16628 (gpc_reg_operand): If VSX allow registers in VSX registers as well
16629 as GPR and floating point registers.
16630 (int_reg_operand): New predicate to match only GPR registers.
16631 (base_reg_operand): New predicate to match base registers.
16632 (quad_int_reg_operand): New predicate to match even GPR registers
16633 for quad memory operations.
16634 (vsx_reg_or_cint_operand): New predicate to allow vector logical
16635 operations in both GPR and VSX registers.
16636 (quad_memory_operand): New predicate for quad memory operations.
16637 (reg_or_indexed_operand): New predicate for direct move support.
16639 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
16640 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
16641 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
16642 (POWERPC_MASKS): Add power8 options.
16643 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
16646 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
16647 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
16649 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
16650 (-mpower8-fusion): New power8 options.
16651 (-mpower8-fusion-sign): Likewise.
16652 (-mpower8-vector): Likewise.
16653 (-mcrypto): Likewise.
16654 (-mdirect-move): Likewise.
16655 (-mquad-memory): Likewise.
16657 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
16658 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
16660 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
16661 (rs6000_debug_vector_unit): Add p8_vector.
16662 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
16663 definitions. Also print fusion state.
16664 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
16665 (rs6000_builtin_mask_calculate): Add power8 builtin support.
16666 (rs6000_option_override_internal): Add support for power8.
16667 (rs6000_common_init_builtins): Add debugging for skipped builtins
16668 if -mdebug=builtin.
16669 (rs6000_adjust_cost): Add power8 support.
16670 (rs6000_issue_rate): Likewise.
16671 (insn_must_be_first_in_group): Likewise.
16672 (insn_must_be_last_in_group): Likewise.
16673 (force_new_group): Likewise.
16674 (rs6000_register_move_cost): Likewise.
16675 (rs6000_opt_masks): Likewise.
16677 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
16678 power8 capable assembler, default to power7 options.
16679 (TARGET_DIRECT_MOVE): Likewise.
16680 (TARGET_CRYPTO): Likewise.
16681 (TARGET_P8_VECTOR): Likewise.
16682 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
16683 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
16684 (VECTOR_MEM_P8_VECTOR_P): Likewise.
16685 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
16686 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
16687 (TARGET_XSCVDPSPN): Likewise.
16688 (TARGET_XSCVSPDPN): Likewsie.
16689 (TARGET_SYNC_HI_QI): Likewise.
16690 (TARGET_SYNC_TI): Likewise.
16691 (MASK_CRYPTO): Likewise.
16692 (MASK_DIRECT_MOVE): Likewise.
16693 (MASK_P8_FUSION): Likewise.
16694 (MASK_P8_VECTOR): Likewise.
16695 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
16696 temporary used by some of the direct move instructions to get two FP
16697 temporary registers does not force creation of a stack frame.
16698 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
16699 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
16700 that any VSX registers are tieable, even if they are also an
16701 Altivec vector mode.
16702 (r6000_reg_class_enum): Add wm, wr, wv constraints.
16703 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
16704 (RS6000_BTM_CRYPTO): Likewise.
16705 (RS6000_BTM_COMMON): Likewise.
16707 * config/rs6000/rs6000.md (cpu attribute): Add power8.
16708 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
16709 (enum rs6000_vector): Add power8 vector support.
16711 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16715 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
16716 (any_sibcall_could_use_r3): this and handle indirect calls.
16717 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
16719 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16721 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
16723 2013-05-22 Richard Biener <rguenther@suse.de>
16725 PR middle-end/57349
16726 * profile.c (branch_prob): Do not split blocks that are
16727 abnormally receiving from ECF_RETURNS_TWICE functions.
16729 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
16731 * recog.c (offsettable_address_addr_space_p): Fix calculation of
16732 address mode. Move pointer mode initialization to the same place.
16734 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
16736 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
16737 while it has any effect.
16739 2013-05-21 Easwaran Raman <eraman@google.com>
16741 PR tree-optimization/57322
16742 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
16743 UID of the statement added to the BB to be 1.
16745 2013-05-21 Jakub Jelinek <jakub@redhat.com>
16747 PR tree-optimization/57331
16748 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
16749 of conversion from pointer type to integral type with integer.
16751 2013-05-21 Martin Jambor <mjambor@suse.cz>
16754 * ipa-prop.c (ipa_read_node_info): Process param_used and
16755 controlled_uses in the same order as when writing.
16757 2013-05-21 Magnus Granberg <baldrick@free.fr>
16760 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
16762 2013-05-21 Richard Biener <rguenther@suse.de>
16764 PR tree-optimization/57318
16765 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
16766 estimate stmts with side-effects as likely eliminated.
16768 2013-05-21 Richard Biener <rguenther@suse.de>
16770 PR tree-optimization/57330
16771 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
16772 preserve the call stmts fntype.
16774 2013-05-21 Richard Biener <rguenther@suse.de>
16776 PR tree-optimization/57303
16777 * tree-ssa-sink.c (statement_sink_location): Improve killing
16778 stmt detection and properly handle self-assignments.
16780 2013-05-21 Christian Bruel <christian.bruel@st.com>
16782 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
16783 spanning registers. LEAF_REG_REMAP is supported only for contiguous
16784 registers. Set register size out of the PARALLEL loop.
16786 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
16789 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
16790 (*fmasf4, *fmasf4_media): New insns.
16792 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16794 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
16795 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
16796 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
16797 (mips_idiv_insns): Update the comments to say that the returned
16798 instruction counts are in units of BASE_INSN_LENGTH.
16799 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
16800 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
16801 using 2 rather than 4 as the length of indirect MIPS16 and
16802 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
16803 length of a NOP. Don't divide MIPS16 lengths by 2.
16804 (mips16_split_long_branches): Assume a branch is long if the
16805 length is greater than 4 rather than 8.
16806 * config/mips/mips.md (length): Give MIPS16 lengths directly,
16807 rather than multiplying them by 2. Multiply instruction counts
16808 by BASE_INSN_LENGTH rather than 4.
16809 (*jump_mips16, tls_get_tp_mips16_<mode>)
16810 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
16812 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16814 * config/mips/mips.md (extended_mips16): Remove branch case.
16815 (length): Remove duplicated extended_mips16 test.
16817 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16819 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
16821 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
16823 * recog.h (Recog_data): Rename to...
16824 (recog_data_d): ...this.
16825 (recog_data): Update accordingly.
16826 * recog.c (recog_data): Likewise.
16827 * reload.c (save_recog_data): Likewise.
16828 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
16829 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
16831 2013-05-17 Julian Brown <julian@codesourcery.com>
16833 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
16834 found in a REG_EQUAL note, invalidate it.
16836 2013-05-17 Easwaran Raman <eraman@google.com>
16838 * tree-ssa-reassoc.c (find_insert_point): New function.
16839 (insert_stmt_after): Likewise.
16840 (get_def_stmt): Likewise.
16841 (ensure_ops_are_available): Likewise.
16842 (not_dominated_by): Likewise.
16843 (rewrite_expr_tree): Do not move statements beyond what is
16844 necessary. Remove call to swap_ops_for_binary_stmt...
16845 (reassociate_bb): ... and move it here.
16846 (build_and_add_sum): Assign UIDs for new statements.
16847 (linearize_expr): Likewise.
16848 (do_reassoc): Renumber gimple statement UIDs.
16850 2013-05-17 Jan Hubicka <jh@suse.cz>
16852 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
16854 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
16855 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
16857 (output_weakrefs): Update.
16859 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
16860 Martin Jambor <mjambor@suse.cz>
16862 PR middle-end/57276
16863 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
16864 value that corresponds to the given aggval is found in values vector.
16866 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
16868 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
16869 sse, sse2, sse3, ssse3 and sse4a flags to options.
16871 2013-05-17 David Malcolm <dmalcolm@redhat.com>
16873 * gengtype-state.c: (s_expr_writer): New class, to handle
16874 prettifying of output layout of s-expressions.
16875 (state_writer): New class, to write out gtype.state.
16876 (state_written_type_count): Move this variable into member data of
16878 (s_expr_writer::s_expr_writer): New code: constructor for new class
16879 (state_writer::state_writer(): ditto
16880 (s_expr_writer::write_new_line): New function
16881 (s_expr_writer::write_any_indent): ditto
16882 (s_expr_writer::begin_s_expr): ditto
16883 (s_expr_writer::end_s_expr): ditto
16884 (write_state_fileloc): convert to method of state_writer...
16885 (state_writer:: write_state_fileloc): ...and use methods of
16886 s_expr_writer to write indentation into the gtype.state output file
16887 to visually represent the hierarchical structure of the list
16889 (write_state_fields): ditto, renaming to...
16890 (state_writer::write_state_fields)
16891 (write_state_a_string): ditto, renaming to...
16892 (state_writer::write_state_a_string)
16893 (write_state_string_option): ditto, renaming to...
16894 (state_writer::write_state_string_option)
16895 (write_state_type_option): ditto, renaming to...
16896 (state_writer::write_state_type_option)
16897 (write_state_nested_option): ditto, renaming to...
16898 (state_writer::write_state_nested_option)
16899 (write_state_option): ditto, renaming to...
16900 (state_writer::write_state_option)
16901 (write_state_options): ditto, renaming to...
16902 (state_writer::write_state_options)
16903 (write_state_lang_bitmap): ditto, renaming to...
16904 (state_writer::write_state_lang_bitmap)
16905 (write_state_version): ditto, renaming to...
16906 (state_writer::write_state_version)
16907 (write_state_scalar_type): ditto, renaming to...
16908 (state_writer::write_state_scalar_type)
16909 (write_state_string_type): ditto, renaming to...
16910 (state_writer::write_state_string_type)
16911 (write_state_undefined_type): ditto, renaming to...
16912 (state_writer::write_state_undefined_type)
16913 (write_state_struct_union_type): ditto, renaming to...
16914 (state_writer::write_state_struct_union_type)
16915 (write_state_struct_type): ditto, renaming to...
16916 (state_writer::write_state_struct_type)
16917 (write_state_user_struct_type): ditto, renaming to...
16918 (state_writer::write_state_user_struct_type)
16919 (write_state_lang_struct_type): ditto, renaming to...
16920 (state_writer::write_state_lang_struct_type)
16921 (write_state_param_struct_type): ditto, renaming to...
16922 (state_writer::write_state_param_struct_type)
16923 (write_state_pointer_type): ditto, renaming to...
16924 (state_writer::write_state_pointer_type)
16925 (write_state_array_type): ditto, renaming to...
16926 (state_writer::write_state_array_type)
16927 (write_state_gc_used): ditto, renaming to...
16928 (state_writer::write_state_gc_used)
16929 (write_state_common_type_content): ditto, renaming to...
16930 (state_writer::write_state_common_type_content)
16931 (write_state_type): ditto, renaming to...
16932 (state_writer::write_state_type)
16933 (write_state_pair_list): ditto, renaming to...
16934 (state_writer::write_state_pair_list)
16935 (write_state_pair): ditto, renaming to...
16936 (state_writer::write_state_pair)
16937 (write_state_typedefs): ditto, renaming to...
16938 (state_writer::write_state_typedefs)
16939 (write_state_structures): ditto, renaming to...
16940 (state_writer::write_state_structures)
16941 (write_state_param_structs): ditto, renaming to...
16942 (state_writer::write_state_param_structs)
16943 (write_state_variables): ditto, renaming to...
16944 (state_writer::write_state_variables)
16945 (write_state_srcdir): ditto, renaming to...
16946 (state_writer::write_state_srcdir)
16947 (write_state_files_list): ditto, renaming to...
16948 (state_writer::write_state_files_list)
16949 (write_state_languages): ditto, renaming to...
16950 (state_writer::write_state_languages)
16951 (write_state): create a state_writer instance and use it when
16952 writing out the state file
16954 2013-05-17 Mike Stump <mikestump@comcast.net>
16956 PR rtl-optimization/57304
16957 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
16958 accessing DF_REF_REAL_LOC.
16960 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16962 PR rtl-optimization/57281
16963 PR rtl-optimization/57300
16964 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
16965 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
16966 what the other splitter did if the registers are dead.
16968 2013-05-17 Richard Biener <rguenther@suse.de>
16970 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
16973 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16975 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
16978 2013-05-17 Marek Polacek <polacek@redhat.com>
16980 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
16981 length when doing non-zero store of storing '\0' to '\0'.
16983 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16985 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
16986 vect_external_def oprnd1 with loop_vinfo, try to emit
16987 optional cast, negation and and stmts on the loop preheader
16988 edge instead of into the pattern def seq.
16990 PR tree-optimization/57051
16991 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
16992 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
16994 2013-05-16 Nick Clifton <nickc@redhat.com>
16996 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
16997 (rl78_is_naked_func): New function.
16998 (rl78_expand_prologue): Skip prologue generation for naked functions.
16999 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
17000 * doc/extend.texi (naked): Add RL78 to the list of processors
17001 that supports this attribute.
17003 2013-05-16 Jeff Law <law@redhat.com>
17005 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
17007 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
17009 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
17010 cache parameters using detect_caches_amd also for CYRIX,
17011 NSC and TM2 signatures.
17013 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
17014 Dzianis Kahanovich <mahatma@eu.by>
17018 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
17019 VIA/Centaur processors and determine their cache parameters
17020 using detect_caches_amd.
17022 2013-05-16 Teresa Johnson <tejohnson@google.com>
17024 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
17025 (rtl_verify_edges): New function.
17026 (rtl_verify_bb_insns): Ditto.
17027 (rtl_verify_bb_pointers): Ditto.
17028 (rtl_verify_bb_insn_chain): Ditto.
17029 (rtl_verify_fallthru): Ditto.
17030 (rtl_verify_bb_layout): Ditto.
17031 (rtl_verify_flow_info_1): Outline checks into new functions.
17032 (rtl_verify_flow_info): Ditto.
17034 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
17036 * cfghooks.c (copy_bbs): Add update_dominance argument.
17037 * cfghooks.h (copy_bbs): Update prototype.
17038 * tree-cfg.c (gimple_duplicate_sese_region):
17039 Add update_dominance argument.
17040 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
17041 * tree-ssa-loop-ch.c (copy_loop_headers): Update
17042 gimple_duplicate_sese_region call.
17043 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
17044 Update copy_bbs call.
17045 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
17046 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
17048 2013-05-16 Jakub Jelinek <jakub@redhat.com>
17050 * tree-vectorizer.h (NUM_PATTERNS): Increment.
17051 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
17052 vect_recog_rotate_pattern.
17053 (vect_recog_rotate_pattern): New function.
17055 2013-05-16 Jason Merrill <jason@redhat.com>
17057 * Makefile.in (LLINKER): New variable.
17058 (mostlyclean): Remove link mutex.
17059 * configure.ac: Handle --enable-link-mutex.
17060 * lock-and-run.sh: New script.
17062 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17065 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
17068 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17070 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
17072 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
17074 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
17075 * config/arm/arm.c (next_consecutive_mem): New function.
17076 (gen_movmem_ldrd_strd): Likewise.
17077 * config/arm/arm.md (movmemqi): Update condition and code.
17078 (unaligned_loaddi, unaligned_storedi): New patterns.
17080 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17082 * config.gcc: Obsolete *-*-solaris2.9*.
17083 * doc/install.texi (Specific, *-*-solaris2*): Document it.
17085 2013-05-16 Richard Biener <rguenther@suse.de>
17087 * passes.c (init_optimization_passes): Move pass_parallelize_loops
17088 earlier, after GRAPHITE transforms and IV canonicalization.
17090 2013-05-16 Jakub Jelinek <jakub@redhat.com>
17092 * omp-low.c (extract_omp_for_data): For collapsed loops,
17093 if at least one of the loops is known at compile time to
17094 iterate zero times, set count to 0.
17095 (expand_omp_regimplify_p): New function.
17096 (expand_omp_for_generic): For collapsed loops, if at least
17097 one of the loops isn't known to iterate at least once,
17098 add runtime check with setting count to 0.
17099 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
17100 For unsigned types if it isn't known at compile time that
17101 the loop will iterate at least once, add runtime check to bypass
17102 the whole loop if initial condition isn't true.
17104 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
17106 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
17108 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
17110 PR middle-end/57286
17111 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
17112 transformations to avoid an infinite loop.
17114 2013-05-16 Marek Polacek <polacek@redhat.com>
17116 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
17118 2013-05-15 Leif Ekblad <leif@rdos.net>
17120 * config/i386/i386.c (ix86_decompose_address): Use
17121 DEFAULT_TLS_SEG_REG to access TLS segment register.
17122 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
17123 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
17124 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
17126 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
17129 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
17130 sibling calls to functions that would normally be lazily bound,
17131 unless $gp is call-clobbered.
17133 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
17135 * config/i386/i386.c (ix86_option_override_internal): Update
17136 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
17137 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
17138 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
17139 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
17141 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
17143 2013-05-15 Andreas Schwab <schwab@suse.de>
17145 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
17146 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
17149 2013-05-15 Teresa Johnson <tejohnson@google.com>
17151 * loop-unroll.c (report_unroll_peel): Check decision before
17152 emitting unroll/peel message.
17154 2013-05-15 Teresa Johnson <tejohnson@google.com>
17156 * function.h (has_bb_partition): New rtl_data flag.
17157 (bb_reorder_complete): Ditto.
17158 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
17159 instead of flag_reorder_blocks_and_partition.
17160 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
17161 with some enhancements.
17162 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
17163 * bb-reorder.c (connect_traces): Check for has_bb_partition
17164 instead of flag_reorder_blocks_and_partition.
17165 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
17166 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
17167 verify_hot_cold_block_grouping.
17168 (partition_hot_cold_basic_blocks): Set has_bb_partition.
17170 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17173 * config/arm/predicates.md (call_insn_operand): New predicate.
17174 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
17175 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
17176 if insn is not a tail call.
17177 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
17179 * config/arm/arm.h (enum reg_class): New caller save register class.
17180 (REG_CLASS_NAMES): Likewise.
17181 (REG_CLASS_CONTENTS): Likewise.
17182 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
17185 2013-05-15 Richard Biener <rguenther@suse.de>
17187 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
17188 of MSG_OPTIMIZED_LOCATIONS.
17189 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
17190 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
17192 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
17193 of MSG_OPTIMIZED_LOCATIONS.
17194 (execute_vect_slp): Likewise.
17195 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
17196 (vect_create_cond_for_alias_checks): Likewise.
17197 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
17198 (vect_recog_widen_mult_pattern): Likewise.
17199 (vect_recog_widen_sum_pattern): Likewise.
17200 (vect_recog_over_widening_pattern): Likewise.
17201 (vect_recog_widen_shift_pattern): Likewise.
17202 (vect_recog_vector_vector_shift_pattern): Likewise.
17203 (vect_recog_divmod_pattern): Likewise.
17204 (vect_recog_mixed_size_cond_pattern): Likewise.
17205 (vect_recog_bool_pattern): Likewise.
17206 (vect_pattern_recog_1): Likewise.
17208 2013-05-15 Martin Jambor <mjambor@suse.cz>
17210 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
17211 non-functions to builtin_unreachable.
17212 * ipa-inline-transform.c (inline_call): Do not assert estimates were
17213 correct when new direct edges were discovered.
17215 2013-05-15 Martin Jambor <mjambor@suse.cz>
17217 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
17218 header, print symbol order instead of node uid, print more information
17219 about indirect edge targets.
17220 (ipa_make_edge_direct_to_target): Print symbol order instead of node
17222 (ipa_make_edge_direct_to_target): Likewise.
17223 (remove_described_reference): Likewise.
17224 (propagate_controlled_uses): Likewise.
17225 (ipa_print_node_params): Also print symbol order.
17226 (ipcp_transform_function): Print symbol order instead of node uids.
17227 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
17228 (cgraph_get_create_real_symbol_node): Likewise.
17229 * ipa-cp.c (print_lattice): Likewise.
17230 (print_all_lattices): Likewise.
17231 (determine_versionability): Likewise.
17232 (initialize_node_lattices): Likewise.
17233 (estimate_local_effects): Likewise.
17234 (update_profiling_info): Likewise.
17235 (create_specialized_node): Likewise.
17236 (perhaps_add_new_callers): Likewise.
17237 (decide_about_value): Likewise.
17238 (decide_whether_version_node): Likewise.
17239 (identify_dead_nodes): Likewise.
17240 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
17241 (dump_inline_summary): Likewise.
17242 (estimate_node_size_and_time): Likewise.
17243 (inline_analyze_function): Likewise.
17244 * ipa-inline.c (report_inline_failed_reason): Likewise.
17245 (want_early_inline_function_p): Likewise.
17246 (edge_badness): Likewise.
17247 (update_edge_key): Likewise.
17248 (inline_small_functions): Likewise. Add dumping of order to two other
17250 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
17251 instead of node uids.
17252 (propagate_pure_const): Likewise.
17253 (propagate_pure_const): Likewise.
17254 * ipa-utils.c (dump_cgraph_node_set): Likewise.
17255 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
17256 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
17258 * tree-pretty-print.c (dump_function_header): Likewise.
17259 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
17260 Print symbol order instead of node uids.
17262 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17264 * config/s390/s390.c (s390_register_move_cost): Don't impose the
17265 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
17267 2013-05-15 Richard Biener <rguenther@suse.de>
17269 PR tree-optimization/57275
17270 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
17271 return value for fail to do runtime alias checks for gather loads.
17273 2013-05-15 Jan Hubicka <jh@suse.cz>
17277 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
17278 weakrefs are not external.
17279 (lto_symtab_merge_decls): Fix thinko when dealing with
17280 non-lto_symtab decls.
17281 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
17282 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
17283 * varpool.c (dump_varpool_node): Dump more flags.
17285 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17287 * config/i386/i386.c (processor_alias_table): Add instruction
17288 FSGSBASE for AMD bdver3 architecture.
17290 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17292 * tree.c (warn_deprecated_use): Print file:line using locus color.
17293 * diagnostic.c (diagnostic_report_current_module): Print file:line
17294 and file:line:column using locus color.
17296 2013-05-14 Mike Stump <mikestump@comcast.net>
17298 * gdbinit.in: Add __null.
17300 2013-05-14 Mike Stump <mikestump@comcast.net>
17302 * recog.h: Rename struct recog_data to Recog_data.
17303 * recog.c: Likewise.
17304 * reload.c (can_reload_into): Likewise.
17305 * config/picochip/picochip.c: Likewise.
17307 2013-05-14 Mike Stump <mikestump@comcast.net>
17309 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
17311 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
17313 * resource.h (struct resources): Remove unch_memory member.
17314 (CLEAR_RESOURCE): Don't clear unch_memory.
17315 * resource.c (mark_referenced_resources): Don't set it.
17316 (mark_set_resources): Likewise.
17317 (mark_target_live_regs): Don't clear it.
17318 (init_resource_info): Likewise.
17319 * reorg.c (resource_conflicts_p): Don't compare it.
17320 (redundant_insn): Don't set it.
17322 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
17324 * emit-rtl.c (next_label): Remove unused function.
17325 (skip_consecutive_labels, link_cc0_insns): Move to ...
17326 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
17327 only place where these functions are used, and make them static.
17329 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17331 * fold-const.c (fold_negate_expr): Handle vectors.
17332 (fold_truth_not_expr): Make it static.
17333 (fold_invert_truthvalue): New static function.
17334 (invert_truthvalue_loc): Handle vectors. Do not call
17335 fold_truth_not_expr directly.
17336 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
17337 <TRUTH_NOT_EXPR>: Do not cast to boolean.
17338 (fold_comparison): Handle vector constants.
17339 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
17340 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
17341 * tree.h (fold_truth_not_expr): Remove declaration.
17343 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
17345 * config/aarch64/aarch64-simd.md
17346 (aarch64_vcond_internal<mode>): Rename to...
17347 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
17348 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
17349 float modes. Clarify all iterator modes.
17350 (vcond<mode><mode>): Use new name for vcond expanders.
17351 (vcond<v_cmp_result><mode>): Likewise.
17352 (vcondu<mode><mode>: Likewise.
17353 * config/aarch64/iterators.md (VDQF_COND): New.
17355 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17358 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
17359 variable for the shift amount. Check that we shift by non-negative
17362 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
17365 * config/arm/arm.h (EPILOGUE_USES): Only return true
17366 for LR_REGNUM after epilogue_completed.
17368 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
17370 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
17371 is error_mark_node.
17373 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17376 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
17377 and Solaris 11+/x86 with gld.
17378 * configure: Regenerate.
17380 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17382 * expmed.c (expand_shift_1): Canonicalize rotates by
17383 constant bitsize / 2 to bitsize - 1.
17384 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
17385 case ROTATERT>: Likewise.
17388 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17390 * config/i386/i386.md (rotateinv): New code attr.
17391 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17392 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17393 roll $31, %eax, etc.
17395 2013-05-14 Richard Biener <rguenther@suse.de>
17397 PR middle-end/57235
17398 * tree-eh.c (sink_clobbers): Give up for successors with
17399 multiple predecessors and no virtual uses.
17401 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
17403 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
17404 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
17406 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17408 PR middle-end/57251
17409 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
17410 the case when both op0 and op1 have VOIDmode.
17412 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17414 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
17415 in multiply-accumulate mode.
17417 2013-05-13 Guozhi Wei <carrot@google.com>
17419 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
17421 2013-05-13 Kai Tietz <ktietz@redhat.com>
17424 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
17425 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
17426 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
17427 * config/i386/i386.c (ix86_option_override_internal): Likewise.
17428 (ix86_expand_prologue): Likewise.
17429 (ix86_expand_split_stack_prologue): Likewise.
17430 (legitimate_pic_address_disp_p): Likewise.
17431 (legitimize_pic_address): Likewise.
17432 (legitimize_tls_address): Likewise.
17433 (legitimize_pe_coff_symbol): Likewise.
17434 (output_pic_addr_const): Likewise.
17435 (construct_plt_address): Likewise.
17436 (ix86_expand_call): Likewise.
17437 (x86_output_mi_thunk): Likewise.
17438 (x86_function_profiler): Likewise.
17440 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
17442 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
17443 similar switch cases.
17444 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
17445 (aarch64_simd_mov_to_<mode>low): Delete.
17446 (aarch64_simd_mov_to_<mode>high): Delete.
17447 (move_lo_quad_<mode>): Add w<-r alternative.
17448 (aarch64_simd_move_hi_quad_<mode>): Likewise.
17449 (aarch64_simd_mov_from_*): Update type attribute.
17450 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
17453 2013-05-13 Jan Hubicka <jh@suse.cz>
17455 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
17456 * config/i386/i386.c (ix86_compute_frame_layout,
17457 ix86_expand_epilogue, emit_i387_cw_initialization,
17458 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
17459 ix86_local_alignment): Fix use of size/speed predicates.
17461 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17463 PR tree-optimization/45216
17464 PR tree-optimization/57157
17465 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
17466 the (-Y) & (B - 1) variant if OP is |.
17467 * expmed.c (expand_shift_1): For rotations by const0_rtx just
17468 return shifted. Use (-op1) & (prec - 1) as other_amount
17469 instead of prec - op1.
17471 2013-05-13 Martin Jambor <mjambor@suse.cz>
17473 PR middle-end/42371
17474 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
17475 (ipa_constant_data): New type.
17476 (ipa_jump_func): Use ipa_constant_data to hold information about
17477 constant jump functions.
17478 (ipa_get_jf_constant): Adjust to jump function type changes.
17479 (ipa_get_jf_constant_rdesc): New function.
17480 (ipa_param_descriptor): New field controlled_uses.
17481 (ipa_get_controlled_uses): New function.
17482 (ipa_set_controlled_uses): Likewise.
17483 * ipa-ref.h (ipa_find_reference): Declare.
17484 * ipa-prop.c (ipa_cst_ref_desc): New type.
17485 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
17487 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
17488 New parameter cs. Adjust all callers.
17489 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
17490 (remove_described_reference): New function.
17491 (jfunc_rdesc_usable): Likewise.
17492 (try_make_edge_direct_simple_call): Decrement controlled use count,
17493 attempt to remove reference if it hits zero.
17494 (combine_controlled_uses_counters): New function.
17495 (propagate_controlled_uses): Likewise.
17496 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
17497 (ipa_edge_duplication_hook): Duplicate reference descriptions.
17498 (ipa_print_node_params): Print described use counter.
17499 (ipa_write_jump_function): Adjust to jump function type changes.
17500 (ipa_read_jump_function): New parameter CS, pass it to
17501 ipa_set_jf_constant. Adjust caller.
17502 (ipa_write_node_info): Stream controlled use count
17503 (ipa_read_node_info): Likewise.
17504 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
17506 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
17507 count. Remove cloning-added reference if it reaches zero.
17508 * ipa-ref.c (ipa_find_reference): New function.
17510 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17512 * config/i386/i386.c (processor_target_table): Modified default
17513 alignment values for AMD BD and BT architectures.
17515 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
17517 * tree-vect-generic.c (uniform_vector_p): Move ...
17518 * tree.c (uniform_vector_p): ... here.
17519 * tree.h (uniform_vector_p): Declare it.
17520 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
17523 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17525 PR tree-optimization/57230
17526 * tree-ssa-strlen.c (handle_char_store): Record length for
17527 array store from STRING_CST.
17529 PR tree-optimization/57230
17530 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
17533 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
17535 * config/epiphany/epiphany.c (epiphany_init): Check size of
17536 NUM_MODES_FOR_MODE_SWITCHING.
17537 (epiphany_expand_prologue):
17538 Remove CONFIG_REGNUM initial value handling code.
17539 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
17540 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
17541 (emit_set_fp_mode, epiphany_mode_after): Likewise.
17542 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
17543 Don't return 1 for FP_MODE_NONE.
17544 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
17545 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
17546 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
17547 * config/epiphany/epiphany.md (save_config): New pattern.
17549 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
17551 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
17553 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
17555 * config/i386/i386.md (memory): Handle sseishft1.
17556 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
17557 (*vec_extractv2di_1): Ditto.
17559 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
17561 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
17564 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
17566 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
17567 Add mthumb/march=armv7-a multilib.
17568 Add mthumb/march=armv7-r multilib.
17569 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
17571 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
17573 * config/v850/t-rtems: Add more multilibs.
17575 2013-05-10 Richard Biener <rguenther@suse.de>
17577 PR tree-optimization/57214
17578 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
17579 not propagate from SSA names that occur in abnormal PHI nodes.
17581 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
17583 * stor-layout.c (element_precision): New function.
17584 * machmode.h (element_precision): Declare it.
17585 * tree.c (build_minus_one_cst): New function.
17586 (element_precision): Likewise.
17587 * tree.h (build_minus_one_cst): Declare new function.
17588 (element_precision): Likewise.
17589 * fold-const.c (operand_equal_p): Use element_precision.
17590 (fold_binary_loc): Handle vector types.
17591 * convert.c (convert_to_integer): Use element_precision.
17592 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
17595 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
17597 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
17598 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
17599 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
17600 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17601 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
17602 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
17603 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
17604 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17605 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
17606 (Uuw8): New constraints.
17607 (Usb4): Move into alphabetical order.
17608 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
17609 (sd8_operand, ub8_operand, uw8_operand): New predicates.
17610 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
17611 previously unnamed patterns.
17612 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
17613 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
17614 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
17615 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
17616 of set_attr_alternative/if_then_else. Use extended_mips16 instead
17617 of specific lengths.
17619 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17621 * config/i386/i386.md (rotateinv): New code attr.
17622 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17623 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17624 roll $31, %eax, etc.
17626 PR tree-optimization/45216
17627 PR tree-optimization/57157
17628 * tree-ssa-forwprop.c (simplify_rotate): New function.
17629 (ssa_forward_propagate_and_combine): Call it.
17631 2013-05-10 Richard Biener <rguenther@suse.de>
17633 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
17634 disable peeling when we version for aliasing.
17635 (vector_alignment_reachable_p): Honor explicit user alignment.
17636 (vect_supportable_dr_alignment): Likewise.
17637 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
17638 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
17639 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
17640 then peeling to arrange for the cost-model check to come first.
17642 2013-05-10 Alan Modra <amodra@gmail.com>
17644 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
17645 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
17646 * configure: Regenerate.
17648 2013-05-10 Alan Modra <amodra@gmail.com>
17651 * varasm.c (default_elf_select_section): Move !DECL_P check..
17652 (get_named_section): ..to here before calling get_section_name.
17654 (default_section_type_flags): Add DECL_P check.
17655 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
17656 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
17658 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
17660 * config/epiphany/epiphany.c (epiphany_expand_prologue):
17661 When using gen_stack_adjust_str with a register offset, add a
17662 REG_FRAME_RELATED_EXPR note.
17664 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
17666 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
17667 (*vec_extractv4si_zext_mem): Ditto.
17668 (*vec_extractv2di): Add 0->x and x->x alternatives.
17669 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
17670 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
17672 2013-05-09 Jason Merrill <jason@redhat.com>
17674 N3639 C++1y VLA support
17675 * gimplify.c (gimplify_vla_decl): Don't touch an existing
17678 * tree.c (build_constructor_va): New.
17679 * tree.h: Declare it.
17681 2013-05-09 Martin Jambor <mjambor@suse.cz>
17684 * gimple-fold.c (canonicalize_constructor_val): Call
17685 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
17687 2013-05-09 Jan Hubicka <jh@suse.cz>
17688 Richard Biener <rguenther@suse.de>
17691 * symtab.c (symtab_make_decl_local): Do not add private names.
17693 2013-05-09 Jan Hubicka <jh@suse.cz>
17696 * symtab.c (insert_to_assembler_name_hash): Handle clones.
17697 (unlink_from_assembler_name_hash): Likewise.
17698 (symtab_prevail_in_asm_name_hash, symtab_register_node,
17699 symtab_unregister_node, symtab_initialize_asm_name_hash,
17700 change_decl_assembler_name): Update.
17702 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
17704 * config/aarch64/aarch64.md: New movtf split.
17705 (*movtf_aarch64): Update.
17706 (aarch64_movdi_tilow): Handle TF modes and rename to
17707 aarch64_movdi_<mode>low.
17708 (aarch64_movdi_tihigh): Handle TF modes and rename to
17709 aarch64_movdi_<mode>high
17710 (aarch64_movtihigh_di): Handle TF modes and rename to
17711 aarch64_mov<mode>high_di
17712 (aarch64_movtilow_di): Handle TF modes and rename to
17713 aarch64_mov<mode>low_di
17714 (aarch64_movtilow_tilow): Remove spurious whitespace.
17715 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
17717 (aarch64_print_operand): Update.
17719 2013-05-09 Alan Modra <amodra@gmail.com>
17721 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
17723 * configure: Regenerate.
17725 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17727 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
17728 splitter preparation statements.
17729 * config/i386/sse.md (*vec_extract* splitters): Ditto.
17730 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
17733 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17735 * gimple-ssa-strength-reduction.c (count_candidates): Change
17736 return value to int.
17737 (analyze_candidates_and_replace): Change type of length to int.
17739 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17741 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
17742 (*vec_extract<mode>): Use VI12_128 mode iterator.
17743 (*vec_extract<mode>_mem): Ditto.
17744 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
17747 2013-05-08 Diego Novillo <dnovillo@google.com>
17752 2012-08-17 Diego Novillo <dnovillo@google.com>
17755 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
17756 * config.in: Regenerate.
17757 * configure: Regenerate.
17758 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
17760 2013-05-08 Jan Hubicka <jh@suse.cz>
17763 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
17764 * cgraph.h (symtab_node_base): Add unique_name.
17765 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17766 input_overwrite_node, input_varpool_node): Stream unique_name.
17767 * cgraphclones.c (cgraph_create_virtual_clone,
17768 cgraph_function_versioning): Set unique_name.
17769 * ipa.c (function_and_variable_visibility): Set unique_name.
17771 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17773 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
17774 (alloc_cand_and_find_basis): Restrict conditional candidate
17775 processing to CAND_MULTs.
17777 2013-05-08 Jan Hubicka <jh@suse.cz>
17780 lto-symtab.c (lto_symtab_symbol_p): New function.
17781 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
17782 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
17783 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
17784 Skip static symbols.
17786 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
17788 PR tree-optimization/57200
17789 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
17790 Only call inform if the preceding warning_at returns true.
17792 2013-05-07 Han Shen <shenhan@google.com>
17794 * cfgexpand.c (record_or_union_type_has_array_p): New function.
17795 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
17796 * common.opt (fstack-protector-strong): New option.
17797 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
17798 * doc/invoke.texi (Optimization Options): Document
17799 "-fstack-protector-strong".
17800 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
17802 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17804 * config/mips/mips.c (mips_machine_reorg2): Return 0.
17806 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
17808 * ira.c (update_equiv_regs): Add insn having equiv memory even if
17809 it is not lhs of the insn.
17810 (setup_reg_equiv): Remove insn having equiv memory which it is not
17812 * lra-constraints.c (process_address): Try to improve generation
17813 code for address base + disp.
17814 (lra_constraints): Make correct the code for checking insn setting
17815 up backward equivalence. Remove insn only if it is in the init
17817 * lra-eliminations.c (update_reg_eliminate): Change return value.
17818 (lra_eliminate): Use the result.
17820 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
17822 * config/i386/sse.md (ssescalarnummask): New mode attribute.
17823 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
17824 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
17825 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
17826 register target operands.
17827 (*vec_extractv8hi_sse2): New pattern.
17828 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
17829 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
17830 (*vec_extract<mode>_mem): New insn and split pattern.
17832 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
17834 * config/arm/arm.c (arm_asan_shadow_offset): New function.
17835 (TARGET_ASAN_SHADOW_OFFSET): Define.
17836 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
17837 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
17839 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17841 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
17842 (incr_vec_index): Return -1 if increment not found.
17843 (create_add_on_incoming_edge): Assert if increment not found.
17844 (record_increment): Limit number of increments recorded.
17845 (all_phi_incrs_profitable): Return false if an increment not found.
17846 (replace_profitable_candidates): Don't process increments that were
17848 (analyze_candidates_and_replace): Limit size of incr_vec.
17850 2013-05-07 Richard Biener <rguenther@suse.de>
17852 * calls.c (special_function_p): setjmp-like functions are leaf.
17853 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
17854 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
17856 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
17858 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
17859 (aarch64_simd_mov<mode>): New expander.
17860 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
17861 (aarch64_simd_mov_to_<mode>high): Likewise.
17862 (aarch64_simd_mov_from_<mode>low): Likewise.
17863 (aarch64_simd_mov_from_<mode>high): Likewise.
17864 (aarch64_dup_lane<mode>): Update.
17865 (aarch64_dup_lanedi): New instruction pattern.
17866 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
17867 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
17869 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17871 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
17872 (replace_mult_candidate): Remove unnecessary argument; remove
17873 unnecessary parameter from call to introduce_cast_before_cand.
17874 (replace_unconditional_candidate): Remove unnecessary parameter
17875 from call to replace_mult_candidate.
17876 (replace_conditional_candidate): Likewise.
17877 (insert_initializers): Use make_temp_ssa_name.
17878 (introduce_cast_before_cand): Remove unnecessary argument; use
17879 make_temp_ssa_name.
17880 (replace_one_candidate): Remove unnecessary argument; remove
17881 unnecessary parameter from calls to introduce_cast_before_cand.
17882 (replace_profitable_candidates): Remove unnecessary parameters
17883 from calls to replace_one_candidate.
17885 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17887 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
17888 phi def as possibly hiding a basis for a CAND_ADD whose operands
17889 have been commuted in the analysis.
17890 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
17892 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17894 * config/aarch64/aarch64.md
17895 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
17896 shift value between 0-4.
17898 2013-05-07 Richard Biener <rguenther@suse.de>
17900 * double-int.h (rshift): New overload.
17901 * double-int.c (rshift): New function.
17902 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
17903 (create_reference_ops_from_ref): Remove.
17904 (vn_reference_insert): Use shared ops for constructing the
17905 reference and copy it.
17907 2013-05-07 Richard Biener <rguenther@suse.de>
17909 PR middle-end/57190
17910 * tree-eh.c (sink_clobbers): Properly propagate
17911 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
17913 2013-05-07 Jakub Jelinek <jakub@redhat.com>
17915 PR tree-optimization/57149
17916 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
17917 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
17918 collect_phi_def_edges, execute_late_warn_uninitialized): Use
17919 uninit_undefined_value_p instead of ssa_undefined_value_p.
17922 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
17923 for modifier == EXPAND_INITIALIZER.
17925 2013-05-07 Anton Blanchard <anton@samba.org>
17927 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
17928 for powerpc64 little endian.
17929 * configure: Regenerate.
17931 2013-05-06 Graham Stott <grahams@btinternet.com>
17933 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
17934 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
17935 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
17936 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
17938 2013-05-06 Graham Stott <grahams@btinternet.com>
17940 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
17941 codes which allow non-lvalues.
17943 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
17945 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
17946 components are all 1s.
17947 (integer_minus_onep): New function.
17948 * tree.h (integer_minus_onep): Declare it.
17949 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
17950 integer_minus_onep instead of integer_all_onesp.
17952 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17955 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
17956 variations of these patterns.
17958 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
17960 * config/i386/i386.md (isa): Add x64_sse4 member.
17961 (enabled): Handle x64_sse4.
17962 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
17963 instruction for 64bit SSE4_1 targets. Update insn attributes.
17964 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
17965 instruction for SSE4_1 targets. Update insn attributes.
17966 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17967 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
17968 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
17970 (*vec_extractv4si): Rename from *sse4_1_pextrd.
17971 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
17972 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
17974 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17977 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
17979 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
17981 * final.c (do_assembler_dialects): Don't handle curly braces and
17982 vertical bar escaped by % as dialect delimiters.
17983 (output_asm_insn): Print curly braces and vertical bar if escaped
17984 by % and ASSEMBLER_DIALECT defined.
17985 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
17986 * doc/tm.texi: Regenerated.
17988 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17990 * config/mips/mips.c: Include tree-pass.h.
17991 (mips_reorg): Split in pre- and post-dbr_schedule parts.
17992 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
17993 (pass_mips_machine_reorg2): New machine specific pass.
17994 (insert_pass_mips_machine_reorg2): New pass plugin definition.
17995 (mips_option_override): Register the new pass.
17996 * rtl.h (cleanup_barriers): Remove prototype.
17997 (dbr_schedule): Likewise.
17998 * jump.c (cleanup_barriers): Make static.
17999 * reorg.c (dbr_schedule): Likewise.
18001 2013-05-06 Richard Biener <rguenther@suse.de>
18003 PR tree-optimization/57185
18004 * tree-parloops.c (add_field_for_reduction): Handle anonymous
18005 SSA names properly.
18007 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
18010 * config/i386/i386.c (add_parameter_dependencies): Add dependence
18011 between "first_arg" and "insn", not "last" and "insn".
18013 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18015 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
18016 (find_candidates_in_block): Re-enable slsr_process_phi.
18017 (create_phi_basis): Fix double counting of candidate adjustment.
18019 2013-05-06 Richard Biener <rguenther@suse.de>
18021 PR middle-end/57147
18022 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
18023 the edge is also fallthru, preserve it and just clear the
18025 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
18026 also complex, preserve that and just clear the fallthru flag.
18027 * tree-inline.c (update_ssa_across_abnormal_edges): Also
18028 update virtual operands.
18030 2013-05-06 Alan Modra <amodra@gmail.com>
18032 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
18033 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
18034 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
18035 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
18036 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
18037 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
18039 2013-05-06 Alan Modra <amodra@gmail.com>
18041 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
18042 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
18043 (DEFAULT_ASM_ENDIAN): Define.
18044 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
18045 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
18046 Update -K PIC clause from sysv4.h.
18047 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
18048 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
18050 2013-05-06 Alan Modra <amodra@gmail.com>
18052 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
18053 twice for little-endian.
18054 (ashrdi3_no_power, ashrdi3): Support little-endian.
18056 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18059 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
18060 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
18062 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
18063 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
18064 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
18067 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
18068 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18070 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
18071 * config/pa/pa.opt: Make mbig-switch a no-op.
18072 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
18073 (CASE_VECTOR_MODE): Always return SImode.
18074 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
18075 for the !TARGET_BIG_SWITCH case.
18076 * config/pa/pa-linux.h: Likewise.
18077 * config/pa/pa-openbsd.h: Likewise.
18078 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
18079 * config/pa/pa.md (short_jump): Remove define_insn.
18080 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
18081 (casesi0): Remove define_insn.
18082 (type): Remove btable_branch.
18083 (pa_combine_type): Likewise.
18084 (in_nullified_branch_delay): Likewise.
18085 (in_call_delay): Likewise.
18086 (define_delay): Likewise.
18087 (define_insn_reservation "Z3"): Likewise.
18088 (define_insn_reservation "Z4"): Likewise.
18089 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
18090 (pa_adjust_insn_length): Remove adjustment for btable branches.
18091 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
18094 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
18096 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
18097 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
18098 Add m->r,x alternatives.
18099 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
18100 splitters using SWI48x mode iterator.
18101 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
18102 TARGET_64BIT. Add m->x alternative.
18103 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
18104 Add o->x alternative. Enable for TARGET_SSE.
18105 (sse_storeq): Remove expander.
18106 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
18107 with memory input operand.
18108 (*vec_extractv2di_1 splitter): New.
18109 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
18110 * config/i386/i386.md (ssevecmodelower): New mode attribute.
18112 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
18114 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
18115 (INT_LOWPART): Delete.
18116 (extract_MB): Adjust.
18117 (extract_ME): Adjust.
18118 (print_operand): Adjust.
18120 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
18122 * config/rs6000/predicates.md (reg_or_add_cint_operand,
18123 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
18124 (reg_or_logical_cint_operand, easy_fp_constant,
18125 logical_const_operand): Delete "CONST_DOUBLE" case.
18126 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
18127 "HOST_BITS_PER_WIDE_INT == 64" test.
18128 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
18129 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
18130 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
18131 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
18133 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
18134 CONST_DOUBLE DImode/VOIDmode case.
18135 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
18136 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
18137 CONST_DOUBLE VOIDmode case.
18138 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
18139 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
18140 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
18141 Delete CONST_DOUBLE case.
18142 (splitters for mov FMOVE64 const_double): Delete
18143 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
18144 "HOST_BITS_PER_WIDE_INT >= 64" test.
18145 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
18147 (mov DI const_double): Delete.
18149 2013-05-04 Jakub Jelinek <jakub@redhat.com>
18151 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
18152 on op shows all bits zero in mode of a lowpart subreg, return zero.
18154 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
18157 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
18158 to save TFmode registers and DImode to save TImode registers for
18159 caller save operations.
18160 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
18161 mark being partially clobbered since they only use the first
18164 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
18165 and TDmode only use the upper 64-bits of each VSX register.
18167 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18169 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
18170 (find_candidates_in_block): Disable slsr_process_phi.
18172 2013-05-03 Guozhi Wei <carrot@google.com>
18174 * coverage.c (coverage_obj_init): Move the construction of gcov
18176 (build_init_ctor): ... here.
18178 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18180 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
18181 (slsr_cand_d): Redefine def_phi.
18182 (stride_status, phi_adjust_status, count_phis_status): New enums.
18183 (find_phi_def): New.
18184 (find_basis_for_base_expr): New.
18185 (find_basis_for_candidate): Handle hidden bases.
18186 (alloc_cand_and_find_basis): Handle phi candidates.
18187 (slsr_process_phi): New.
18188 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
18189 (create_mul_imm_cand): Likewise.
18190 (create_add_ssa_cand): Exclude phi base candidates.
18191 (create_add_imm_cand): Likewise.
18192 (slsr_process_cast): Likewise.
18193 (slsr_process_copy): Likewise.
18194 (find_candidates_in_block): Handle phi candidates.
18195 (dump_candidate): Likewise.
18196 (unconditional_cands): Delete.
18197 (unconditional_cands_with_known_stride_p): Delete.
18198 (phi_dependent_cand_p): New.
18199 (cand_increment): Handle phi-dependent candidates.
18200 (replace_dependent): Delete.
18201 (replace_mult_candidate): New.
18202 (replace_unconditional_candidate): New.
18203 (incr_vec_index): Move to avoid forward reference.
18204 (create_add_on_incoming_edge): New.
18205 (create_phi_basis): New.
18206 (replace_dependents): Delete.
18207 (replace_conditional_candidate): New.
18208 (phi_add_costs): New.
18209 (replace_uncond_cands_and_profitable_phis): New.
18210 (record_increment): Handle phi adjustments.
18211 (record_phi_increments): New.
18212 (record_increments): Handle phi adjustments.
18213 (phi_incr_cost): New.
18214 (lowest_cost_path): Handle phis.
18215 (total_savings): Likewise.
18216 (analyze_increments): Likewise.
18217 (ncd_with_phi): New.
18218 (ncd_of_cand_and_phis): New.
18219 (nearest_common_dominator_for_cands): Handle phi increments.
18220 (all_phi_incrs_profitable): New.
18221 (replace_profitable_candidates): Handle phi-dependent candidates.
18222 (analyze_candidates_and_replace): Likewise.
18224 2013-05-03 Teresa Johnson <tejohnson@google.com>
18227 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
18228 do not exceed REG_BR_PROB_BASE.
18230 2013-05-03 Jeff Law <law@redhat.com>
18232 PR tree-optimization/57144
18233 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
18234 operand of the condition will bit into the new type when eliminating
18235 a cast feeding a condition.
18237 2013-05-03 Jakub Jelinek <jakub@redhat.com>
18239 PR rtl-optimization/57130
18240 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
18241 of COMPARE as in_code to the recursive call if needed.
18243 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
18245 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
18246 (enabled): Handle new members.
18247 * config/i386/sse.md (*vec_concatv2si): Merge from
18248 *vec_concatv2si_sse2 and vec_concatv2si_sse.
18249 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
18251 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
18253 PR tree-optimization/57027
18254 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
18255 for fnms opportunity, check we got the prerequisite kind
18256 of tree / gimple before using accessor functions.
18258 2013-05-03 Richard Biener <rguenther@suse.de>
18260 * double-int.h (lshift): New overload without precision
18261 and arith argument.
18262 (operator *=, operator +=, operator -=): Move ...
18263 * double-int.c (operator *=, operator +=, operator -=): ... here
18264 and implement more efficiently.
18265 (mul_double_with_sign): Remove.
18266 (lshift_double): Adjust to take unsinged shift argument, push
18267 dispatching code to callers.
18268 (mul_double_wide_with_sign): Add early out for callers that
18269 are not interested in high parts or overflow.
18270 (lshift): New function.
18271 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
18272 dispatch code here.
18273 (lrotate, rrotate): Use logical shifts.
18274 * expr.c (get_inner_reference): Use lshift.
18275 * fixed-value.c (do_fixed_divide): Likewise.
18276 * tree-dfa.c (get_ref_base_and_extent): Likewise.
18277 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
18278 (indirect_refs_may_alias_p): Likewise.
18279 (stmt_kills_ref_p_1): Likewise.
18281 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18283 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
18285 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18287 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
18288 scalar form of FABD instruction.
18290 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
18292 * lra-constraints.c (process_alt_operands): Add checking alt
18293 number to choose the best alternative.
18295 2013-05-02 Richard Biener <rguenther@suse.de>
18297 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
18298 bitmap and its handling.
18299 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
18301 2013-05-02 Richard Biener <rguenther@suse.de>
18303 PR middle-end/57140
18304 * tree-inline.c (copy_loops): Properly handle removed loops.
18305 (copy_cfg_body): Mark destination loops for fixup if source
18306 loops needed fixup.
18308 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
18311 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
18312 generating simple_return for naked functions.
18314 2013-05-02 Martin Jambor <mjambor@suse.cz>
18316 PR middle-end/56988
18317 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
18318 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
18320 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
18321 ipa_agg_replacement_value structures.
18322 (known_aggs_to_agg_replacement_list): Likewise.
18323 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
18324 (read_agg_replacement_chain): Likewise.
18325 (ipcp_transform_function): Also check that by_ref flags match.
18327 2013-05-02 Richard Biener <rguenther@suse.de>
18329 * graphds.h (struct graph): Add obstack member.
18330 * graphds.c (new_graph): Initialize obstack and allocate
18332 (add_edge): Allocate edge from the obstack.
18333 (free_graph): Free the obstack instead of all edges and vertices.
18335 2013-05-02 Teresa Johnson <tejohnson@google.com>
18337 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
18339 * cfg.c (update_bb_profile_for_threading): Ditto.
18340 * tree-inline.c (copy_bb): Ditto.
18341 (copy_edges_for_bb): Ditto.
18342 (initialize_cfun): Ditto.
18343 (copy_cfg_body): Ditto.
18344 (expand_call_inline): Ditto.
18345 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
18346 (estimate_node_size_and_time): Ditto.
18347 (inline_merge_summary): Ditto.
18348 * cgraphclones.c (cgraph_clone_edge): Ditto.
18349 (cgraph_clone_node): Ditto.
18350 * sched-rgn.c (compute_dom_prob_ps): Ditto.
18351 (compute_trg_info): Ditto.
18353 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18355 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
18356 S reg when fp attribute set.
18357 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
18359 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18361 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
18363 (*and_one_cmplsi3_compare0_uxtw): Likewise.
18364 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
18365 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
18367 2013-05-02 Richard Biener <rguenther@suse.de>
18369 * tree-scalar-evolution.c (scev_info_hasher): Remove.
18370 (struct instantiate_cache_entry): New type.
18371 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
18372 (struct instantiate_cache_type): New type.
18373 (set_instantiated_value, get_instantiated_value): Remove.
18374 (get_instantiated_value_entry): New function.
18375 (instantiate_scev_name): Use the new cache and adjust.
18376 (instantiate_scev_poly): Adjust.
18377 (instantiate_scev_binary): Likewise.
18378 (instantiate_array_ref): Likewise.
18379 (instantiate_scev_convert): Likewise.
18380 (instantiate_scev_not): Likewise.
18381 (instantiate_scev_3): Likewise.
18382 (instantiate_scev_2): Likewise.
18383 (instantiate_scev_r): Likewise.
18384 (instantiate_scev): Likewise.
18385 (resolve_mixers): Likewise.
18387 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
18390 * lra-constraints.c (best_small_class_operands_num): Remove.
18391 (process_alt_operands): Remove small_class_operands_num. Take
18392 small classes operands into losers and only if the operand is not
18393 matched. Modify debugging output.
18394 (curr_insn_transform): Remove best_small_class_operands_num.
18397 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18399 * config/aarch64/aarch64-builtins.c
18400 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
18401 * config/aarch64/aarch64-simd-builtins.def
18402 (reduc_splus_): Add new modes.
18403 (reduc_uplus_): New.
18404 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
18405 (reduc_uplus_v4sf): Likewise.
18406 (reduc_splus_v4sf): Likewise.
18407 (aarch64_addv<mode>): Likewise.
18408 (reduc_uplus_<mode>): Likewise.
18409 (reduc_splus_<mode>): Likewise.
18410 (aarch64_addvv2di): Likewise.
18411 (reduc_uplus_v2di): Likewise.
18412 (reduc_splus_v2di): Likewise.
18413 (aarch64_addvv2si): Likewise.
18414 (reduc_uplus_v2si): Likewise.
18415 (reduc_splus_v2si): Likewise.
18416 (reduc_<sur>plus_<mode>): New.
18417 (reduc_<sur>plus_v2di): Likewise.
18418 (reduc_<sur>plus_v2si): Likewise.
18419 (reduc_<sur>plus_v4sf): Likewise.
18420 (aarch64_addpv4sf): Likewise.
18421 * config/aarch64/arm_neon.h
18422 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
18423 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
18424 add UNSPEC_SADDV, UNSPEC_UADDV.
18426 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
18428 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18430 * config/aarch64/arm_neon.h
18431 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
18433 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18435 * config/aarch64/aarch64-builtins
18436 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
18438 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18440 * config/aarch64/aarch64-simd-builtins.def
18441 (reduc_smax_): New.
18442 (reduc_smin_): Likewise.
18443 (reduc_umax_): Likewise.
18444 (reduc_umin_): Likewise.
18445 (reduc_smax_nan_): Likewise.
18446 (reduc_smin_nan_): Likewise.
18449 (smax): Update for V2SF, V4SF and V2DF modes.
18452 (smin_nan): Likewise.
18453 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
18454 (<su><maxmin><mode>3): ...This, refactor.
18455 (s<maxmin><mode>3): New.
18456 (<maxmin_uns><mode>3): Likewise.
18457 (reduc_<maxmin_uns>_<mode>): Refactor.
18458 (reduc_<maxmin_uns>_v4sf): Likewise.
18459 (reduc_<maxmin_uns>_v2si): Likewise.
18460 (aarch64_<fmaxmin><mode>: Remove.
18461 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
18463 (vmin<q>_f<32,64>): Likewise.
18464 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18466 (su): Add mappings for smax, smin, umax, umin.
18468 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18469 (FMAXMIN): Rename as...
18470 (FMAXMIN_UNS): ...This.
18472 (fmaxminv): Likewise.
18473 (fmaxmin): Likewise.
18475 (maxmin_uns_op): Likewise.
18477 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18479 * config/aarch64/arm_neon.h
18480 (vac<ge, gt><sd>_f<32, 64>): Rename to...
18481 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
18482 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
18484 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18486 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
18487 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
18489 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18491 * config/aarch64/aarch64-simd.md
18492 (vcond<mode>_internal): Handle special cases for constant masks.
18493 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
18494 (vcondu<mode><mode>): Likewise.
18495 (vcond<v_cmp_result><mode>): New.
18497 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18499 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
18500 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
18501 * config/aarch64/aarch64-simd-builtins.def
18502 (cmeq): Update to BUILTIN_VALLDI.
18507 * config/aarch64/arm_neon.h
18508 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
18509 to builtins or C as appropriate.
18511 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18513 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
18515 (cmhi): Rename to...
18517 * config/aarch64/aarch64-simd.md
18518 (simd_mode): Add SF.
18519 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
18520 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
18521 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
18522 (cstore<mode>_neg): ...This.
18523 * config/aarch64/iterators.md
18525 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
18526 (COMPARISONS): New.
18527 (UCOMPARISONS): Likewise.
18528 (optab): Add missing comparisons.
18534 (VCMP_S): Likewise.
18535 (VCMP_U): Likewise.
18536 (V_cmp_result): Add DF, SF modes.
18537 (v_cmp_result): Likewise.
18539 (vmtype): Likewise.
18540 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
18542 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
18544 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
18545 define_insn to define_insn_and_split.
18546 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
18547 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
18548 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
18549 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
18550 (thumb2_negscc): Likewise.
18552 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18554 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
18556 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18558 * config/arm/thumb2.md: Remove trailing whitespaces.
18560 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
18562 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
18563 Use gen_int_mode rather than GEN_INT.
18565 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
18567 * value-prof.c (stream_in_histogram_value): Remove the strayed
18570 2013-04-30 Richard Biener <rguenther@suse.de>
18572 PR middle-end/57122
18573 * cfghooks.c (split_edge): Properly check for the loop latch edge.
18575 2013-04-30 Richard Biener <rguenther@suse.de>
18577 PR middle-end/57107
18578 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
18580 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
18582 PR rtl-optimization/56957
18583 PR rtl-optimization/57105
18584 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
18585 variable. Use just INSN_UID for determining whether an insn
18586 should be only disconnected from the insn stream.
18587 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
18589 2013-04-30 Jakub Jelinek <jakub@redhat.com>
18591 PR tree-optimization/57104
18592 * tsan.c (instrument_expr): Don't instrument accesses to
18593 DECL_HARD_REGISTER VAR_DECLs.
18595 2013-04-30 Richard Biener <rguenther@suse.de>
18597 * function.h (loops_for_fn): New inline function.
18598 (set_loops_for_fn): Likewise.
18599 * cfgloop.h (place_new_loop): Add struct function parameter.
18600 (get_loop): Likewise.
18601 (get_loops): Likewise.
18602 (number_of_loops): Likewise.
18603 (fel_next): Adjust.
18604 (fel_init): Likewise.
18605 * cfg.c (get_loop_copy): Adjust.
18606 * cfgloop.c (flow_loops_dump): Likewise.
18607 (record_loop_exits): Likewise.
18608 (verify_loop_structure): Likewise.
18609 * cfgloopanal.c (mark_irreducible_loops): Likewise.
18610 (estimate_reg_pressure_cost): Likewise.
18611 (mark_loop_exit_edges): Likewise.
18612 * cfgloopmanip.c (place_new_loop): Likewise.
18613 (add_loop): Likewise.
18614 (duplicate_loop): Likewise.
18615 * graph.c (draw_cfg_nodes): Likewise.
18616 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
18617 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
18618 (extract_affine_chrec): Likewise.
18619 (build_scop_iteration_domain): Likewise.
18620 * graphite.c (graphite_initialize): Likewise.
18621 * ira-build.c (create_loop_tree_nodes): Likewise.
18622 (more_one_region_p): Likewise.
18623 (rebuild_regno_allocno_maps): Likewise.
18624 (mark_loops_for_removal): Likewise.
18625 (mark_all_loops_for_removal): Likewise.
18626 (remove_unnecessary_regions): Likewise.
18627 (ira_build): Likewise.
18628 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
18629 * loop-init.c (fix_loop_structure): Likewise.
18630 (gate_rtl_move_loop_invariants): Likewise.
18631 (gate_rtl_unswitch): Likewise.
18632 (gate_rtl_unroll_and_peel_loops): Likewise.
18633 (rtl_doloop): Likewise.
18634 * lto-streamer-in.c (input_cfg): Likewise.
18635 * lto-streamer-out.c (output_cfg): Likewise.
18636 * modulo-sched.c (sms_schedule): Likewise.
18637 * predict.c (tree_estimate_probability): Likewise.
18638 (tree_estimate_probability_driver): Likewise.
18639 (estimate_loops): Likewise.
18640 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
18641 (move_sese_region_to_fn): Likewise.
18642 (debug_loop_num): Likewise.
18643 * tree-chrec.c (chrec_evaluate): Likewise.
18644 (hide_evolution_in_other_loops_than_loop): Likewise.
18645 (chrec_component_in_loop_num): Likewise.
18646 (reset_evolution_in_loop): Likewise.
18647 (evolution_function_is_invariant_rec_p): Likewise.
18648 * tree-if-conv.c (main_tree_if_conversion): Likewise.
18649 * tree-inline.c (copy_loops): Likewise.
18650 (copy_cfg_body): Likewise.
18651 (tree_function_versioning): Likewise.
18652 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
18653 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
18655 (add_to_evolution_1): Likewise.
18656 (scev_const_prop): Likewise.
18657 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
18658 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
18659 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
18660 (tree_ssa_lim_initialize): Likewise.
18661 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
18662 (verify_loop_closed_ssa): Likewise.
18663 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
18664 (tree_ssa_loop_im): Likewise.
18665 (tree_ssa_loop_unswitch): Likewise.
18666 (tree_vectorize): Likewise.
18667 (check_data_deps): Likewise.
18668 (tree_ssa_loop_ivcanon): Likewise.
18669 (tree_ssa_loop_bounds): Likewise.
18670 (tree_complete_unroll): Likewise.
18671 (tree_complete_unroll_inner): Likewise.
18672 (tree_parallelize_loops): Likewise.
18673 (tree_ssa_loop_prefetch): Likewise.
18674 (tree_ssa_loop_ivopts): Likewise.
18675 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
18676 * tree-vectorizer.c (vectorize_loops): Likewise.
18678 2013-04-29 Mike Frysinger <vapier@gentoo.org>
18680 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
18681 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
18682 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
18683 with EABI_LINK_SPEC.
18685 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18688 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
18691 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
18694 * lra-constraints.c (process_alt_operands): Discourage a bit more
18695 using memory for pseudos. Print cost dump for alternatives.
18696 Modify cost values for conflicts with early clobbers.
18697 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
18699 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18702 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
18704 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18706 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
18707 from/to S register.
18708 (movdi_aarch64): Support LDR/STR from/to D register.
18710 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18712 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
18713 or higher by default.
18715 2013-04-29 Richard Biener <rguenther@suse.de>
18717 PR middle-end/57075
18718 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
18719 even if not adding abnormal edges for calls that can make
18722 2013-04-29 Richard Biener <rguenther@suse.de>
18724 PR middle-end/57103
18725 * tree-cfg.c (move_stmt_op): Fix condition under which to update
18727 (move_stmt_r): Remove redundant checking.
18729 2013-04-29 Teresa Johnson <tejohnson@google.com>
18732 * basic-block.h (apply_scale): New function.
18733 (apply_probability): Use apply_scale.
18734 * gimple-streamer-in.c (input_bb): Ditto.
18735 * lto-streamer-in.c (input_cfg): Ditto.
18736 * lto-cgraph.c (merge_profile_summaries): Ditto.
18737 * tree-optimize.c (execute_fixup_cfg): Ditto.
18738 * tree-inline.c (copy_bb): Update comment to use apply_scale.
18739 (copy_edges_for_bb): Ditto.
18740 (copy_cfg_body): Ditto.
18742 2013-04-29 Tom de Vries <tom@codesourcery.com>
18744 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
18745 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
18746 (tail_merge_optimize): Handle current_loops == NULL.
18748 2013-04-26 Jeff Law <law@redhat.com>
18750 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
18751 (simplify_cond_using_ranges): Generalize code to simplify
18752 COND_EXPRs where one argument is a constant and the other
18753 is an SSA_NAME created by an integral type conversion.
18755 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18757 * config/arm/arm.md (store_minmaxsi): Use only when
18758 optimize_insn_for_size_p.
18760 2013-04-29 Christian Bruel <christian.bruel@st.com>
18763 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
18765 2013-04-29 Richard Biener <rguenther@suse.de>
18767 PR middle-end/57089
18768 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
18769 loop tree make sure to schedule a fixup for the child as well.
18770 (expand_omp_for_generic): Properly add loops.
18771 (expand_omp_for_static_nochunk): Likewise.
18772 (expand_omp_for_static_chunk): Likewise.
18773 (expand_omp_for): For the degenerate case fixup loops.
18774 (expand_omp_sections): Fix default bb placement in loops.
18775 (expand_omp_atomic_pipeline): Properly add loops.
18777 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18779 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
18781 2013-04-29 Tom de Vries <tom@codesourcery.com>
18783 * tree-ssa-tail-merge.c: Update header comment.
18785 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18787 * config/aarch64/arm_neon.h
18788 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
18789 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
18790 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
18791 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
18792 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
18793 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
18794 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
18795 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
18797 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18799 * config/aarch64/aarch64-simd.md
18800 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
18801 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
18802 fix_trunc, fixuns_trunc.
18803 (ftrunc<VDQF:mode>2): New.
18804 * config/aarch64/iterators.md (optab): Add fix, fixuns.
18805 (fix_trunc_optab): New.
18807 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18809 * config/aarch64/aarch64-builtins.c
18810 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
18811 iceilf, lround, iroundf.
18813 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18816 * config/i386/i386.h (enum ix86_tune_indices)
18817 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
18818 New, split from X86_TUNE_INTER_UNIT_MOVES.
18819 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
18820 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
18821 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
18822 (TARGET_INTER_UNIT_MOVES): Remove.
18823 * config/i386/i386.c (initial_ix86_tune_features): Update.
18824 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
18825 (ix86_expand_convert_uns_didf_sse): Use
18826 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18827 (ix86_expand_vector_init_one_nonzero): Ditto.
18828 (ix86_expand_vector_init_interleave): Ditto.
18829 (inline_secondary_memory_needed): Return true for moves from SSE class
18830 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
18831 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
18832 * config/i386/constraints.md (Yi, Ym): Depend on
18833 TARGET_INTER_UNIT_MOVES_TO_VEC.
18834 (Yj, Yn): New constraints.
18835 * config/i386/i386.md (*movdi_internal): Change constraints of
18836 operand 1 from Yi to Yj and from Ym to Yn.
18837 (*movsi_internal): Ditto.
18838 (*movdf_internal): Ditto.
18839 (*movsf_internal): Ditto.
18840 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
18841 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18842 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
18843 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
18844 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
18845 * config/i386/sse.md (movdi_to_sse): Ditto.
18846 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
18847 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
18848 TARGET_INTER_UNIT_MOVES.
18849 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
18850 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
18851 instead of TARGET_INTER_UNIT_MOVES.
18852 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
18853 operand 1 from Yi to Yj and from Ym to Yn.
18855 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18857 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
18858 (float_truncate_hi_): Likewise.
18859 (float_extend_lo_): Likewise.
18860 (float_truncate_lo_): Likewise.
18861 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
18862 (aarch64_float_extend_lo_v2df): Likewise.
18863 (vec_unpacks_hi_v4sf): Likewise.
18864 (aarch64_float_truncate_lo_v2sf): Likewise.
18865 (aarch64_float_truncate_hi_v4sf): Likewise.
18866 (vec_pack_trunc_v2df): Likewise.
18867 (vec_pack_trunc_df): Likewise.
18869 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18871 * config/aarch64/aarch64-builtins.c
18872 (aarch64_fold_builtin): Fold float conversions.
18873 * config/aarch64/aarch64-simd-builtins.def
18874 (floatv2si, floatv4si, floatv2di): New.
18875 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
18876 * config/aarch64/aarch64-simd.md
18877 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
18878 * config/aarch64/iterators.md (FLOATUORS): New.
18879 (optab): Add float, floatuns.
18880 (su_optab): Likewise.
18882 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18884 * config/aarch64/aarch64-builtins.c
18885 (aarch64_builtin_vectorized_function): Use new names for
18887 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
18888 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
18889 (fcvtzu): Split as...
18890 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
18891 (fcvtas): Split as...
18892 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
18893 (fcvtau): Split as...
18894 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
18895 (fcvtps): Split as...
18896 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
18897 (fcvtpu): Split as...
18898 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
18899 (fcvtms): Split as...
18900 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
18901 (fcvtmu): Split as...
18902 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
18903 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
18904 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
18905 (lfrintnusf, lfrintnudf): Likewise.
18906 * config/aarch64/aarch64-simd.md
18907 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
18909 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
18910 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
18911 (fcvt_pattern): Likewise.
18913 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18915 * config/aarch64/aarch64-simd.md
18916 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
18917 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
18919 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18921 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
18922 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
18923 (vrnd<a,m,n,p>_f32): Implement using builtins.
18924 (vrnd<i,x><q>_f<32, 64>): New.
18926 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18928 * config/aarch64/aarch64-builtins.c
18929 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
18930 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
18931 (frintz): Rename to...
18933 (frintp): Rename to...
18935 (frintm): Rename to...
18937 (frinti): Rename to...
18938 (nearbyint): ...this.
18939 (frintx): Rename to...
18941 (frinta): Rename to...
18943 * config/aarch64/aarch64-simd.md
18944 (aarch64_frint<frint_suffix><mode>): Delete.
18945 (<frint_pattern><mode>2): Convert to insn.
18946 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
18947 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
18948 (frint_pattern): Likewise.
18949 (frint_suffix): Likewise.
18951 2013-04-29 Richard Biener <rguenther@suse.de>
18953 PR tree-optimization/57081
18954 * loop-init.c: Include tree-flow.h.
18955 (loop_optimizer_finalize): Free number of iteration estimates.
18956 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
18958 2013-04-29 Jakub Jelinek <jakub@redhat.com>
18960 PR tree-optimization/57083
18961 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
18962 non-singleton shift count range, zero extend low_bound for uns case.
18964 * config/i386/predicates.md (general_vector_operand): New predicate.
18965 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
18966 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
18967 if they aren't nonimmediate operands. If their original values
18968 satisfy const_vector_equal_evenodd_p, don't shift them.
18969 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
18970 predicates. For the SSE4.1 case force operands[{1,2}] into registers
18971 if not nonimmediate_operand.
18972 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
18973 instead of register_operand.
18974 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
18976 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
18978 * stor-layout.c (finalize_size_functions): Allocate a structure and
18979 reset cfun before dumping the functions.
18981 2013-04-27 Jakub Jelinek <jakub@redhat.com>
18983 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
18986 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
18987 use xop_pmacsdqh if uns_p.
18988 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
18989 the immediate rotate count.
18991 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
18993 * rtl.h (struct rtx_def): Add comment for field jump.
18994 (LRA_SUBREG_P): New macro.
18995 * recog.c (register_operand): Check LRA_SUBREG_P.
18996 * lra.c (lra): Add note at the end of RTL code. Align non-empty
18998 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
18999 (lra_final_code_change): Skip subreg change for operators.
19000 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19001 if there are no operand changes.
19002 * lra-constraints.c (curr_insn_set): New.
19003 (match_reload): Set LRA_SUBREG_P.
19004 (emit_spill_move): Ditto.
19005 (check_and_process_move): Use curr_insn_set. Process only single
19006 set insns. Don't initialize sec_mem_p and change_p.
19007 (simplify_operand_subreg): Use LRA_SUBREG_P.
19008 (reg_in_class_p): New function.
19009 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19010 of #ifdef. Add code to remove cycling.
19011 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19012 non-null disp. Reload inner instead of disp when base and index
19013 are null. Try to put lo_sum into register.
19014 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19015 (check_and_process_move): Move code for move cost check to
19016 simple_move_p. Remove equiv_substitution.
19017 (simple_move_p): New function.
19018 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19019 curr_insn_set. Call check_and_process_move only for single set
19020 insns. Use the new function. Move call of check_and_process_move
19021 after operand equiv substitution and address process.
19023 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19026 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
19027 with nonlocal goto receivers or returns twice calls, ignore
19028 unininitialized values from abnormal edges to nl goto receiver
19029 or returns twice call.
19031 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19033 PR tree-optimization/57051
19034 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
19035 and VEC_RSHIFT_EXPR if shift count is a multiple of element
19038 2013-04-26 Richard Biener <rguenther@suse.de>
19040 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
19041 (expand_omp_taskreg): Likewise. Mark loops for fixup.
19042 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
19043 (fixup_loop_arrays_after_move): New function.
19044 (move_sese_region_to_fn): Properly outline the loop tree parts
19045 of the SESE region.
19047 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
19049 * config/i386/i386.md (type, unit): Fix long lines.
19051 2013-04-26 Richard Biener <rguenther@suse.de>
19053 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
19054 (lto-streamer-out.o): Likewise.
19055 * cfgloop.c (init_loops_structure): Export, add struct function
19056 argument and adjust.
19057 (flow_loops_find): Adjust.
19058 * cfgloop.h (enum loop_estimation): Add EST_LAST.
19059 (init_loops_structure): Declare.
19060 * lto-streamer-in.c: Include cfgloop.h.
19061 (input_cfg): Input the loop tree.
19062 * lto-streamer-out.c: Include cfgloop.h.
19063 (output_cfg): Output the loop tree.
19064 (output_struct_function_base): Do not drop PROP_loops.
19066 2013-03-26 Richard Biener <rguenther@suse.de>
19068 * tree-cfg.c (execute_build_cfg): Build the loop tree.
19069 (pass_build_cfg): Provide PROP_loops.
19070 (move_sese_region_to_fn): Remove loops that are outlined into fn
19072 * tree-inline.c: Include cfgloop.h.
19073 (initialize_cfun): Do not drop PROP_loops.
19074 (copy_loops): New function.
19075 (copy_cfg_body): Copy loop structure.
19076 (tree_function_versioning): Initialize destination loop tree.
19077 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
19078 (pass_parallelize_loops): Do IL verification.
19079 * loop-init.c (loop_optimizer_init): Fixup loops if required.
19080 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
19081 the CFG make sure we fixup loops as well.
19082 * tree-ssa-tail-merge.c: Include cfgloop.h.
19083 (replace_block_by): When merging loop latches mark loops for fixup.
19084 * lto-streamer-out.c (output_struct_function_base): Drop
19085 PROP_loops for now.
19086 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
19087 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
19088 * ipa-split.c: Include cfgloop.h.
19089 (split_function): Add the new return block to the loop tree root.
19090 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
19091 whether we have removed the forwarder block.
19092 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
19093 * cfgloop.h (place_new_loop): Declare.
19094 * cfgloopmanip.c (place_new_loop): Export.
19095 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
19096 (tree-switch-conversion.o): Likewise.
19097 (tree-complex.o): Likewise.
19098 (tree-inline.o): Likewise.
19099 (tree-ssa-tailmerge.o): Likewise.
19100 (ipa-split.o): Likewise.
19101 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
19102 (tree-ssa-copy.o): Likewise.
19103 * tree-switch-conversion.c: Include cfgloop.h
19104 (process_switch): If we emit a bit-test cascade, schedule loops
19106 * tree-complex.c: Include cfgloop.h.
19107 (expand_complex_div_wide): Properly add new basic-blocks to loops.
19108 * asan.c: Include cfgloop.h.
19109 (create_cond_insert_point): Properly add new basic-blocks to
19110 loops, schedule loop fixup.
19111 * cfgloop.c (verify_loop_structure): Check that looks are not
19113 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
19115 (expand_omp_for_generic): Likewise.
19116 (expand_omp_sections): Likewise.
19117 (expand_omp_atomic_pipeline): Schedule loops for fixup.
19118 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
19119 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
19120 is initialized, not when loops are present.
19121 * tree-parloops.c (parallelize_loops): Remove checking here.
19122 * passes.c (init_optimization_passes): Schedule a copy-propagation
19123 pass before complete unrolling of inner loops.
19125 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19127 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
19128 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
19129 (colorize_init): Add argument to _WIN32 version.
19130 * toplev.c: Include diagnostic-color.h.
19131 (process_options): Default to -fdiagnostics-color=auto if
19132 GCC_COLORS env var is in the environment.
19133 * common.opt (fdiagnostics-color=): Add Var and Init.
19134 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
19135 env var is in the environment, the default is auto rather than never.
19137 * diagnostic.h (file_name_as_prefix): Add context argument.
19138 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
19139 the string as locus.
19140 * langhooks.c (lhd_print_error_function): Adjust caller.
19142 2013-04-25 Lawrence Crowl <crowl@google.com>
19144 * var-tracking.c (shared_hash_def::htab):
19145 Change type to hash_table. Update dependent calls and types.
19147 2013-04-25 Lawrence Crowl <crowl@google.com>
19149 * Makefile.in: Update as needed below.
19151 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
19152 Move declaration to after the type's method definitons.
19154 * attribs.c (htab_t scoped_attributes::attribute_hash):
19155 Change type to hash_table. Update dependent calls and types.
19157 * bitmap.c (htab_t bitmap_desc_hash):
19158 Change type to hash_table. Update dependent calls and types.
19160 * cselib.c (htab_t cselib_hash_table):
19161 Change type to hash_table. Update dependent calls and types.
19163 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
19164 (hash_string_slot_node): Move implementation into lto-streamer.h
19165 struct string_slot_hasher.
19166 (eq_string_slot_node): Likewise.
19168 * data-streamer-out.c: Update output_block::string_hash_table
19169 dependent calls and types.
19171 * dwarf2cfi.c (htab_t trace_index):
19172 Change type to hash_table. Update dependent calls and types.
19174 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
19175 Change type to hash_table. Update dependent calls and types.
19176 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
19177 (htab_t optimize_external_refs::map): Likewise.
19178 (htab_t output_comp_unit::extern_map): Likewise.
19179 (htab_t output_comdat_type_unit::extern_map): Likewise.
19180 (htab_t output_macinfo::macinfo_htab): Likewise.
19181 (htab_t optimize_location_lists::htab): Likewise.
19182 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
19184 * except.c (htab_t ehspec_hash_type):
19185 Change type to hash_table. Update dependent calls and types.
19186 (assign_filter_values::ttypes): Likewise.
19187 (assign_filter_values::ehspec): Likewise.
19188 (sjlj_assign_call_site_values::ar_hash): Likewise.
19189 (convert_to_eh_region_ranges::ar_hash): Likewise.
19191 * gcse.c (htab_t pre_ldst_table):
19192 Change type to hash_table. Update dependent calls and types.
19194 * ggc-common.c (htab_t saving_htab):
19195 Change type to hash_table. Update dependent calls and types.
19196 (htab_t loc_hash): Likewise.
19197 (htab_t ptr_hash): Likewise.
19198 (call_count): Rename ggc_call_count.
19199 (call_alloc): Rename ggc_call_alloc.
19200 (loc_descriptor): Rename make_loc_descriptor.
19201 (add_statistics): Rename ggc_add_statistics.
19203 * ggc-common.c (saving_htab):
19204 Change type to hash_table. Update dependent calls and types.
19206 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
19207 (push_gimplify_context): Likewise.
19208 (pop_gimplify_context): Likewise.
19209 (struct gimple_temp_hash_elt): Added.
19210 (struct gimplify_hasher): Likewise.
19211 (struct gimplify_ctx.temp_htab):
19212 Change type to hash_table. Update dependent calls and types.
19214 * gimple-fold.c: Include gimplify-ctx.h.
19216 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
19217 Change type to hash_table. Update dependent calls and types.
19218 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
19219 avoid potential global name collision.
19221 * gimplify.c: Include gimplify-ctx.h.
19222 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
19223 (htab_t gimplify_ctx::temp_htab):
19224 Update dependent calls and types for new type hash_table.
19225 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
19226 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
19228 * gimplify-ctx.h: New.
19229 (struct gimple_temp_hash_elt): Move from gimplify.c.
19230 (class gimplify_hasher): New.
19231 (struct gimplify_ctx): Move from gimple.h.
19232 (htab_t gimplify_ctx::temp_htab):
19233 Change type to hash_table. Update dependent calls and types.
19235 * graphite-clast-to-gimple.c: Include graphite-htab.h.
19236 (htab_t ivs_params::newivs_index):
19237 Change type to hash_table. Update dependent calls and types.
19238 (htab_t ivs_params::params_index): Likewise.
19239 (htab_t print_generated_program::params_index): Likewise.
19240 (htab_t gloog::newivs_index): Likewise.
19241 (htab_t gloog::params_index): Likewise.
19243 * graphite.c: Include graphite-htab.h.
19244 4htab_t graphite_transform_loops::bb_pbb_mapping):
19245 Change type to hash_table. Update dependent calls and types.
19247 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
19248 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
19249 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
19251 * graphite-dependences.c: Include graphite-htab.h.
19252 (loop_is_parallel_p): Change hash table type of parameter.
19254 * graphite-htab.h: New.
19255 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
19256 (extern find_pbb_via_hash): Move from graphite-poly.h.
19257 (extern loop_is_parallel_p): Move from graphite-poly.h.
19258 (extern get_loop_body_pbbs): Move from graphite-poly.h.
19260 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
19261 (extern loop_is_parallel_p): Move to graphite-htab.h.
19262 (extern get_loop_body_pbbs): Move to graphite-htab.h.
19264 * haifa-sched.c (htab_t delay_htab):
19265 Change type to hash_table. Update dependent calls and types.
19266 (htab_t delay_htab_i2): Likewise.
19268 * ira-color.c (htab_t allocno_hard_regs_htab):
19269 Change type to hash_table. Update dependent calls and types.
19271 * ira-costs.c (htab_t cost_classes_htab):
19272 Change type to hash_table. Update dependent calls and types.
19274 * loop-invariant.c (htab_t merge_identical_invariants::eq):
19275 Change type to hash_table. Update dependent calls and types.
19277 * loop-iv.c (htab_t bivs):
19278 Change type to hash_table. Update dependent calls and types.
19280 * loop-unroll.c (htab_t opt_info::insns_to_split):
19281 Change type to hash_table. Update dependent calls and types.
19282 (htab_t opt_info::insns_with_var_to_expand): Likewise.
19284 * lto-streamer.h (struct string_slot): Move from data-streamer.h
19285 (struct string_slot_hasher): New.
19286 (htab_t output_block::string_hash_table):
19287 Change type to hash_table. Update dependent calls and types.
19289 * lto-streamer-in.c (freeing_string_slot_hasher): New.
19290 (htab_t file_name_hash_table):
19291 Change type to hash_table. Update dependent calls and types.
19293 * lto-streamer-out.c: Update output_block::string_hash_table dependent
19296 * lto-streamer.c (htab_t tree_htab):
19297 Change type to hash_table. Update dependent calls and types.
19299 * omp-low.c: Include gimplify-ctx.h.
19301 * passes.c (htab_t name_to_pass_map):
19302 Change type to hash_table. Update dependent calls and types.
19303 (pass_traverse): Rename to passes_pass_traverse.
19305 * plugin.c (htab_t event_tab):
19306 Change type to hash_table. Update dependent calls and types.
19308 * postreload-gcse.c (htab_t expr_table):
19309 Change type to hash_table. Update dependent calls and types.
19310 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
19312 * sese.c (debug_rename_map_1): Make extern.
19313 (htab_t copy_bb_and_scalar_dependences::rename_map):
19314 Change type to hash_table. Update dependent calls and types.
19316 * sese.h (extern debug_rename_map): Move to .c file.
19318 * store-motion.c (htab_t store_motion_mems_table):
19319 Change type to hash_table. Update dependent calls and types.
19321 * trans-mem.c (htab_t tm_new_mem_hash):
19322 Change type to hash_table. Update dependent calls and types.
19324 * tree-browser.c (htab_t TB_up_ht):
19325 Change type to hash_table. Update dependent calls and types.
19327 * tree-cfg.c (htab_t discriminator_per_locus):
19328 Change type to hash_table. Update dependent calls and types.
19330 * tree-complex.c: Include tree-hasher.h
19331 (htab_t complex_variable_components):
19332 Change type to hash_table. Update dependent calls and types.
19334 * tree-eh.c (htab_t finally_tree):
19335 Change type to hash_table. Update dependent calls and types.
19337 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
19338 struct int_tree_hasher.
19339 (extern int_tree_map_eq): Likewise.
19340 (uid_decl_map_hash): Removed.
19341 (extern decl_tree_map_eq): Likewise.
19343 * tree-hasher.h: New.
19344 (struct int_tree_hasher): New.
19345 (typedef int_tree_htab_type): New.
19347 * tree-inline.c: Include gimplify-ctx.h.
19349 * tree-mudflap.c: Include gimplify-ctx.h.
19351 * tree-parloops.c: Include tree-hasher.h.
19352 (htab_t eliminate_local_variables_stmt::decl_address):
19353 Change type to hash_table. Update dependent calls and types.
19354 (htab_t separate_decls_in_region::decl_copies): Likewise.
19356 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
19357 Change type to hash_table. Update dependent calls and types.
19359 * tree-sra.c (candidates):
19360 Change type to hash_table. Update dependent calls and types.
19362 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
19364 (int_tree_map_hash): Likewise.
19366 * tree-ssa-dom.c (htab_t avail_exprs):
19367 Change type to hash_table. Update dependent calls and types.
19369 * tree-ssa-live.c (var_map_base_init::tree_to_index):
19370 Change type to hash_table. Update dependent calls and types.
19372 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
19373 Change type to hash_table. Update dependent calls and types.
19375 * tree-ssa-phiopt.c (seen_ssa_names):
19376 Change type to hash_table. Update dependent calls and types.
19378 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
19379 Change type to hash_table. Update dependent calls and types.
19381 * tree-ssa-uncprop.c (equiv):
19382 Change type to hash_table. Update dependent calls and types.
19384 2013-04-25 Jakub Jelinek <jakub@redhat.com>
19386 PR rtl-optimization/57003
19387 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
19388 call note_stores with kill_clobbered_value callback again after
19389 killing regs_invalidated_by_call.
19391 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19393 * config/aarch64/aarch64-simd.md
19394 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
19395 (aarch64_simd_bsl<mode>): Likewise.
19396 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
19398 2013-04-25 Marek Polacek <polacek@redhat.com>
19400 PR tree-optimization/57066
19401 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
19403 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
19405 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
19407 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19409 * config/aarch64/aarch64-builtins.c
19410 (aarch64_fold_builtin): New.
19411 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
19412 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
19413 * config/aarch64/aarch64-simd-builtins.def (abs): New.
19414 * config/aarch64/arm_neon.h
19415 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
19417 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19418 Tejas Belagod <tejas.belagod@arm.com>
19420 * config/aarch64/aarch64-builtins.c
19421 (aarch64_gimple_fold_builtin): New.
19422 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
19423 * config/aarch64/aarch64-simd-builtins.def (addv): New.
19424 * config/aarch64/aarch64-simd.md (addpv4sf): New.
19425 (addvv4sf): Update.
19426 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
19428 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19430 * config/aarch64/aarch64.md
19431 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
19433 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19435 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
19436 (*ngcsi_uxtw): New pattern.
19438 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19439 Julian Brown <julian@codesourcery.com>
19441 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
19442 (TB_DREG): Add T_V4HF.
19443 (v4hf_UP): New macro.
19444 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19445 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19446 Handle initialisation of V4HF. Adjust initialisation of reinterpret
19448 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19449 (arm_vector_mode_supported_p): Handle V4HF.
19450 (arm_mangle_map): Handle V4HFmode.
19451 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
19452 * config/arm/arm_neon_builtins.def: Add entries for
19453 vcvtv4hfv4sf, vcvtv4sfv4hf.
19454 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
19455 (neon_vcvtv4hfv4sf): Likewise.
19456 * config/arm/neon-gen.ml: Handle half-precision floating point
19458 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
19459 * config/arm/arm_neon.h: Regenerate.
19460 * config/arm/neon.ml (type elts): Add F16.
19461 (type vectype): Add T_float16x4, T_floatHF.
19462 (type vecmode): Add V4HF.
19463 (type features): Add Requires_FP_bit feature.
19464 (elt_width): Handle F16.
19465 (elt_class): Likewise.
19466 (elt_of_class_width): Likewise.
19467 (mode_of_elt): Refactor.
19468 (type_for_elt): Handle F16, fix error messages.
19469 (vectype_size): Handle T_float16x4.
19470 (vcvt_sh): New function.
19471 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
19472 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
19473 (string_of_mode): Handle V4HF.
19474 * doc/arm-neon-intrinsics.texi: Regenerate.
19476 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19478 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
19479 format specifier in 'X' case.
19481 2013-04-25 Alan Modra <amodra@gmail.com>
19484 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
19485 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
19486 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
19487 Repeat for many other rotate/shift and mask patterns using subregs.
19488 Name lshiftrt insns.
19489 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
19490 on WORDS_BIG_ENDIAN.
19492 2013-04-25 Alan Modra <amodra@gmail.com>
19494 * config.gcc: Support little-endian powerpc-linux targets.
19495 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
19496 (LINK_OS_LINUX_SPEC): Define.
19497 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
19498 Preserve MASK_LITTLE_ENDIAN.
19499 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
19500 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
19501 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
19502 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
19503 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
19504 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
19505 Correct fp word order for little-endian. Don't shift toc entries
19506 smaller than a word for little-endian.
19507 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
19508 (bswapdi2 splits): Correct low-part subreg for little-endian.
19509 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
19510 low/high where such is correct only for be.
19511 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
19512 little-endian for -mcall-aixdesc.
19514 2013-04-25 Alan Modra <amodra@gmail.com>
19516 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
19517 replace_equiv_address_nv.
19519 2013-04-25 Alan Modra <amodra@gmail.com>
19521 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
19523 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19526 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19527 * rtl.h (struct rtx_def): ...
19529 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19531 PR rtl-optimizations/57046
19532 * lra-constraints (split_reg): Set up lra_risky_transformations_p
19533 for multi-reg splits.
19535 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
19537 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
19539 2013-04-24 Sterling Augustine <saugustine@google.com>
19541 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
19542 (comp_dir_string, debug_str_dwo_section): New.
19543 (DEBUG_STR_DWO_SECTION): Rename to ...
19544 (DEBUG_DWO_STR_SECTION): ... this.
19545 (DEBUG_NORM_STR_SECTION): Delete.
19546 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
19547 (DEBUG_STR_DWO_SECTION_FLAGS): New.
19548 (find_AT_string): Move most logic to ...
19549 (find_AT_string_in_table): ... here. New.
19550 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
19551 add_skeleton_AT_string. Delete logic.
19552 (output_skeleton_debug_sections): Remove call to
19553 add_top_level_skeleton_die_attrs.
19554 (add_comp_dir_attribute): Move logic to comp_dir_string.
19555 (dwarf2out_init): Initialize debug_str_dwo_section.
19556 (output_indirect_string): Call find_string_form.
19557 (output_indirect_strings): Rewrite.
19558 (prune_unused_types): Empty skeleton_debug_str_hash.
19559 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
19560 (dwarf2out_finish): Call output_indirect_strings.
19562 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19564 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
19566 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19568 * rtl.h (struct rtx_def): Add comment for field jump.
19569 (LRA_SUBREG_P): New macro.
19570 * recog.c (register_operand): Check LRA_SUBREG_P.
19571 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19573 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19574 (lra_final_code_change): Skip subreg change for operators.
19575 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19576 if there are no operand changes.
19577 * lra-constraints.c (curr_insn_set): New.
19578 (match_reload): Set LRA_SUBREG_P.
19579 (emit_spill_move): Ditto.
19580 (check_and_process_move): Use curr_insn_set. Process only single
19581 set insns. Don't initialize sec_mem_p and change_p.
19582 (simplify_operand_subreg): Use LRA_SUBREG_P.
19583 (reg_in_class_p): New function.
19584 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19585 of #ifdef. Add code to remove cycling.
19586 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19587 non-null disp. Reload inner instead of disp when base and index
19588 are null. Try to put lo_sum into register.
19589 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19590 (check_and_process_move): Move code for move cost check to
19591 simple_move_p. Remove equiv_substitution.
19592 (simple_move_p): New function.
19593 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19594 curr_insn_set. Call check_and_process_move only for single set
19595 insns. Use the new function. Move call of check_and_process_move
19596 after operand equiv substitution and address process.
19598 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
19600 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
19601 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
19602 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
19604 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19606 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
19608 2013-04-24 Marek Polacek <polacek@redhat.com>
19610 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
19611 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
19612 (select_loops_exit_conditions): Likewise.
19613 (number_of_iterations_for_all_loops): Likewise.
19614 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
19615 (scev_analysis): Likewise.
19617 2013-04-02 Catherine Moore <clm@codesourcery.com>
19618 Chao-ying Fu <fu@mips.com>
19620 * config/mips/micromips.md (jraddiusp): New pattern.
19621 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
19622 instruction if possible.
19624 2013-04-24 Alan Modra <amodra@gmail.com>
19626 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
19628 2013-04-24 Julian Brown <julian@codesourcery.com>
19629 Chung-Lin Tang <cltang@codesourcery.com>
19631 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
19632 dependency behavior in enumeration type DIE generation. Add TODO note
19633 to comments about future DW_FORM_sdata/udata re-work of related code.
19635 2013-04-23 Lawrence Crowl <crowl@google.com>
19637 * Makefile.in: Update as needed below.
19639 * hash-table.h (class hash_table):
19640 Correct many methods with parameter types compare_type to the correct
19641 value_type. (Correct code was unlikely to notice the change.)
19642 (hash_table::elements_with_deleted) New.
19643 (class hashtable::iterator): New.
19644 (hashtable::begin()): New.
19645 (hashtable::end()): New.
19646 (FOR_EACH_HASH_TABLE_ELEMENT): New.
19648 * statistics.c (statistics_hashes):
19649 Change type to hash_table. Update dependent calls and types.
19651 * tree-into-ssa.c (var_infos):
19652 Change type to hash_table. Update dependent calls and types.
19654 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
19655 Change type to hash_table. Update dependent calls and types.
19657 * tree-ssa-loop-im.c (struct mem_ref.refs):
19658 Change type to hash_table. Update dependent calls and types.
19660 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
19661 Change type to hash_table. Update dependent calls and types.
19663 * tree-ssa-sccvn.c (vn_tables_s::nary):
19664 Change type to hash_table. Update dependent calls and types.
19665 (vn_tables_s::phis): Likewise.
19666 (vn_tables_s::references): Likewise.
19668 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
19669 (vn_reference_eq): Update parameter and return types.
19671 * tree-ssa-structalias.c (pointer_equiv_class_table):
19672 Change type to hash_table. Update dependent calls and types.
19673 (location_equiv_class_table): Likewise.
19675 * tree-vect-data-refs.c: Consequential changes for making
19676 peeling a hash_table.
19678 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
19679 (destroy_loop_vec_info): Dependent hash_table update.
19681 * tree-vectorizer.h (peeling_htab):
19682 Change type to hash_table. Update dependent calls and types.
19684 2013-04-23 Shiva Chen <shiva0217@gmail.com>
19686 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
19687 to check the register content is equal or not.
19688 * lra-constraints.c (match_reload): Use lra_assign_reg_val
19689 to assign register content record.
19690 * lra-eliminations.c (update_reg_eliminate): Use
19691 lra_update_reg_val_offset to update register content offset.
19692 * lra-int.h (struct lra_reg): Add offset member.
19693 (lra_reg_val_equal_p): New static inline function.
19694 (lra_update_reg_val_offset): New static inline function.
19695 (lra_assign_reg_val): New static inline function.
19696 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
19697 to assign register content record.
19698 (initialize_lra_reg_info_element): Initial offset to zero.
19700 2013-04-23 Catherine Moore <clm@codesourcery.com>
19702 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
19703 operands. Record compression.
19705 2013-04-23 Xinliang David Li <davidxl@google.com>
19707 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
19709 2013-04-23 Richard Biener <rguenther@suse.de>
19711 PR middle-end/57036
19712 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
19713 parameter, only add abnormal goto edges from the copied body
19714 if the call could perform abnormal gotos.
19715 (copy_cfg_body): Adjust.
19717 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
19719 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
19721 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
19723 * coretypes.h (gimple_stmt_iterator): Add struct to make
19726 2013-04-23 Richard Biener <rguenther@suse.de>
19728 PR tree-optimization/57026
19729 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
19730 from SSA names occuring in abnormal PHI nodes.
19732 2013-04-22 Andi Kleen <ak@linux.intel.com>
19734 * lto/lto.c (print_lto_report_1): Fix LTO report names.
19736 2013-04-22 Andi Kleen <ak@linux.intel.com>
19738 * lto/lto.c (print_lto_report_1): Declare early.
19739 (read_cgraph_and_symbols): Call print_lto_report_1 early.
19741 2013-04-22 Andi Kleen <ak@linux.intel.com>
19743 * common.opt (-flto-report-wpa): Add.
19744 * doc/invoke.texi (-flto-report-wpa): Add.
19745 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
19748 2013-04-22 Xinliang David Li <davidxl@google.com>
19750 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
19751 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
19752 * Makefile.in: New dependency
19754 David Daney <ddaney.cavm@gmail.com>
19756 * configure.ac (gcc_cv_as_micromips_support): Use the
19757 --fatal-warnings option.
19758 * configure: Regenerate.
19760 2013-04-22 Marek Polacek <polacek@redhat.com>
19763 * tsan.c (instrument_expr): Don't instrument expression
19764 in case its size is zero.
19766 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
19770 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19772 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19774 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19776 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
19777 (gimple_stmt_iterator): New typedef.
19778 * gimple.h (gimple_stmt_iterator): Rename to...
19779 (gimple_stmt_iterator_d): ... This.
19780 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
19781 trees be valid for GIMPLE and GENERIC.
19782 (TARGET_GIMPLE_FOLD_BUILTIN): New.
19783 * gimple-fold.c (gimple_fold_call): Call target hook
19784 gimple_fold_builtin.
19785 * hooks.c (hook_bool_gsiptr_false): New.
19786 * hooks.h (hook_bool_gsiptr_false): New.
19787 * target.def (fold_stmt): New.
19788 * doc/tm.texi: Regenerate.
19790 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
19793 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
19794 a set sp if no stack realignment.
19796 2013-04-22 Nick Clifton <nickc@redhat.com>
19798 * config.gcc (tilegx-linux): Extend extra_objs rather than
19800 (tilepro-linux): Likewise.
19802 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19804 * config/aarch64/aarch64-builtins.c
19806 (CF0, CF1, CF2, CF3, CF4, CF10): New.
19807 (VAR<1-12>): Add MAP parameter.
19808 (BUILTIN_*): Likewise.
19809 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
19810 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
19811 (aarch64_ushl_n<mode>): Likewise.
19812 (aarch64_sshr_n<mode>): Likewise.
19813 (aarch64_ushr_n<mode>): Likewise.
19814 (aarch64_<maxmin><mode>): Likewise.
19815 (aarch64_sqrt<mode>): Likewise.
19816 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
19817 (vshr<q>_n_*): Likewise.
19819 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19821 * config/aarch64/aarch64-builtins.c
19822 (aarch64_simd_builtin_type_mode): Handle SF types.
19824 (BUILTIN_GPF): Define.
19825 (aarch64_init_simd_builtins): Handle SF types.
19826 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
19827 (frecps): Likewise.
19828 (frecpx): Likewise.
19829 * config/aarch64/aarch64-simd.md
19830 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
19831 (aarch64_frecpe<mode>): New.
19832 (aarch64_frecps<mode>): Likewise.
19833 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
19834 (v8type): Add frecp<esx>.
19835 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
19836 (aarch64_frecps<mode>): Likewise.
19837 * config/aarch64/iterators.md (FRECP): New.
19838 (frecp_suffix): Likewise.
19839 * config/aarch64/arm_neon.h
19840 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
19842 2013-04-22 Christian Bruel <christian.bruel@st.com>
19845 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
19846 (REG_CLASS_NAMES): Idem.
19847 (REG_CLASS_CONTENTS): Idem.
19848 (REGCLASS_HAS_FP_REG): Idem.
19849 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
19850 (sh_conditional_register_usage): Idem.
19852 2013-04-21 Jeff Law <law@redhat.com>
19854 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
19855 (ssa_forward_propagate_and_combine): Use it.
19857 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19859 * lra.c: Update the flow chart diagram.
19861 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19863 PR rtl-optimization/56847
19864 * lra-constraints.c (process_alt_operands): Discourage alternative
19865 with non-matche doffsettable memory constraint fro memory with
19868 2013-04-19 Richard Biener <rguenther@suse.de>
19870 PR tree-optimization/56982
19871 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
19873 * gimplify.c (gimplify_call_expr): Notice special calls.
19874 (gimplify_modify_expr): Likewise.
19875 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
19876 abnormal control flow receivers.
19877 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
19878 in the same way as cfun->has_nonlocal_labels.
19879 (gimple_purge_dead_abnormal_call_edges): Likewise.
19880 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
19881 receivers start a basic-block.
19883 2013-04-19 Richard Biener <rguenther@suse.de>
19885 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
19887 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
19888 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
19889 (SLP_TREE_LOAD_PERMUTATION): Add.
19890 (vect_transform_slp_perm_load): Adjust prototype.
19891 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
19892 (vect_free_slp_instance): Likewise.
19893 (vect_create_new_slp_node): Likewise.
19894 (vect_supported_slp_permutation_p): Remove.
19895 (vect_slp_rearrange_stmts): Adjust.
19896 (vect_supported_load_permutation_p): Likewise. Inline
19897 vect_supported_slp_permutation_p here.
19898 (vect_analyze_slp_instance): Compute load permutations per
19899 slp node instead of per instance.
19900 (vect_get_slp_defs): Adjust.
19901 (vect_transform_slp_perm_load): Likewise.
19902 (vect_schedule_slp_instance): Remove redundant code.
19903 (vect_schedule_slp): Remove hack for PR56270, add it ...
19904 * tree-vect-stmts.c (vectorizable_load): ... here, do not
19905 CSE loads for SLP. Adjust.
19907 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19909 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
19910 spelling in two comments.
19912 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19915 * config/arm/arm.c (load_multiple_sequence): Require SP
19916 as base register for loads if SP is in the register list.
19918 2013-04-19 Martin Jambor <mjambor@suse.cz>
19920 PR tree-optimization/56718
19921 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
19922 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
19923 and made public. Adjusted all callers.
19924 (ipa_intraprocedural_devirtualization): New function.
19925 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
19926 (ipa_intraprocedural_devirtualization): Likewise.
19927 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
19929 2013-04-19 Richard Biener <rguenther@suse.de>
19931 PR tree-optimization/57000
19932 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
19934 2013-04-19 Terry Guo <terry.guo@arm.com>
19936 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
19938 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
19939 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
19940 (cortex_m4_fmacs): Use new reservations.
19941 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
19943 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
19945 PR rtl-optimization/56999
19946 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
19948 (lra_coalesce): Remove split_origin_bitmap and related code.
19949 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
19950 ranges if necessary.
19952 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
19954 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
19956 (ix86_expand_call): Remove clobbered_registers array and use
19957 x86_64_ms_sysv_extra_clobbered_registers instead.
19958 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
19960 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
19962 * config/i386/i386.md (*call_rex64_ms_sysv): Use
19963 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
19964 (*call_value_rex64_ms_sysv): Ditto.
19966 2013-04-18 Cary Coutant <ccoutant@google.com>
19968 * dwarf2out.c (output_pubnames): Check die_perennial_p of
19969 parent instead of die_mark.
19971 2013-04-18 Diego Novillo <dnovillo@google.com>
19973 * gimple.c (create_gimple_tmp): New.
19974 (get_expr_type): New.
19975 (build_assign): New.
19976 (build_type_cast): New.
19977 * gimple.h (enum ssa_mode): Define.
19978 (gimple_seq_set_location): New.
19979 * asan.c (build_check_stmt): Change some gimple_build_* calls
19980 to use build_assign and build_type_cast.
19982 2013-04-18 Richard Biener <rguenther@suse.de>
19984 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
19985 handle negative step. Remove redundant checks.
19986 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
19987 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
19988 for negative step and grouped loads fail to vectorize.
19990 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19992 * emit-rtl.c (reset_insn_used_flags): New function.
19993 (reset_all_used_flags): Use it.
19994 (verify_insn_sharing): New function.
19995 (verify_rtl_sharing): Fix verification for SEQUENCEs.
19997 2013-04-18 Jakub Jelinek <jakub@redhat.com>
19999 PR tree-optimization/56984
20000 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
20001 and (x >> M) >= N don't register any assertion if N << M is the
20004 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
20006 * lower-subreg.c (resolve_simple_move): If called self-recursive,
20007 do not delete_insn insns that have not yet been emitted, only
20008 unlink them with remove_insn.
20009 * df-scan.c (df_insn_delete): Revert r197492.
20011 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
20013 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
20014 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
20016 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
20018 * config/arm/arm.md (movsicc_insn): Convert define_insn into
20019 define_insn_and_split.
20020 (and_scc,ior_scc,negscc): Likewise.
20021 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
20023 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
20025 * config/arm/arm.c (use_return_insn): Return 0 for targets that
20026 can benefit from using a sequence of LDRD instructions in epilogue
20027 instead of a single LDM instruction.
20029 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
20032 * doc/extend.texi: Fix typo.
20034 2013-04-17 Richard Biener <rguenther@suse.de>
20036 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
20037 (vect_build_slp_tree): ... here.
20038 (vect_build_slp_tree_1): Compute which stmts of the SLP group
20039 match. Remove special-casing of mismatched complex loads.
20040 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
20041 re-try the match with swapped commutative operands.
20042 (vect_supported_load_permutation_p): Remove special-casing of
20043 mismatched complex loads.
20044 (vect_analyze_slp_instance): Adjust.
20046 2013-04-17 Richard Biener <rguenther@suse.de>
20048 PR rtl-optimization/56921
20049 * cfgloop.h (struct loop): Add simple_loop_desc member.
20050 (struct niter_desc): Mark with GTY(()).
20051 (simple_loop_desc): Do not use aux field but simple_loop_desc.
20052 * loop-iv.c (get_simple_loop_desc): Likewise.
20053 (free_simple_loop_desc): Likewise.
20056 2013-04-16 Richard Biener <rguenther@suse.de>
20058 PR rtl-optimization/56921
20059 * loop-init.c (pass_rtl_move_loop_invariants): Add
20060 TODO_do_not_ggc_collect to todo_flags_finish.
20061 (pass_rtl_unswitch): Same.
20062 (pass_rtl_unroll_and_peel_loops): Same.
20063 (pass_rtl_doloop): Same.
20065 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
20067 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
20068 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
20069 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
20071 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
20072 * tree-streamer.c (record_common_node): Adjust reference in comment.
20074 2013-04-17 Terry Guo <terry.guo@arm.com>
20076 * config/arm/cortex-m4.md: Add a new bypass.
20078 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20080 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
20082 (*subs_<optab><mode>_multp2): New pattern.
20083 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
20084 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
20086 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20088 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
20089 (*subs_mul_imm_<mode>): New pattern.
20091 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
20094 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
20095 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
20096 (vsx_movti_32bit): Change j->wa to O->wa.
20098 2013-04-16 Richard Biener <rguenther@suse.de>
20100 PR rtl-optimization/56921
20101 * loop-init.c (pass_rtl_move_loop_invariants): Add
20102 TODO_do_not_ggc_collect to todo_flags_finish.
20103 (pass_rtl_unswitch): Same.
20104 (pass_rtl_unroll_and_peel_loops): Same.
20105 (pass_rtl_doloop): Same.
20107 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
20109 * config/arm/arm.c (emit_multi_reg_push): New declaration
20110 for an existing function.
20111 (arm_emit_strd_push): New function.
20112 (arm_expand_prologue): Used here.
20113 (arm_emit_ldrd_pop): New function.
20114 (arm_expand_epilogue): Used here.
20115 (arm_get_frame_offsets): Update condition.
20116 (arm_emit_multi_reg_pop): Add a special case for load of a single
20117 register with writeback.
20119 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
20121 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
20124 2013-04-16 Richard Biener <rguenther@suse.de>
20126 PR tree-optimization/56756
20127 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
20128 (first_mem_ref_loc): New.
20129 (execute_sm): Place the load temporarily before a previous
20130 access instead of in the latch edge to ensure its SSA dependencies
20131 are defined at points dominating the load.
20133 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
20135 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
20136 correct fix by moving header and footer insn to the footer of
20137 the merged basic block. Clear BB_END of the merged-away block.
20139 PR middle-end/43631
20140 * emit-rtl.c (make_note_raw): New function.
20141 (link_insn_into_chain): New static inline function.
20142 (add_insn): Use it.
20143 (add_insn_before, add_insn_after): Factor insn chain linking code...
20144 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
20145 using link_insn_into_chain.
20146 (note_outside_basic_block_p): New helper function for emit_note_after
20147 and emit_note_before.
20148 (emit_note_after): Use nobb variant of add_insn_after if the note
20149 should not be contained in a basic block.
20150 (emit_note_before): Use nobb variant of add_insn_before if the note
20151 should not be contained in a basic block.
20152 (emit_note_copy): Use make_note_raw.
20153 (emit_note): Likewise.
20154 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
20155 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
20156 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
20157 the moved barrier the tail of the basic block it follows.
20158 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
20160 2013-04-15 Jakub Jelinek <jakub@redhat.com>
20162 PR tree-optimization/56962
20163 * gimple-ssa-strength-reduction.c (record_increment): Only set
20164 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
20165 either rhs1 or rhs2 is equal to c->base_expr.
20167 2013-04-15 Richard Biener <rguenther@suse.de>
20169 PR tree-optimization/56933
20170 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
20172 (GROUP_READ_WRITE_DEPENDENCE): Remove.
20173 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
20174 * tree-vect-data-refs.c (vect_analyze_group_access): Move
20175 dependence check ...
20176 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
20178 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
20179 GROUP_READ_WRITE_DEPENDENCE.
20181 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20183 * emit-rtl.c (reset_all_used_flags): New function.
20184 (verify_rtl_sharing): Call reset_all_used_flags before and after
20185 performing the checks.
20187 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20189 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
20190 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
20191 * config/arm/constraints.md (De): New constraint.
20192 * config/arm/neon.md (anddi3_neon): Delete.
20193 (neon_vand<mode>): Expand to standard anddi3 pattern.
20194 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
20195 Move earlier in the file.
20196 (neon_inv_logic_op2): Likewise.
20197 (arm_anddi_operand_neon): New predicate.
20199 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20201 * configure.ac (gcc_cv_ld_as_needed): Set
20202 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
20203 Use -z ignore, -z record on *-*-solaris2*.
20204 (HAVE_LD_AS_NEEDED): Update comment.
20205 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
20206 * configure: Regenerate.
20207 * config.in: Regenerate.
20208 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
20209 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
20210 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
20211 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
20212 equivalents. Fix markup.
20213 * doc/tm.texi: Regenerate.
20215 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
20217 * config/i386/i386.opt: New option mstack-protector-guard=.
20218 * config/i386/i386-opts.h: Add enum stack_protector_guard.
20219 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
20220 TARGET_SSP_TLS_GUARD.
20221 * config/i386/i386.c (ix86_option_override_internal): Set
20222 ix86_stack_protector_guard.
20223 * config/i386/i386.md (stack_protect_set): Enable for
20224 TARGET_SSP_TLS_GUARD only.
20225 (stack_protect_set_<mode>): Ditto.
20226 (stack_protect_test): Ditto.
20227 (stack_protect_test_<mode>): Ditto.
20228 * doc/invoke.texi (i386 Option): Document.
20230 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
20233 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
20234 (S_MODES): Set H_MODE bit.
20235 (SF_MODES): Set only S_MODE and SF_MODE bits.
20236 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
20237 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
20238 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
20239 <MODE_FLOAT>: Likewise.
20241 2013-04-15 Joey Ye <joey.ye@arm.com>
20243 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
20245 2013-04-15 Joey Ye <joey.ye@arm.com>
20247 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
20249 (thumb_far_jump_used_p): Count instruction size and set
20252 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
20254 * reorg.c (fill_simple_delay_slots): Reindent block of code.
20255 * resource.c (mark_target_live_regs): Reformat conditional block.
20257 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20259 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
20260 notes, they are emitted only just before final.
20261 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
20263 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20265 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
20266 * cfgrtl.c (delete_insn): Call it here instead.
20267 * lra-spills.c (lra_final_code_change): Use delete_insn.
20268 * haifa-sched.c (sched_remove_insn): Likewise.
20269 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
20270 returning to the nop pool.
20271 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
20272 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
20274 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20276 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
20277 * doc/tm.texi: Regenerated.
20279 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
20281 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
20284 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20286 * df-core.c (df_find_def): Compare register numbers.
20287 (df_find_use): Likewise.
20289 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
20292 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
20293 lra_in_progress for return.
20295 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20297 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
20298 define_insn into define_insn_and_split and emit movsicc patterns.
20300 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20302 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
20304 2013-04-12 Richard Biener <rguenther@suse.de>
20306 * tree-pass.h (TODO_do_not_ggc_collect): New.
20307 * passes.c (execute_one_ipa_transform_pass): Honor
20308 TODO_do_not_ggc_collect.
20309 (execute_one_pass): Likewise.
20312 2013-04-10 Richard Biener <rguenther@suse.de>
20314 * passes.c (init_optimization_passes): Remove reload pass.
20315 * ira.c (do_reload): Merge into ...
20317 (rest_of_handle_reload): Remove.
20318 (pass_reload): Likewise.
20319 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20320 of reload for vzeroupper pass placement.
20322 2013-04-12 Jakub Jelinek <jakub@redhat.com>
20324 PR tree-optimization/56918
20325 PR tree-optimization/56920
20326 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
20327 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
20328 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
20329 use wide_mul_with_sign method.
20331 2013-04-12 Richard Biener <rguenther@suse.de>
20333 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
20334 not be considered a gimple constant.
20336 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
20338 * fold-const.c (const_binop): Handle vector shifts by a scalar.
20339 (fold_binary_loc): Call const_binop also for mixed vector-scalar
20342 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
20343 Jakub Jelinek <jakub@redhat.com>
20345 * opts.c: Include diagnostic-color.h.
20346 (common_handle_option): Handle OPT_fdiagnostics_color_.
20347 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
20348 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
20349 (diagnostic-color.o): New.
20350 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
20351 (diagnostic_color_rule): New enum.
20352 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
20353 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
20354 the location string.
20355 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
20356 either NULL, or color kind.
20357 * diagnostic-color.c: New file.
20358 * diagnostic-color.h: New file.
20359 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
20361 * doc/invoke.texi (-fdiagnostics-color): Document.
20362 * pretty-print.h (pp_show_color): Define.
20363 (struct pretty_print_info): Add show_color field.
20364 * diagnostic.c: Include diagnostic-color.h.
20365 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
20366 macros. Colorize error:, warning: etc. strings and also the location
20368 (diagnostic_show_locus): Colorize the caret line.
20369 * pretty-print.c: Include diagnostic-color.h.
20370 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
20371 inside of %< %> quotes or quoted through q format modifier.
20373 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20375 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
20377 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20379 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
20380 code in CC_NZ mode.
20381 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
20384 2013-04-11 Marek Polacek <polacek@redhat.com>
20386 PR tree-optimization/48184
20387 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
20389 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
20391 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
20392 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
20393 (skip_simple_arithmetic): Tidy up.
20394 * tree.h (skip_simple_constant_arithmetic): Declare.
20396 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20398 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
20400 2013-04-11 Richard Biener <rguenther@suse.de>
20402 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20403 generate vector constants.
20405 2013-04-11 Richard Biener <rguenther@suse.de>
20407 PR tree-optimization/56878
20408 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
20409 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
20411 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
20412 Prefer to align the DR with the most invariant base address.
20414 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20416 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
20419 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
20421 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
20422 floating-point vector comparisons against 0.
20424 2013-04-11 Jakub Jelinek <jakub@redhat.com>
20426 PR tree-optimization/56899
20427 * fold-const.c (extract_muldiv_1): Apply distributive law
20428 only if TYPE_OVERFLOW_WRAPS (ctype).
20430 2013-04-11 Bin Cheng <bin.cheng@arm.com>
20433 * ira-costs.c (scan_one_insn): Check whether the source rtx of
20434 loading has side effect.
20436 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
20438 * config/sparc/sparc.c: Include tree-pass.h.
20439 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
20440 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
20441 head of file. Change return type. Split off gate function.
20442 (sparc_gate_work_around_errata): New function.
20443 (pass_work_around_errata): New pass definition.
20444 (insert_pass_work_around_errata) New pass insert definition to
20445 insert pass_work_around_errata just after delayed-branch scheduling.
20446 (sparc_option_override): Insert the pass.
20447 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
20449 2013-04-10 David S. Miller <davem@davemloft.net>
20451 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
20452 or -mcpu=hypersparc.
20454 * target.def (cstore_mode): New hook.
20455 * target.h: Include insn-codes.h
20456 * targhooks.c: Likewise.
20457 (default_cstore_mode): New function.
20458 * targhooks.h: Declare it.
20459 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
20460 * doc/tm.texi: Rebuild.
20461 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
20462 target hook, rather than inspecting the insn_data.
20463 * config/sparc/sparc.c (sparc_cstore_mode): New function.
20464 (TARGET_CSTORE_MODE): Redefine.
20465 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
20467 * config/sparc/predicates.md (cstore_result_operand): New special
20469 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
20470 Use it for operand 0.
20471 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
20472 (*snesi_special): Likewise.
20473 (*snesi_zero): Likewise.
20474 (*seqsi_zero): Likewise.
20475 (*sltu_insn): Likewise.
20476 (*sgeu_insn): Likewise.
20477 (*seqdi_special): Make operand 0 and comparison operation be of
20479 (*snedi_special): Likewise.
20480 (*snedi_special_vis3): Likewise.
20481 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
20482 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
20483 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
20484 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
20485 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
20486 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
20487 (*sltu_extend_sp64): Likewise.
20488 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
20489 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
20490 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
20491 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
20492 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
20494 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
20496 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
20497 (aarch64_start_file): Use the new function.
20499 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20500 Jason Merrill <jason@redhat.com>
20502 * common.opt: Add -gdwarf.
20503 * opts.c (common_handle_option): Handle it.
20504 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
20506 2013-04-10 Richard Biener <rguenther@suse.de>
20508 * passes.c (execute_todo): Do not call ggc_collect conditional here.
20509 (execute_one_ipa_transform_pass): But unconditionally here.
20510 (execute_one_pass): And here.
20511 (init_optimization_passes): Remove reload pass.
20512 * tree-pass.h (TODO_ggc_collect): Remove.
20513 (pass_reload): Likewise.
20514 * ira.c (do_reload): Merge into ...
20516 (rest_of_handle_reload): Remove.
20517 (pass_reload): Likewise.
20518 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20519 of reload for vzeroupper pass placement.
20520 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
20521 and todo_flags_finish of all passes.
20523 2013-04-10 Richard Biener <rguenther@suse.de>
20525 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
20526 first_const_oprnd field, rename first_def_type to first_op_type.
20527 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
20528 (vect_get_and_check_slp_defs): Always use the type of the
20529 operand. Allow mixed vect_external_def, vect_constant_def types.
20530 (vect_get_constant_vectors): Handle mixed vect_external_def,
20531 vect_constant_def types.
20533 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
20535 PR tree-optimization/55524
20536 * tree-ssa-math-opts.c
20537 (convert_mult_to_fma): Don't use an fms construct
20538 when we don't have an fms operation, but fnma, and it looks
20539 likely that we'll be able to use the latter.
20541 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
20543 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
20545 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
20546 inline fail caused by overwritable functions.
20548 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
20550 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
20551 unnecessary bits in the constant power of two case.
20553 2013-04-10 Richard Biener <rguenther@suse.de>
20555 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
20556 broken code swapping operands.
20557 (vect_build_slp_tree): Do not compute load permutations here.
20558 (vect_analyze_slp_instance): Compute load permutations here,
20559 after building the SLP tree.
20561 2013-04-09 Christian Bruel <christian.bruel@st.com>
20563 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
20564 of next/prev_real_insn.
20566 2013-04-09 Jan Hubicka <jh@suse.cz>
20568 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
20569 Drop aliased parameter.
20570 (function_and_variable_visibility): Do not handle alias pairs.
20571 * cgraph.c (varpool_externally_visible_p): Update prototype.
20572 * varpool.c (varpool_add_new_variable): Update.
20574 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20576 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
20578 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
20580 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
20582 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
20584 2013-04-09 Marek Polacek <polacek@redhat.com>
20586 PR tree-optimization/48762
20587 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
20589 2013-04-09 Richard Biener <rguenther@suse.de>
20591 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
20593 (vect_build_slp_tree): Likewise.
20594 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
20595 calculating the cost of a SLP instance.
20596 (vect_analyze_slp_instance): Use it from here, after building
20599 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20601 PR middle-end/56883
20602 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
20603 expand_omp_for_static_chunk): Use simple_p = true in
20604 force_gimple_operand_gsi calls when assigning to addressable decls.
20606 2013-04-09 Jeff Law <law@redhat.com>
20608 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
20609 when the boolean was created by converting a wider object which
20610 had a boolean range.
20612 2013-04-09 Richard Biener <rguenther@suse.de>
20614 * tree-vectorizer.h (slp_void_p): Remove.
20615 (slp_tree): Typedef before _slp_tree declaration.
20616 (struct _slp_tree): Use a vector of slp_tree as children.
20617 (vect_get_place_in_interleaving_chain): Remove.
20618 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
20620 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
20622 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
20623 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
20624 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
20625 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
20626 Use slp_node instead of slp_void_p and adjust.
20628 2013-04-09 Richard Biener <rguenther@suse.de>
20630 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
20631 work that is not necessary.
20633 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20635 PR tree-optimization/56854
20636 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
20637 forward into clobber stmts if it would change MEM_REF lhs into
20640 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
20642 * tree.c (type_hash_lookup, type_hash_add): Make static.
20643 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
20645 2013-04-09 Richard Biener <rguenther@suse.de>
20647 * tree.h (unsave_expr_now): Remove.
20648 * tree-inline.c (mark_local_for_remap_r): Remove.
20649 (unsave_expr_1): Likewise.
20650 (unsave_r): Likewise.
20651 (unsave_expr_now): Likewise.
20652 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
20653 (propagate_tree_value): Likewise.
20655 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
20657 * doc/rtl.texi (sequence): Rewrite documentation to match the
20658 current use of SEQUENCE rtl objects.
20659 * rtl.def (SEQUENCE): Likewise.
20661 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
20662 Update documentation.
20663 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
20664 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
20666 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
20668 2013-04-08 Teresa Johnson <tejohnson@google.com>
20670 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
20671 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
20673 (estimate_edge_size_and_time): Add comment to suggest using rounding
20675 (estimate_node_size_and_time): Ditto.
20676 (remap_edge_change_prob): Use helper rounding divide methods.
20677 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
20678 (gimple_mod_pow2_value_transform): Ditto.
20679 (gimple_mod_subtract_transform): Ditto.
20680 (gimple_ic_transform): Ditto.
20681 (gimple_stringops_transform): Ditto.
20682 * stmt.c (conditional_probability): Ditto.
20683 (emit_case_dispatch_table): Ditto.
20684 * lto-cgraph.c (merge_profile_summaries): Ditto.
20685 * tree-optimize.c (execute_fixup_cfg): Ditto.
20686 * cfgcleanup.c (try_forward_edges): Ditto.
20687 * cfgloopmanip.c (scale_loop_profile): Ditto.
20689 (duplicate_loop_to_header_edge): Ditto.
20690 (lv_adjust_loop_entry_edge): Ditto.
20691 * tree-vect-loop.c (vect_transform_loop): Ditto.
20692 * profile.c (compute_branch_probabilities): Ditto.
20693 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
20694 * lto-streamer-in.c (input_cfg): Ditto.
20695 * gimple-streamer-in.c (input_bb): Ditto.
20696 * ipa-cp.c (update_profiling_info): Ditto.
20697 (update_specialized_profile): Ditto.
20698 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
20699 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
20701 * sched-rgn.c (compute_dom_prob_ps): Ditto.
20702 (compute_trg_info): Ditto.
20703 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
20704 (purge_dead_edges): Ditto.
20705 * loop-unswitch.c (unswitch_loop): Ditto.
20706 * cgraphclones.c (cgraph_clone_edge): Ditto.
20707 (cgraph_clone_node): Ditto.
20708 * tree-inline.c (copy_bb): Ditto.
20709 (copy_edges_for_bb): Ditto.
20710 (initialize_cfun): Ditto.
20711 (copy_cfg_body): Ditto.
20712 (expand_call_inline): Ditto.
20714 2013-04-08 Kai Tietz <ktietz@redhat.com>
20716 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
20717 TARGET_CYGWIN64 by TARGET_64BIT.
20719 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20721 * config/epiphany/epiphany.md (GPR_1): New constant.
20722 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
20723 * config/epiphany/epiphany.c (gen_compare_reg):
20724 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
20725 is already in place.
20726 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
20727 Don't require being called during rtl expansion; If y operlaps r0,
20729 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
20730 (epiphany_expand_epilogue): Likewise.
20732 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
20733 Don't use CC_FPmode for ORDERED / UNORDERED.
20734 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
20736 * config/epiphany/constraints.md (CnL): New constraint.
20737 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
20738 * config/epiphany/predicates.md (add_operand): Allow 1024.
20740 * config/epiphany/epiphany.md (logical_op): New code iterator.
20741 (op_mnc): New code attribute.
20742 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
20743 (mov_f+1, mov_f+2): New peephole2 patterns.
20745 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
20746 (cstoresi4): Also allow re-use of zero result when doing a NE
20747 comparison to a non-zero operand.
20748 Use (clobber (scratch)) for first insn if the gpr output is not needed.
20750 * config/epiphany/epiphany.md (<insn_opname>v2si3):
20751 Use gen_addsi3_i / gen_subsi3_i.
20753 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20757 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
20758 contain anything but clobbers, at most one __builtin_stack_restore,
20759 optionally debug stmts and final resx, and if it has at least one
20760 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
20761 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
20762 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
20763 which isn't defaut definition, remove them.
20764 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
20765 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
20766 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
20767 with MEM_REF LHS with SSA_NAME address.
20769 2013-04-08 Jeff Law <law@redhat.com>
20771 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
20773 2013-04-08 Richard Biener <rguenther@suse.de>
20775 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
20777 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
20778 determined vector type.
20779 (vect_analyze_data_refs): Likewise.
20780 (vect_get_new_vect_var): Adjust.
20781 (vect_create_destination_var): Preserve SSA name versions.
20782 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
20783 not dump anything here.
20785 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20787 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
20788 Add member lr_slot_known.
20789 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
20791 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
20792 Remove code that sets lr_slot_offset according to what a previous
20793 version of epiphany_emit_save_restore used to do.
20794 (epiphany_emit_save_restore): When doing an lr save or restore,
20795 set/verify lr_slot_known and lr_slot_offset.
20797 2013-04-08 Xinyu Qi <xyqi@marvell.com>
20800 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
20803 2013-04-08 Richard Biener <rguenther@suse.de>
20805 * alias.c (find_base_term): Fix thinko in previous change.
20807 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20809 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
20810 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
20811 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
20812 if possible to compute val.
20813 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
20814 For QImode integers don't require anything about precision. Use
20815 const_with_all_bytes_same to find out if the constant doesn't have
20816 repeated bytes in it.
20818 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20820 * config/s390/s390.c (s390_expand_insv): Only accept insertions
20823 2013-04-08 Marek Polacek <polacek@redhat.com>
20825 PR rtl-optimization/48182
20826 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
20829 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20832 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
20833 nuses, make sure we have a label.
20835 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20838 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
20839 (rs6000_emit_swdiv_low_precision): Remove.
20840 (rs6000_emit_swdiv): Rewrite to handle between one and four
20841 iterations of Newton-Raphson generally; modify required number of
20842 iterations for some cases.
20843 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
20845 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
20847 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
20848 set-but-unused variable.
20850 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
20851 basic blocks of released function bodies garbage-collectable.
20853 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
20854 (struct rtl_opt_pass): Add TODO_df_finish.
20856 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
20858 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20860 * config/arm/constraints.md (q): New constraint.
20861 * config/arm/ldrdstrd.md: New file.
20862 * config/arm/arm.md (ldrdstrd.md) New include.
20863 (arm_movdi): Use "q" instead of "r" constraint
20864 for double-word memory access.
20865 (movdf_soft_insn): Likewise.
20866 * config/arm/vfp.md (movdi_vfp): Likewise.
20867 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
20868 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
20869 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
20870 (mem_ok_for_ldrd_strd): Likewise.
20871 (output_move_double): Update assertion.
20873 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20875 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
20877 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20879 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
20880 define_insn_and_split.
20881 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
20883 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20885 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
20886 define_insn_and_split.
20887 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
20888 (shiftsi3_compare): New pattern.
20889 (rrx): New pattern.
20890 * config/arm/unspecs.md (UNSPEC_RRX): New.
20892 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20894 * config/arm/arm.md (negdi_extendsidi): New pattern.
20895 (negdi_zero_extendsidi): Likewise.
20897 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20899 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
20900 define_insn_and_split.
20901 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
20902 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
20904 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20906 * config/arm/arm.md (arm_subdi3): Convert define_insn into
20907 define_insn_and_split.
20908 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
20909 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
20911 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20913 * config/arm/arm.md (subsi3_carryin): New pattern.
20914 (subsi3_carryin_const): Likewise.
20915 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
20916 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
20918 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20920 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
20922 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20924 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
20925 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
20927 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20929 * config/arm/arm.c (arm_expand_builtin): Change fcode
20930 type to unsigned int.
20932 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20934 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
20936 2013-04-04 Ian Lance Taylor <iant@google.com>
20938 * doc/standards.texi (Standards): The Go frontend supports the Go 1
20941 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
20943 PR middle-end/56729
20944 * df-scan.c (df_insn_delete): Disable failing assert.
20946 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20948 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
20949 New function prototype.
20950 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
20951 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
20952 (arm_builtin_vectorized_function): New function.
20954 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20956 * config/arm/arm_neon_builtins.def: New file.
20957 * config/arm/arm.c (neon_builtin_data): Move contents to
20958 arm_neon_builtins.def.
20959 (enum arm_builtins): Include neon builtin definitions.
20960 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
20961 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
20963 2013-04-04 Marek Polacek <polacek@redhat.com>
20965 PR tree-optimization/48186
20966 * predict.c (maybe_hot_frequency_p): Return false if
20967 HOT_BB_FREQUENCY_FRACTION is 0.
20968 (cgraph_maybe_hot_edge_p): Likewise.
20970 2013-04-04 Richard Biener <rguenther@suse.de>
20972 PR tree-optimization/56826
20973 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
20976 2013-04-04 Richard Biener <rguenther@suse.de>
20978 PR tree-optimization/56213
20979 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
20980 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
20982 2013-04-04 Richard Biener <rguenther@suse.de>
20984 PR tree-optimization/56837
20985 * tree-loop-distribution.c (classify_partition): For non-zero
20986 values require that the value has the same precision as its
20987 mode to be useful as memset value.
20989 2013-04-03 Nick Clifton <nickc@redhat.com>
20991 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
20992 (fmssf4): Use fmsf.s on E3V5 architectures.
20993 (fnmasf4): Use fnmaf.s on E3V5 architectures.
20994 (fnmssf4): Use fnmsf.s on E3V5 architectures.
20996 2013-04-03 Jeff Law <law@redhat.com>
20998 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
20999 (lra-eliminations.o): Likewise.
21001 2013-04-03 Teresa Johnson <tejohnson@google.com>
21003 * gcov-io.c (compute_working_sets): Moved most of body of old
21004 compute_working_sets here from profile.c.
21005 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
21006 (gcov_working_set_t): Moved typedef here from basic-block.h
21007 (compute_working_set): Declare.
21008 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
21009 (get_working_sets): Renamed from compute_working_set,
21010 replace most of body with call to new compute_working_sets.
21011 (get_exec_counts): Replace call to compute_working_sets
21012 to get_working_sets.
21013 * profile.h (get_working_sets): Renamed from compute_working_set.
21014 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
21015 to get_working_sets.
21016 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
21017 * gcov-dump.c (dump_working_sets): New function.
21019 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
21021 * hwint.c (sext_hwi, zext_hwi): New functions.
21022 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
21023 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
21024 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
21025 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
21026 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
21027 (sext_hwi, zext_hwi): New functions.
21029 2013-04-03 Jeff Law <law@redhat.com>
21031 PR tree-optimization/56799
21032 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
21033 back test for widening conversion erroneously dropped in prior change.
21035 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21038 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
21039 instead of next_real_insn.
21041 2013-04-03 Marek Polacek <polacek@redhat.com>
21044 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
21046 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21049 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
21051 (thumb1_output_casesi): Likewise.
21052 (thumb2_output_casesi): Likewise.
21054 2013-04-03 Richard Biener <rguenther@suse.de>
21056 PR tree-optimization/56817
21057 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
21059 (tree_unroll_loops_completely_1): ... new function to manually
21060 walk the loop tree, properly defering outer loops of unrolled
21061 loops to later iterations.
21063 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21065 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
21066 (vectorizable_load): Likewise.
21067 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
21068 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
21070 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21072 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
21075 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
21077 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
21079 2013-04-03 Bin Cheng <bin.cheng@arm.com>
21081 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
21083 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21085 PR tree-optimization/56790
21086 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
21089 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21091 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
21093 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
21094 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
21097 2013-04-03 Jakub Jelinek <jakub@redhat.com>
21100 * tree.h (force_folding_builtin_constant_p): New decl.
21101 * builtins.c (force_folding_builtin_constant_p): New variable.
21102 (fold_builtin_constant_p): Fold immediately also if
21103 force_folding_builtin_constant_p.
21105 2013-04-03 Richard Biener <rguenther@suse.de>
21107 PR tree-optimization/56812
21108 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
21109 DRs of the same interleaving chain are independent.
21111 2013-04-02 Jason Merrill <jason@redhat.com>
21113 * gdbinit.in (pbb): Use debug fn.
21115 2013-04-02 Lawrence Crowl <crowl@google.com>
21117 * sese.h (struct ivtype_map_elt_s): Remove unused.
21118 (extern debug_ivtype_map): Remove unused.
21119 (extern eq_ivtype_map_elts): Remove unused.
21120 * sese.c (debug_ivtype_map): Removed unused.
21121 (debug_ivtype_map_1): Removed unused.
21122 (debug_ivtype_elt): Remove unused.
21123 (eq_ivtype_map_elts): Remove unused.
21126 2013-04-02 Kai Tietz <ktietz@redhat.com>
21129 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
21130 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
21131 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
21133 (legitimize_pe_coff_symbol): Likewise.
21134 (is_imported_p): New helper-function.
21135 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
21136 for Windows x64 targets.
21137 (ix86_expand_prologue): Optimize for pe-coff targets.
21138 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
21139 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
21140 medium/large code-model.
21141 (legitimize_pic_address): Likewise.
21142 (legitimize_tls_address): Likewise.
21143 (ix86_expand_call): Likewise.
21144 (x86_output_mi_thunk): Likewise.
21145 (get_dllimport_decl): Add new beimport argument.
21146 (construct_plt_address): Don't assert for x64 pe-coff targets.
21147 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
21149 (SYMBOL_FLAG_STUBVAR): New macro.
21150 (SYMBOL_REF_STUBVAR_P): Likewise.
21151 * config/i386/winnt.c (stub_list): New structure.
21152 (stub_head): New local variable.
21153 (i386_pe_record_stub): New function.
21154 (i386_pe_file_end): Emit refptr-stubs.
21156 2013-04-02 Jakub Jelinek <jakub@redhat.com>
21158 PR rtl-optimization/56745
21159 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
21160 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
21163 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
21164 and both of them are MEM_REFs, just compare first argument for
21165 equality and attempt to deal even with differing offsets.
21168 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
21169 of gimple_clobber_p to be MEM_REF.
21170 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
21171 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
21172 after gimplification.
21173 * asan.c (get_mem_ref_of_assignment): Don't instrument
21174 gimple_clobber_p stmts.
21175 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
21176 gimple_clobber_p stmt if they have MEM_REF lhs and
21177 are dead because of another gimple_clobber_p stmt.
21178 * tree-ssa-live.c (clear_unused_block_pointer): Treat
21179 gimple_clobber_p stmts like debug stmts.
21180 (remove_unused_locals): Remove clobbers with MEM_REF lhs
21181 that refer to unused VAR_DECLs or uninitialized values.
21182 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
21183 gimple_clobber_p stmts if they refer to removed parameters.
21184 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
21187 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
21189 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
21190 using SWI48 mode attribute.
21192 2013-04-02 Wei Mi <wmi@google.com>
21194 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
21195 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
21196 *<rotate_insn><mode>3_mask in i386.md.
21198 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
21200 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
21202 2013-04-02 Richard Biener <rguenther@suse.de>
21204 PR tree-optimization/56778
21205 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
21206 Runtime alias tests are not supported for gather loads.
21207 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
21208 stmts referenced from SSA operands before updating SSA form.
21210 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
21211 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21213 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
21214 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
21215 * config/arm/cortex-a53.md: New file.
21216 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
21217 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
21218 * config/arm/arm.c (arm_issue_rate): Likewise.
21219 * config/arm/arm-tune.md: Regenerate
21220 * config/arm/arm-tables.opt: Regenerate.
21221 * config/arm/arm-cores.def: Add cortex-a53.
21223 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
21225 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
21228 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21230 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
21231 scalar load/store operations using B/H registers.
21232 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
21234 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21236 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
21238 * config/aarch64/aarch64.c
21239 (aarch64_simd_scalar_immediate_valid_for_move): New.
21240 * config/aarch64/aarch64-protos.h
21241 (aarch64_simd_scalar_immediate_valid_for_move): New.
21242 * config/aarch64/constraints.md (Dh, Dq): New.
21243 * config/aarch64/iterators.md (hq): New.
21245 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
21247 * reorg.c (get_branch_condition): Deal with conditional returns.
21248 (fill_simple_delay_slots): Remove dead code dealing with jumps.
21250 2013-04-01 Wei Mi <wmi@google.com>
21252 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
21253 Truncate operand 2 using %b asm operand modifier.
21254 (*<shift_insn><mode>3_mask): Ditto.
21255 (*<rotate_insn><mode>3_mask): Ditto.
21257 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
21259 PR middle-end/56798
21260 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
21262 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
21264 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
21266 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
21268 2013-03-30 Lawrence Crowl <crowl@google.com>
21270 * dse.c (clear_alias_sets): Remove never set.
21271 (disqualified_clear_alias_sets): Remove never set.
21272 (clear_alias_mode_pool): Remove never set.
21273 (dse_step0): Remove condition that is never true.
21274 (canon_address): Remove condition that is never true.
21275 (dse_step7): Remove condition that is never true.
21276 (rest_of_handle_dse): Remove condition that is never true.
21277 (rest_of_handle_dse::did_global): Remove never read from above.
21278 (dse_step2_spill): Remove never called from above.
21279 (dse_step5_spill): Remove never called from above.
21281 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
21283 * doc/md.texi (Standard Names) <casesi>: Update documentation for
21284 JUMP_TABLE_DATA changes.
21285 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
21286 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
21287 (Insns) <jump_table_data>: New entry.
21288 * doc/tm.texi: Regenerate.
21290 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
21292 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
21293 for table jump at the end of a basic block using tablejump_p.
21294 * targhooks.c (default_invalid_within_doloop): Likewise.
21295 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
21296 target hook implementation that is identical to the default hook.
21297 (rs6000_invalid_within_doloop): Remove.
21299 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
21300 unused variable from tablejump_p call.
21302 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
21303 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
21304 (INSN_DELETED_P): Likewise.
21305 (emit_jump_table_data): New prototype.
21306 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
21307 after 4th as unused.
21308 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
21309 * sched-vis.c (print_insn): Likewise.
21310 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
21311 insn for compatibility with back ends that use next_active_insn to
21312 identify jump table data.
21313 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
21314 (remove_insn): Likewise.
21315 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
21317 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
21318 (emit_jump_table_data): New function.
21320 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
21321 basic block, a JUMP_TABLE_DATA never is.
21322 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
21323 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
21324 off from code handling real insns.
21325 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
21326 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
21327 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
21328 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
21329 is not a NONDEBUG_INSN_P.
21330 * ira-costs.c (scan_one_insn): Likewise.
21331 * jump.c (mark_all_labels): Likewise.
21332 (mark_jump_label_1): Likewise.
21333 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21334 * lra.c (get_insn_freq): Expect all insns reaching here to be in
21336 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
21337 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
21338 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
21339 JUMP_TABLE_DATA_P insns.
21340 (calculate_elim_costs_all_insns): Likewise.
21341 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
21342 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
21343 (delete_output_reload): Code style fixups.
21344 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
21345 insn flags on this non-insn.
21346 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
21347 as scheduling barriers, for pre-change compatibility.
21348 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
21349 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
21351 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
21352 redundant JUMP_TABLE_DATA_P test.
21353 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
21354 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21355 (frv_for_each_packet): Likewise.
21356 * config/i386/i386.c (min_insn_size): Likewise.
21357 (ix86_avoid_jump_mispredicts): Likewise.
21358 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21359 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
21360 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21361 (mips16_insn_length): Robustify.
21362 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
21363 (mips16_split_long_branches): Likewise.
21364 * config/pa/pa.c (pa_combine_instructions): Likewise.
21365 * config/rs6000/rs6000.c (get_next_active_insn): Treat
21366 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
21367 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
21368 as contributing to pool range lengths.
21369 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
21370 Remove redundant JUMP_TABLE_DATA_P test.
21371 (sh_loop_align): Likewise.
21372 (split_branches): Likewise.
21373 (sh_insn_length_adjustment): Likewise.
21374 * config/spu/spu.c (get_branch_target): Likewise.
21376 2013-03-29 Jan Hubicka <jh@suse.cz>
21378 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
21379 gcov streaming; stream hot bb threshold to ltrans.
21380 * predict.c (get_hot_bb_threshold): Break out from ....
21381 (maybe_hot_count_p): ... here.
21382 (set_hot_bb_threshold): New function.
21383 * lto-section-in.c (lto_section_name): Add profile.
21384 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
21385 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
21386 and data-streamer.h
21387 (histogram_entry): New structure.
21388 (histogram, histogram_pool): New global vars.
21389 (histogram_hash): New structure.
21390 (histogram_hash::hash): New method.
21391 (histogram_hash::equal): Likewise.
21392 (account_time_size): New function.
21393 (cmp_counts): New function.
21394 (dump_histogram): New function.
21395 (ipa_profile_generate_summary): New function.
21396 (ipa_profile_write_summary): New function.
21397 (ipa_profile_read_summary): New function.
21398 (ipa_profile): Decide on threshold.
21399 (pass_ipa_profile): Add ipa_profile_write_summary and
21400 ipa_profile_read_summary.
21401 * Makefile.in (ipa.o): Update dependencies.
21402 * lto-streamer.h (LTO_section_ipa_profile): New section.
21404 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
21406 * tree.h (VAR_P): New.
21408 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
21411 * doc/invoke.texi ([-fwhole-program]): Fix typo.
21413 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
21415 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
21416 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21417 (control_flow_insn_p): Likewise.
21418 * cfgrtl.c (duplicate_insn_chain): Likewise.
21419 * final.c (get_attr_length_1): Likewise.
21420 (shorten_branches): Likewise.
21421 (final_scan_insn): Likewise.
21422 * function.c (instantiate_virtual_regs): Likewise.
21423 * gcse.c (insert_insn_end_basic_block): Likewise.
21424 * ira-costs.c (scan_one_insn): Likewise.
21425 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21426 * lra.c (check_rtl): Likewise.
21427 * reload1.c (elimination_costs_in_insn): Likewise.
21428 * reorg.c (follow_jumps): Likewise.
21430 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
21431 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21432 (thumb_far_jump_used_p): Likewise.
21433 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
21434 (workaround_speculation): Likewise.
21435 (add_sched_insns_for_speculation): Likewise.
21436 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
21437 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21438 (frv_for_each_packet): Likewise.
21439 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
21440 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
21441 (final_emit_insn_group_barriers): Likewise.
21442 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21443 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21444 (mips16_insn_length): Likewise.
21445 * config/pa/pa.c (pa_reorg): Likewise.
21446 (pa_combine_instructions): Likewise.
21447 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
21448 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
21449 (sh_reorg): Likewise.
21450 (split_branches): Likewise.
21451 * config/spu/spu.c (get_branch_target): Likewise.
21453 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
21456 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
21458 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
21459 Fix declaration name.
21461 2013-03-28 Lawrence Crowl <crowl@google.com>
21463 * graphds.h (struct graph.indicies): Remove unused.
21464 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
21465 (SCOP_ORIGINAL_PDDRS): Remove unused.
21466 * sese.h (extern insert_loop_close_phis): Removed unused.
21467 (extern insert_guard_phis): Removed unused.
21468 (extern ivtype_map_elt_info): Removed unused.
21469 (new_ivtype_map_elt): Removed unused.
21470 * sese.c (ivtype_map_elt_info): Removed unused.
21472 2013-03-28 Lawrence Crowl <crowl@google.com>
21474 * Makefile.in: Add several missing include dependences.
21476 (test-dump.o): New. This object is not added to any executable,
21477 but is present for ad-hoc testing.
21479 (debug (const bitmap_head_def &)): New.
21480 (debug (const bitmap_head_def *)): New.
21482 (extern debug (const bitmap_head_def &)): New.
21483 (extern debug (const bitmap_head_def *)): New.
21485 (debug (edge_def &)): New.
21486 (debug (edge_def *)): New.
21488 (debug (basic_block_def &)): New.
21489 (debug (basic_block_def *)): New.
21491 (dump_node (const_tree, int, FILE *)): Correct source file.
21493 (debug (die_struct &)): New.
21494 (debug (die_struct *)): New.
21496 (extern debug (die_struct &)): New.
21497 (extern debug (die_struct *)): New.
21498 * gimple-pretty-print.c
21499 (debug (gimple_statement_d &)): New.
21500 (debug (gimple_statement_d *)): New.
21501 * gimple-pretty-print.h
21502 (extern debug (gimple_statement_d &)): New.
21503 (extern debug (gimple_statement_d *)): New.
21505 (debug (ira_allocno_copy &)): New.
21506 (debug (ira_allocno_copy *)): New.
21507 (debug (ira_allocno &)): New.
21508 (debug (ira_allocno *)): New.
21510 (extern debug (ira_allocno_copy &)): New.
21511 (extern debug (ira_allocno_copy *)): New.
21512 (extern debug (ira_allocno &)): New.
21513 (extern debug (ira_allocno *)): New.
21515 (debug (live_range &)): New.
21516 (debug (live_range *)): New.
21518 (debug (lra_live_range &)): New.
21519 (debug (lra_live_range *)): New.
21521 (debug (lra_live_range &)): New.
21522 (debug (lra_live_range *)): New.
21524 (debug (omega_pb_d &)): New.
21525 (debug (omega_pb_d *)): New.
21527 (extern debug (omega_pb_d &)): New.
21528 (extern debug (omega_pb_d *)): New.
21530 (debug (const rtx_def &)): New.
21531 (debug (const rtx_def *)): New.
21533 (debug_tree (tree): Move within file.
21534 (debug_raw (const tree_node &)): New.
21535 (debug_raw (const tree_node *)): New.
21536 (dump_tree_via_hooks (const tree_node *, int)): New.
21537 (debug (const tree_node &)): New.
21538 (debug (const tree_node *)): New.
21539 (debug_verbose (const tree_node &)): New.
21540 (debug_verbose (const tree_node *)): New.
21541 (debug_head (const tree_node &)): New.
21542 (debug_head (const tree_node *)): New.
21543 (debug_body (const tree_node &)): New.
21544 (debug_body (const tree_node *)): New.
21545 (debug_vec_tree (tree): Move and reimplement in terms of dump.
21546 (debug (vec<tree, va_gc> &)): New.
21547 (debug (vec<tree, va_gc> *)): New.
21549 (extern debug (const rtx_def &)): New.
21550 (extern debug (const rtx_def *)): New.
21552 (debug_raw (simple_bitmap_def &)): New.
21553 (debug_raw (simple_bitmap_def *)): New.
21554 (debug (simple_bitmap_def &)): New.
21555 (debug (simple_bitmap_def *)): New.
21557 (extern debug (simple_bitmap_def &)): New.
21558 (extern debug (simple_bitmap_def *)): New.
21559 (extern debug_raw (simple_bitmap_def &)): New.
21560 (extern debug_raw (simple_bitmap_def *)): New.
21562 (debug (vinsn_def &)): New.
21563 (debug (vinsn_def *)): New.
21564 (debug_verbose (vinsn_def &)): New.
21565 (debug_verbose (vinsn_def *)): New.
21566 (debug (expr_def &)): New.
21567 (debug (expr_def *)): New.
21568 (debug_verbose (expr_def &)): New.
21569 (debug_verbose (expr_def *)): New.
21570 (debug (vec<rtx> &)): New.
21571 (debug (vec<rtx> *)): New.
21573 (extern debug (vinsn_def &)): New.
21574 (extern debug (vinsn_def *)): New.
21575 (extern debug_verbose (vinsn_def &)): New.
21576 (extern debug_verbose (vinsn_def *)): New.
21577 (extern debug (expr_def &)): New.
21578 (extern debug (expr_def *)): New.
21579 (extern debug_verbose (expr_def &)): New.
21580 (extern debug_verbose (expr_def *)): New.
21581 (extern debug (vec<rtx> &)): New.
21582 (extern debug (vec<rtx> *)): New.
21584 (_list_iter_cond_expr): Make inline instead of static.
21586 (debug (sreal &)): New.
21587 (debug (sreal *)): New.
21589 (extern debug (sreal &)): New.
21590 (extern debug (sreal *)): New.
21592 (extern debug_raw (const tree_node &)): New.
21593 (extern debug_raw (const tree_node *)): New.
21594 (extern debug (const tree_node &)): New.
21595 (extern debug (const tree_node *)): New.
21596 (extern debug_verbose (const tree_node &)): New.
21597 (extern debug_verbose (const tree_node *)): New.
21598 (extern debug_head (const tree_node &)): New.
21599 (extern debug_head (const tree_node *)): New.
21600 (extern debug_body (const tree_node &)): New.
21601 (extern debug_body (const tree_node *)): New.
21602 (extern debug (vec<tree, va_gc> &)): New.
21603 (extern debug (vec<tree, va_gc> *)): New.
21605 (debug (struct loop &)): New.
21606 (debug (struct loop *)): New.
21607 (debug_verbose (struct loop &)): New.
21608 (debug_verbose (struct loop *)): New.
21609 * tree-dump.c: Add header dependence.
21611 (extern debug (struct loop &)): New.
21612 (extern debug (struct loop *)): New.
21613 (extern debug_verbose (struct loop &)): New.
21614 (extern debug_verbose (struct loop *)): New.
21616 (debug (data_reference &)): New.
21617 (debug (data_reference *)): New.
21618 (debug (vec<data_reference_p> &)): New.
21619 (debug (vec<data_reference_p> *)): New.
21620 (debug (vec<ddr_p> &)): New.
21621 (debug (vec<ddr_p> *)): New.
21623 (extern debug (data_reference &)): New.
21624 (extern debug (data_reference *)): New.
21625 (extern debug (vec<data_reference_p> &)): New.
21626 (extern debug (vec<data_reference_p> *)): New.
21627 (extern debug (vec<ddr_p> &)): New.
21628 (extern debug (vec<ddr_p> *)): New.
21630 (debug (pt_solution &)): New.
21631 (debug (pt_solution *)): New.
21633 (extern debug (pt_solution &)): New.
21634 (extern debug (pt_solution *)): New.
21636 (debug (_var_map &)): New.
21637 (debug (_var_map *)): New.
21638 (debug (tree_live_info_d &)): New.
21639 (debug (tree_live_info_d *)): New.
21641 (extern debug (_var_map &)): New.
21642 (extern debug (_var_map *)): New.
21643 (extern debug (tree_live_info_d &)): New.
21644 (extern debug (tree_live_info_d *)): New.
21646 2013-03-28 Jan Hubicka <jh@suse.cz>
21648 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
21650 2013-03-28 Ian Bolton <ian.bolton@arm.com>
21652 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
21653 record only when desired or required.
21655 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
21657 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
21658 *vec_extractv2di_1_rex64. Use x64 isa attribute.
21660 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21662 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
21663 (*andsi3_compare0_uxtw): New pattern.
21664 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
21665 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
21667 2013-03-28 Jan Hubicka <jh@suse.cz>
21669 * data-streamer-in.c (streamer_read_gcov_count): New function.
21670 * gimple-streamer-out.c: Include value-prof.h.
21671 (output_gimple_stmt): Output histogram.
21672 (output_bb): Use streamer_write_gcov_count.
21673 * value-prof.c: Include data-streamer.h
21674 (dump_histogram_value): Add HIST_TYPE_MAX.
21675 (stream_out_histogram_value): New function.
21676 (stream_in_histogram_value): New function.
21677 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
21678 (stream_out_histogram_value, stream_in_histogram_value): Declare.
21679 * data-streamer-out.c (streamer_write_gcov_count): New function.
21680 (streamer_write_gcov_count_stream): New function.
21681 * lto-cgraph.c (lto_output_edge): Update counter streaming.
21682 (lto_output_node): Likewise.
21683 (input_node, input_edge): Likewise.
21684 * lto-streamer-out.c (output_cfg): Update streaming.
21685 * lto-streamer-in.c (input_cfg): Likewise.
21686 * data-streamer.h (streamer_write_gcov_count,
21687 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
21688 * gimple-streamer-in.c: Include value-prof.h
21689 (input_gimple_stmt): Input histograms.
21690 (input_bb): Update profile streaming.
21692 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21694 * genmodes.c (emit_max_int): New function.
21695 (emit_insn_modes_h): Added call to emit_max_function.
21696 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
21698 * machmode.def: Fixed comment.
21700 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21702 * combine.c (try_combine): Removed useless assert.
21703 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
21705 2013-03-28 Marek Polacek <polacek@redhat.com>
21706 Richard Biener <rguenther@suse.de>
21708 PR tree-optimization/56695
21709 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
21710 build signed result of a vector comparison.
21711 * tree-cfg.c (verify_gimple_comparison): Check that a result
21712 of a vector comparison has signed type.
21714 2013-03-28 Richard Biener <rguenther@suse.de>
21716 PR tree-optimization/37021
21717 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
21718 do not restrict gaps between groups.
21719 * tree-vect-stmts.c (vectorizable_load): Properly account for
21720 a gap between groups.
21722 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
21724 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
21725 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
21728 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
21730 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
21731 * doc/extend.texi (Named Address Spaces): Ditto.
21732 (Variable Attributes): Ditto.
21734 2013-03-27 Kai Tietz <ktietz@redhat.com>
21736 * config.build: Add support for cygwin x64 target.
21737 * config.gcc: Likewise.
21738 * config.host: Likewise.
21739 * configure.ac: Likewise
21740 * configure: Regenerated.
21742 2013-03-27 Kai Tietz <ktietz@redhat.com>
21744 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
21745 * config/i386/t-cygwin-w64: New file.
21746 * config/i386/cygwin-w64.h: New file.
21747 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
21748 and add support for x64-cygwin target.
21749 (CPP_SPEC): Likewise.
21750 (CXX_WRAP_SPEC_LIST): Undefine before define.
21751 (LIBGCJ_SONAME): Use 15 as version.
21753 2013-03-27 Richard Biener <rguenther@suse.de>
21755 PR tree-optimization/56716
21756 * tree-ssa-structalias.c (perform_var_substitution): Adjust
21757 dumping for ref nodes.
21759 2013-03-27 Martin Jambor <mjambor@suse.cz>
21761 PR tree-optimization/55334
21762 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
21763 restricted pointers to arrays.
21765 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
21767 * Makefile.in (.SUFFIXES): Add .cc.
21768 (.c.o): Apply same recipe for implicit rule .cc.o.
21770 2013-03-27 Richard Biener <rguenther@suse.de>
21772 PR tree-optimization/37021
21773 * tree-vect-data-refs.c (vect_check_strided_load): Allow
21774 REALPART/IMAGPART_EXPRs around the supported refs.
21775 * tree-ssa-structalias.c (find_func_aliases): Assume that
21776 floating-point values are not used to transfer pointers.
21778 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
21780 * target.def (TARGET_HAS_IFUNC_P): New target hook.
21781 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
21782 * doc/tm.texi: Regenerate.
21783 * targhooks.h (default_has_ifunc_p): New.
21784 * targhooks.c (default_has_ifunc_p): Ditto.
21785 * config/linux-protos.h: New file.
21786 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
21787 hook for linux which disables support of indirect functions in android.
21788 * config/linux-android.c: New file.
21789 * config/t-linux-android.c: Ditto.
21790 * config.gcc: Added new object file linux-android.o.
21791 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
21792 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
21793 * varasm.c (do_assemble_alias): Likewise.
21794 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
21795 doesn't support indirect functions.
21796 * configure: Regenerate.
21798 2013-03-27 Bin Cheng <bin.cheng@arm.com>
21801 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
21802 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
21805 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21807 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
21809 2013-03-27 Terry Guo <terry.guo@arm.com>
21811 * config/arm/arm-cores.def: Added core cortex-r7.
21812 * config/arm/arm-tune.md: Regenerated.
21813 * config/arm/arm-tables.opt: Regenerated.
21814 * doc/invoke.texi: Added entry for core cortex-r7.
21816 2013-03-27 Walter Lee <walt@tilera.com>
21818 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
21819 double-decrement of next_scratch_regno.
21821 2013-03-27 Walter Lee <walt@tilera.com>
21823 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
21825 (insn_v1mulus): Ditto.
21826 (insn_v2muls): Ditto.
21828 2013-03-27 Walter Lee <walt@tilera.com>
21830 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
21831 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
21833 2013-03-27 Walter Lee <walt@tilera.com>
21835 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
21836 (*sibcall_value): Ditto.
21838 2013-03-27 Walter Lee <walt@tilera.com>
21840 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
21841 (insn_mnz_v8qi): ... this ...
21842 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
21844 (insn_v<n>mnz): Replaced by ...
21845 (insn_v1mnz): ... this ...
21846 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
21848 (insn_mz_<mode>): Replaced by ...
21849 (insn_mz_v8qi): ... this ...
21850 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
21852 (insn_v<n>mz): Replaced by ...
21853 (insn_v1mz): ... this ...
21854 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
21857 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
21859 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
21861 2013-03-26 Roland McGrath <mcgrathr@google.com>
21863 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
21864 than fprintf with a non-constant, non-format string.
21866 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21868 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
21869 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
21870 operand 0 predicate.
21871 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
21872 attribute. Use general_x64nomem_operand as operand 1 predicate.
21873 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
21874 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21875 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
21876 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21877 (mov<mode>_insv_1): Remove expander. Merge insn with
21878 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
21879 Use general_x64nomem_operand as operand 1 predicate.
21880 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
21881 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
21882 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21883 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
21884 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
21885 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
21886 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21887 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
21888 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
21889 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
21890 (general_x64nomem_operand): Ditto.
21892 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
21894 * config/rtems.opt: Add -pthread option.
21896 2013-03-26 Richard Biener <rguenther@suse.de>
21898 * alias.c (find_base_term): Avoid redundant and not used recursion.
21899 (base_alias_check): Get the initial base term from the caller.
21900 (true_dependence_1): Compute and pass base terms to base_alias_check.
21901 (write_dependence_p): Likewise.
21902 (may_alias_p): Likewise.
21904 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
21906 * config/aarch64/aarch64.c (aarch64_classify_address): Support
21907 PC-relative load in SI modes and above only.
21909 2013-03-26 Xinyu Qi <xyqi@marvell.com>
21911 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
21912 * config/arm/iwmmxt.md (WCGR0): Update.
21913 (WCGR1, WCGR2, WCGR3): Likewise.
21915 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21917 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
21918 Use x64 and nox64 isa attributes.
21920 2013-03-26 Richard Biener <rguenther@suse.de>
21922 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
21923 alignment computations and rely on get_object_alignment_1
21924 for the !TYPE_P case.
21925 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
21927 2013-03-26 Walter Lee <walt@tilera.com>
21929 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
21930 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
21932 2013-03-25 Jeff Law <law@redhat.com>
21934 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
21935 check for INTEGRAL_TYPE_P that was missing due to checking in
21936 wrong version of prior patch.
21938 2013-03-25 Walter Lee <walt@tilera.com>
21940 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
21941 TILEGX_INSN_SHUFFLEBYTES1.
21942 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
21944 (tilegx_builtins): Ditto.
21945 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
21947 2013-03-25 Walter Lee <walt@tilera.com>
21949 * config/tilegx/tilegx.md (floatsisf2): New pattern.
21950 (floatunssisf2): New pattern.
21951 (floatsidf2): New pattern.
21952 (floatunssidf2): New pattern.
21954 2013-03-25 Walter Lee <walt@tilera.com>
21956 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
21957 tests for constraint J, K, N, P.
21959 2013-03-25 Walter Lee <walt@tilera.com>
21961 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
21962 Use indirect/pcrel encoding.
21963 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
21966 2013-03-25 Steve Ellcey <sellcey@mips.com>
21968 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
21969 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
21970 * config/mips/mips.c (mips_option_override): Set IMADD default.
21971 * config/mips/mips.h (PTF_AVOID_IMADD): New.
21972 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
21973 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
21974 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
21975 * doc/invoke.texi (-mimadd/-mno-imadd): New.
21977 2013-03-25 Jeff Law <law@redhat.com>
21979 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
21980 slightly to avoid creating and folding useless trees. Simplify
21981 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
21983 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
21985 * config/i386/i386.md (*zero_extendsidi2): Merge with
21986 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
21987 * config/i386/predicates.md (x86_64_zext_operand): Rename from
21988 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
21989 targets. Clarify comment.
21991 2013-03-25 Martin Jambor <mjambor@suse.cz>
21993 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
21994 pass-through jump functions differently.
21995 (ipa_read_jump_function): Likewise. Also use setter functions to set
21998 2013-03-25 Martin Jambor <mjambor@suse.cz>
22000 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
22001 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
22003 (ipa_get_indirect_edge_target): New function.
22004 (devirtualization_time_bonus): New parameter known_aggs, pass it to
22005 ipa_get_indirect_edge_target. Update all callers.
22006 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
22007 ipa_get_indirect_edge_target_1 instead of calling
22008 ipa_get_indirect_edge_target.
22009 (create_specialized_node): Pass aggvlas to
22010 ipcp_discover_new_direct_edges.
22012 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22014 * config/arm/arm.md (f_sels, f_seld): New types.
22015 (*cmov<mode>): New pattern.
22016 * config/arm/predicates.md (arm_vsel_comparison_operator): New
22019 2013-03-25 Kai Tietz <ktietz@redhat.com>
22021 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
22022 POSIX-printf for mingw-hosted builds.
22024 2013-03-25 Richard Biener <rguenther@suse.de>
22026 PR middle-end/56694
22027 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
22028 must-not-throw stmt location.
22030 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22032 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
22033 Emit load-acquire versions when acq is true.
22034 (arm_emit_store_exclusive): Add rel parameter.
22035 Emit store-release versions when rel is true.
22036 (arm_split_compare_and_swap): Use acquire-release instructions
22038 of barriers when appropriate.
22039 (arm_split_atomic_op): Likewise.
22040 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
22041 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
22042 (VUNSPEC_SLX): Likewise.
22043 (VUNSPEC_LDA): Likewise.
22044 (VUNSPEC_STL): Likewise.
22045 * config/arm/sync.md (atomic_load<mode>): New pattern.
22046 (atomic_store<mode>): Likewise.
22047 (arm_load_acquire_exclusive<mode>): Likewise.
22048 (arm_load_acquire_exclusivesi): Likewise.
22049 (arm_load_acquire_exclusivedi): Likewise.
22050 (arm_store_release_exclusive<mode>): Likewise.
22052 2013-03-25 Catherine Moore <clm@codesourcery.com>
22054 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
22055 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
22056 * config/mip/predicates.md (lwsp_swsp_operand,
22057 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
22058 sb16_operand, db4_operand, db7_operand, ib3_operand,
22059 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
22060 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
22061 andi16_operand): New predicates.
22062 * config/mips/mips.md (compression): New attribute.
22063 (enabled): New attribute.
22064 (length): Consider compression in computing length.
22065 (shift_compression): New code attribute.
22066 (*add<mode>3): New operands. Record compression.
22067 (sub<mode>3): Likewise.
22068 (one_cmpl<mode>2): Likewise.
22069 (*and<mode>3): Likewise.
22070 (*ior<mode>3): Likewise.
22071 (unnamed pattern for xor): Likewise.
22072 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
22073 (*<optab><mode>3): Likewise.
22074 (*mov<mode>_internal: Likewise.
22075 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
22076 (mips_unsigned_immediate_p): New.
22077 (umips_lwsp_swsp_address_p): New.
22078 (m16_based_address_p): New.
22079 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
22080 (mips_unsigned_immediate_p): New prototype.
22081 (lwsp_swsp_address_p): New prototype.
22082 (m16_based_address_p): New prototype.
22083 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
22084 (mips_signed_immediate_p): New function.
22085 (m16_based_address_p): New function.
22086 (lwsp_swsp_address_p): New function.
22087 (mips_print_operand_punctuation): Recognize short delay slot insns
22088 for microMIPS.add<mode>3"
22090 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22093 * config/arm/iterators.md (v_cmp_result): New mode attribute.
22094 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
22096 2013-03-25 Richard Biener <rguenther@suse.de>
22098 PR tree-optimization/56689
22099 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
22102 2013-03-25 Richard Biener <rguenther@suse.de>
22104 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
22106 (memory_references): Likewise.
22107 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
22108 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
22109 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
22110 (gather_mem_refs_in_loops): Fold into ...
22111 (analyze_memory_references): ... this. Move initialization
22112 to tree_ssa_lim_initialize.
22113 (fill_always_executed_in): Rename to ...
22114 (fill_always_executed_in_1): ... this.
22115 (fill_always_executed_in): Move contains_call computation to
22116 this new function from ...
22117 (tree_ssa_lim_initialize): ... here.
22118 (tree_ssa_lim): Call fill_always_executed_in.
22120 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
22122 * postreload.c (reload_combine): Fix code detecting returns.
22124 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
22126 * function.c (emit_use_return_register_into_block): On cc0 targets,
22127 do not emit the sequence between cc0 setter and user.
22129 2013-03-25 Kai Tietz <ktietz@redhat.com>
22131 * config/i386/predicates.md (local_symbolic_operand): Interpret
22132 dll-imported symbols as none-local.
22134 2013-03-25 Richard Biener <rguenther@suse.de>
22136 * tree-ssa-loop-im.c (struct depend): Remove.
22137 (struct lim_aux_data): Make depends a vec of gimples.
22138 (free_lim_aux_data): Adjust.
22139 (add_dependency): Likewise.
22140 (set_level): Likewise.
22142 2013-03-25 Richard Biener <rguenther@suse.de>
22144 PR middle-end/56434
22145 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
22146 the pointer returned by calls with ECF_MALLOC set.
22148 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
22150 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
22152 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
22154 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
22155 using MMXMODE mode iterator.
22156 (*move<mode>_internal): Merge with *movv2sf_internal and
22157 *movv2sf_internal_rex64 using MMXMODE mode iterator.
22159 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22161 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
22162 (record_last_mem_set_info): Likewise.
22164 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
22165 of XNEWVEC followed by memset.
22166 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
22168 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22170 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
22171 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
22172 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
22173 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
22174 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
22175 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
22176 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
22177 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
22178 BARRIER_P instead of GET_CODE.
22180 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22182 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
22183 inaccuracy in the probing code.
22185 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
22186 (ctrapdi4): Likewise.
22188 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22190 * calls.c (expand_call): Add missing guard to code handling return
22191 of non-BLKmode structures in MSB.
22192 * function.c (expand_function_end): Likewise.
22194 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22196 * combine.c (try_combine): Adjust comment. Do not add the set of
22197 insn #0 if the destination indirectly is set or dies in insn #2.
22198 Tidy up code to distribute a new note.
22200 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22202 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
22203 also for alternatives 16 and 17.
22205 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22207 * config/i386/sse.md (*mov<mode>_internal): Merge with
22208 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
22209 Emit insn template depending on type attribute. Use
22210 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22211 movd instead of movq mnemonic for interunit moves. Rewrite mode
22212 attribute calculation. Remove unit attribute calculation.
22213 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
22214 Set prefix_data16 attribute for DImode ssemov types.
22215 Use Ym instead of y for SSE-MMX conversion alternatives.
22216 Reorder operand constraints.
22218 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
22220 * df.h (df_insn_delete): Adjust prototype.
22221 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
22222 and let it decide whether mark the basic block dirty.
22223 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
22224 * df-scan.c (df_insn_info_delete): New helper function, split
22225 off from df_insn_delete.
22226 (df_scan_free_bb_info): Use it.
22227 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
22229 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
22230 that the insn is actually an insn and it has a non-NULL basic block.
22231 Do not mark basic block dirty if only deleting a DEBUG_INSN.
22233 2013-03-22 Richard Biener <rguenther@suse.de>
22235 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
22237 (mem_ref_alloc): Do not allocate them.
22238 (refs_independent_p): Do not query or maintain a cache.
22240 2013-03-22 Richard Biener <rguenther@suse.de>
22242 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
22243 (gather_mem_refs_in_loops): Do not compute it.
22244 (analyze_memory_references): Do not allocate it.
22245 (tree_ssa_lim_finalize): Do not free it.
22246 (for_all_locs_in_loop): Do not query all_refs_in_loop.
22248 2013-03-22 Richard Biener <rguenther@suse.de>
22250 * is-a.h (as_a): Use gcc_checking_assert.
22252 2013-03-22 Ian Bolton <ian.bolton@arm.com>
22254 * config/aarch64/aarch64.c (aarch64_print_operand): New
22255 format specifier for printing a constant in hex.
22256 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
22257 format specifier for printing second operand.
22259 2013-03-22 Richard Biener <rguenther@suse.de>
22261 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
22263 (gather_mem_refs_in_loops): Perform store accumulation here.
22264 (create_vop_ref_mapping_loop): Remove.
22265 (create_vop_ref_mapping): Likewise.
22266 (analyze_memory_references): Initialize refs_stored_in_loop.
22267 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
22268 (record_indep_loop): Remove.
22269 (record_dep_loop): New function.
22270 (ref_indep_loop_p_1): Adjust to only walk over references
22271 in the loop, not its subloops.
22272 (ref_indep_loop_p): Rename to ...
22273 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
22274 maintaining a more fine-grained cache.
22275 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
22276 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
22278 2013-03-22 Richard Biener <rguenther@suse.de>
22280 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
22281 (struct mem_ref): Make accesses_in_loop a vec of a vec of
22282 aggregate mem_ref_loc.
22283 (free_mem_ref_locs): Inline into ...
22284 (memref_free): ... this and adjust.
22285 (mem_ref_alloc): Adjust.
22286 (mem_ref_locs_alloc): Remove.
22287 (record_mem_ref_loc): Adjust.
22288 (get_all_locs_in_loop): Rewrite into ...
22289 (for_all_locs_in_loop): ... this iterator.
22290 (rewrite_mem_ref_loc): New functor.
22291 (rewrite_mem_refs): Use for_all_locs_in_loop.
22292 (sm_set_flag_if_changed): New functor.
22293 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
22294 (ref_always_accessed): New functor.
22295 (ref_always_accessed_p): Use for_all_locs_in_loop.
22297 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22299 * tree-pass.h (PROP_gimple_lvec): New.
22300 * passes.c (dump_properties): Handle PROP_gimple_lvec.
22301 (init_optimization_passes): Move pass_lower_vector.
22302 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
22304 (pass_lower_vector): Provide PROP_gimple_lvec.
22305 (pass_lower_vector_ssa): Likewise.
22306 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
22308 2013-03-21 Mark Wielaard <mjw@redhat.com>
22310 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
22312 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22314 * config/i386/i386.md (*movdi_internal): Disparage slightly
22315 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
22316 conversion alternatives.
22318 2013-03-21 Jakub Jelinek <jakub@redhat.com>
22320 PR middle-end/48087
22321 * diagnostic.def (DK_WERROR): New kind.
22322 * diagnostic.h (werrorcount): Define.
22323 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
22324 promoted to DK_ERROR, increment DK_WERROR counter instead of
22326 * toplev.c (toplev_main): Call print_ignored_options even if
22327 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
22328 even if just werrorcount is non-zero.
22331 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
22333 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
22334 (string_cst_pool_decl): New function.
22335 (optimize_one_addr_into_implicit_ptr): New function.
22336 (resolve_addr_in_expr): Optimize DWARF location expression
22337 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
22338 which doesn't live in memory, but has DW_AT_location or
22339 DW_AT_const_value, or refers to a string literal, into
22340 DW_OP_GNU_implicit_pointer.
22341 (optimize_location_into_implicit_ptr): New function.
22342 (resolve_addr): If removing DW_AT_location of a variable because
22343 it was DW_OP_addr of address of the variable, but the variable doesn't
22344 live in memory, try to emit const value attribute for the initializer.
22346 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22348 * tree.h (VECTOR_TYPE_P): New macro.
22349 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
22350 TYPE_MODE): Use it.
22351 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
22352 VEC_COND_EXPR cannot be lvalues.
22353 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
22355 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22357 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
22358 Restrict the transformation to equal modes.
22360 2013-03-21 Richard Biener <rguenther@suse.de>
22362 PR tree-optimization/39326
22363 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
22364 (MEM_ANALYZABLE): Adjust.
22365 (record_mem_ref_loc): Move bitmap ops ...
22366 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
22367 unanalyzable refs, do not record locations for it.
22368 (analyze_memory_references): Allocate ref zero as shared
22370 (refs_independent_p): Do not test for unanalyzed mems here.
22371 (ref_indep_loop_p_1): Special-case disambiguation against
22372 the unanalyzed ref.
22373 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
22375 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
22377 * config/arm/arm-protos.h (tune_params): Add
22378 prefer_neon_for_64bits field.
22379 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
22380 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
22381 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
22382 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
22383 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
22384 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
22385 (arm_option_override): Handle -mneon-for-64bits new option.
22386 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
22387 (prefer_neon_for_64bits): Declare new variable.
22388 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
22389 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
22390 (arch_enabled): Handle new arch types. Remove support for onlya8
22392 (one_cmpldi2): Use new arch names.
22393 (zero_extend<mode>di2, extend<mode>di2): Ditto.
22394 * config/arm/arm.opt (mneon-for-64bits): Add option.
22395 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
22396 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
22397 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
22399 * doc/invoke.texi (-mneon-for-64bits): Document.
22401 2013-03-21 Richard Biener <rguenther@suse.de>
22403 PR tree-optimization/39326
22404 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
22405 (sort_bbs_in_loop_postorder_cmp): New function.
22406 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
22408 2013-03-21 Richard Biener <rguenther@suse.de>
22410 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
22411 (vect_insert_into_interleaving_chain): Likewise.
22412 (vect_drs_dependent_in_basic_block): Inline ...
22413 (vect_slp_analyze_data_ref_dependence): ... here. New function,
22415 (vect_analyze_data_ref_dependence): ... here. Simplify.
22416 (vect_check_interleaving): Simplify.
22417 (vect_analyze_data_ref_dependences): Likewise. Split out ...
22418 (vect_slp_analyze_data_ref_dependences): ... this new function.
22419 (dr_group_sort_cmp): New function.
22420 (vect_analyze_data_ref_accesses): Compute data-reference groups
22421 here instead of in vect_analyze_data_ref_dependence. Use
22422 a more efficient algorithm.
22423 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
22424 vect_slp_analyze_data_ref_dependences. Call
22425 vect_analyze_data_ref_accesses earlier.
22426 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
22427 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
22428 (vect_slp_analyze_data_ref_dependences): New prototype.
22430 2013-03-21 Richard Biener <rguenther@suse.de>
22432 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
22433 ref is stored in the loop.
22434 (find_refs_for_sm): Walk only over all stores.
22435 (store_motion_loop): Allocate from lim_bitmap_obstack.
22436 (store_motion): Likewise.
22438 2013-03-21 Richard Biener <rguenther@suse.de>
22440 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
22441 Update virtual SSA form.
22443 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22445 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
22446 * configure: Regenerate.
22447 * config.in: Regenerate.
22448 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
22449 if !HAVE_LD_EH_FRAME_CIEV3.
22451 2013-03-21 Richard Biener <rguenther@suse.de>
22453 * tree-cfg.c (verify_expr_no_block): New function.
22454 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
22455 nor DECL_VALUE_EXPR have locations with associated blocks.
22456 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
22457 (clear_unused_block_pointer): Remove code dealing with
22458 blocks in DECL_DEBUG_EXPR locations.
22460 2013-03-21 Richard Biener <rguenther@suse.de>
22462 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
22463 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
22464 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
22465 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
22466 instead of DECL_DEBUG_EXPR_IS_FROM.
22467 * gimplify.c (gimplify_modify_expr): Likewise.
22468 * tree-cfg.c (verify_expr_location_1): Likewise.
22469 * tree-complex.c (create_one_component_var): Likewise.
22470 * tree-sra.c (create_access_replacement): Likewise.
22471 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
22472 (clear_unused_block_pointer): Likewise.
22473 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
22474 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
22475 * var-tracking.c (var_debug_decl): Likewise.
22476 (track_expr_p): Likewise.
22477 * tree-inline.c (add_local_variables): Likewise. Set
22478 DECL_HAS_DEBUG_EXPR_P after copying it.
22479 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
22480 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
22482 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22485 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
22486 * configure: Regenerate.
22487 * config.in: Regenerate.
22488 * config/i386/i386.md (*movdf_internal): Use
22489 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22490 movd instead of movq mnemonic for interunit moves.
22491 (*movdi_internal): Ditto.
22493 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22495 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
22496 (abd<mode>_3): New pattern.
22497 (aba<mode>_3): New pattern.
22498 (fabd<mode>_3): New pattern.
22500 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22502 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
22503 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
22504 occurrence of REGISTER_PREFIX as its empty string.
22506 2013-03-20 Jeff Law <law@redhat.com>
22508 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
22509 addititional equivalences for equality comparisons between an SSA_NAME
22510 and a constant where the SSA_NAME was set from a widening conversion.
22512 2013-03-20 Walter Lee <walt@tilera.com>
22514 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
22516 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22518 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
22519 depending on type attribute.
22520 (*movti_internal): Ditto.
22521 (*movtf_internal): Ditto.
22522 (*movxf_internal): Ditto.
22523 (*movdf_internal): Ditto.
22524 (*movsf_internal): Ditto.
22526 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22528 * config/i386/i386.md (*movti_internal): Set prefix attribute to
22529 maybe_vex for sselog1 and ssemov types.
22530 (*movdi_internal): Reorder operand constraints.
22531 (*movsi_internal): Ditto. Set prefix attribute to
22532 maybe_vex for sselog1 and ssemov types.
22533 (*movtf_internal): Set prefix attribute to maybe_vex
22534 for sselog1 and ssemov types.
22535 (*movdf_internal): Ditto. Set prefix_data16 attribute for
22536 DImode ssemov types. Reorder operand constraints.
22537 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
22538 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
22539 attribute for SImode ssemov types. Reorder operand constraints.
22541 2013-03-20 Martin Jambor <mjambor@suse.cz>
22543 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
22544 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
22546 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
22548 * config/rs6000/predicates.md (indexed_address, update_address_mem
22549 update_indexed_address_mem): New predicates.
22550 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
22551 attribute for load/store instructions.
22552 * config/rs6000/dfp.md (movsd_store): Likewise.
22553 (movsd_load): Likewise.
22554 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
22555 (unnamed HI->DI extend define_insn): Likewise.
22556 (unnamed SI->DI extend define_insn): Likewise.
22557 (unnamed QI->SI extend define_insn): Likewise.
22558 (unnamed QI->HI extend define_insn): Likewise.
22559 (unnamed HI->SI extend define_insn): Likewise.
22560 (unnamed HI->SI extend define_insn): Likewise.
22561 (extendsfdf2_fpr): Likewise.
22562 (movsi_internal1): Likewise.
22563 (movsi_internal1_single): Likewise.
22564 (movhi_internal): Likewise.
22565 (movqi_internal): Likewise.
22566 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
22567 attribute for load/store instructions.
22568 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
22570 (mov<mode>_softfloat): Likewise.
22571 (mov<mode>_hardfloat32): Likewise.
22572 (mov<mode>_hardfloat64): Likewise.
22573 (mov<mode>_softfloat64): Likewise.
22574 (movdi_internal32): Likewise.
22575 (movdi_internal64): Likewise.
22576 (probe_stack_<mode>): Likewise.
22578 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
22580 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
22581 floating point, and decimal floating point to reload iterator.
22583 * config/rs6000/constraints.md (wl constraint): New constraints to
22584 return FLOAT_REGS if certain options are used to reduce the number
22585 of separate patterns that exist in the file.
22586 (wx constraint): Likewise.
22587 (wz constraint): Likewise.
22589 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22590 -mdebug=reg, print wg, wl, wx, and wz constraints.
22591 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22592 Initialize the reload functions for 64-bit binary/decimal floating
22594 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22595 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22596 create the buffer on the stack to overcome not having a 32-bit
22598 (rs6000_emit_move): Likewise.
22599 (rs6000_secondary_memory_needed_rtx): Likewise.
22600 (rs6000_alloc_sdmode_stack_slot): Likewise.
22601 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22602 via xxlxor, just like DFmode 0.0.
22604 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
22605 define as 1 if we are running on a power7 or newer.
22606 (enum r6000_reg_class_enum): Add new constraints.
22608 * config/rs6000/dfp.md (movsd): Delete, combine with binary
22609 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22610 with other moves by using conditional constraits (wg). Use LFIWZX
22611 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
22612 (movsd splitter): Likewise.
22613 (movsd_hardfloat): Likewise.
22614 (movsd_softfloat): Likewise.
22616 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
22617 binary and decimal floating point moves.
22618 (fmove_ok): New attributes to combine binary and decimal floating
22619 point moves, and to combine power6x (mfpgpr) moves along normal
22621 (real_value_to_target): Likewise.
22622 (f32_lr): Likewise.
22623 (f32_lm): Likewise.
22624 (f32_li): Likewise.
22625 (f32_sr): Likewise.
22626 (f32_sm): Likewise.
22627 (f32_si): Likewise.
22628 (movsf): Combine binary and decimal floating point moves. Combine
22629 power6x (mfpgpr) moves with other moves by using conditional
22630 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
22631 (mov<mode> for SFmode/SDmode); Likewise.
22632 (SFmode/SDmode splitters): Likewise.
22633 (movsf_hardfloat): Likewise.
22634 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
22635 (movsf_softfloat): Likewise.
22636 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
22638 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
22639 wx and wz constraints.
22641 * config/rs6000/constraints.md (wg constraint): New constraint to
22642 return FLOAT_REGS if -mmfpgpr (power6x) was used.
22644 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
22647 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22648 -mdebug=reg, print wg, wl, wx, and wz constraints.
22649 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22650 Initialize the reload functions for 64-bit binary/decimal floating
22652 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22653 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22654 create the buffer on the stack to overcome not having a 32-bit
22656 (rs6000_emit_move): Likewise.
22657 (rs6000_secondary_memory_needed_rtx): Likewise.
22658 (rs6000_alloc_sdmode_stack_slot): Likewise.
22659 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22660 via xxlxor, just like DFmode 0.0.
22662 * config/rs6000/dfp.md (movdd): Delete, combine with binary
22663 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22664 with other moves by using conditional constraits (wg). Use LFIWZX
22665 and STFIWX for loading SDmode on power7.
22666 (movdd splitters): Likewise.
22667 (movdd_hardfloat32): Likewise.
22668 (movdd_softfloat32): Likewise.
22669 (movdd_hardfloat64_mfpgpr): Likewise.
22670 (movdd_hardfloat64): Likewise.
22671 (movdd_softfloat64): Likewise.
22673 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
22674 64-bit binary and decimal floating point moves.
22675 (FMOVE64X): Likewise.
22676 (movdf): Combine 64-bit binary and decimal floating point moves.
22677 Combine power6x (mfpgpr) moves with other moves by using
22678 conditional constraits (wg).
22679 (mov<mode> for DFmode/DDmode): Likewise.
22680 (DFmode/DDmode splitters): Likewise.
22681 (movdf_hardfloat32): Likewise.
22682 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
22683 (movdf_softfloat32): Likewise.
22684 (movdf_hardfloat64_mfpgpr): Likewise.
22685 (movdf_hardfloat64): Likewise.
22686 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
22687 (movdf_softfloat64): Likewise.
22688 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
22689 (reload_<mode>_load): Move to later in the file so they aren't in
22690 the middle of the floating point move insns.
22691 (reload_<mode>_store): Likewise.
22693 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
22696 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
22697 constraint if -mdebug=reg.
22698 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
22699 Enable using dd reload support if needed.
22701 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
22702 binary and decimal floating point moves in rs6000.md.
22703 (movtd_internal): Likewise.
22705 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
22706 decimal floating point moves.
22708 (movtf_internal): Likewise.
22709 (mov<mode>_internal, TDmode/TFmode): Likewise.
22710 (movtf_softfloat): Likewise.
22711 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
22713 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
22714 movdi_internal64, using wg constraint for move direct operations.
22715 (movdi_internal64): Likewise.
22717 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
22718 MODES_TIEABLE_P for selected modes. Print the numerical value of
22719 the various virtual registers. Use GPR/FPR first/last values,
22720 instead of hard coding the register numbers. Print which modes
22721 have reload functions registered.
22722 (rs6000_option_override_internal): If -mdebug=reg, trace the options
22723 settings before/after setting cpu, target and subtarget settings.
22724 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
22725 and for secondary reload failures in rs6000_secondary_reload_inner.
22726 (rs6000_secondary_reload_fail): Likewise.
22727 (rs6000_secondary_reload_inner): Likewise.
22729 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
22730 macros for first/last GPR and FPR registers.
22731 (LAST_GPR_REGNO): Likewise.
22732 (FIRST_FPR_REGNO): Likewise.
22733 (LAST_FPR_REGNO): Likewise.
22735 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
22736 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
22737 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
22738 (vcond<mode><mode>): Likewise.
22739 (vcondu<mode><mode>): Likewise.
22740 (vector_gtu<mode>): Likewise.
22741 (vector_gte<mode>): Likewise.
22742 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
22743 to prevent the compiler from converting DImode operations to TImode.
22744 (ior<mode>3): Likewise.
22745 (and<mode>3): Likewise.
22746 (one_cmpl<mode>2): Likewise.
22747 (nor<mode>3): Likewise.
22748 (andc<mode>3): Likewise.
22750 * config/rs6000/constraints.md (wt constraint): New constraint
22751 that returns VSX_REGS if TImode is allowed in VSX registers.
22753 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
22754 constant under VSX.
22756 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
22757 similar to TImode, but it is restricted to being in the GPRs.
22759 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
22760 TImode to occupy a single VSX register.
22762 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
22763 -mvsx-timode for power7/power8.
22764 (power7 cpu): Likewise.
22765 (power8 cpu): Likewise.
22767 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
22768 sure that TFmode/TDmode take up two registers if they are ever
22769 allowed in the upper VSX registers.
22770 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
22772 (rs6000_init_hard_regno_mode_ok): Likewise.
22773 (rs6000_debug_reg_global): Add debugging for PTImode and wt
22774 constraint. Print if LRA is turned on.
22775 (rs6000_option_override_internal): Give an error if -mvsx-timode
22776 and VSX is not enabled.
22777 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
22778 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
22779 to reg+offset addressing. Use PTImode when checking offset
22780 addresses for validity.
22781 (reg_offset_addressing_ok_p): Likewise.
22782 (rs6000_legitimate_offset_address_p): Likewise.
22783 (rs6000_legitimize_address): Likewise.
22784 (rs6000_legitimize_reload_address): Likewise.
22785 (rs6000_legitimate_address_p): Likewise.
22786 (rs6000_eliminate_indexed_memrefs): Likewise.
22787 (rs6000_emit_move): Likewise.
22788 (rs6000_secondary_reload): Likewise.
22789 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
22790 reloads to fpr registers to continue to use reg+offset addressing,
22791 but 64-bit reloads to altivec registers need reg+reg addressing.
22792 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
22793 it. Treat LO_SUM like a PLUS operation.
22794 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
22795 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
22796 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
22797 registers to share a register with a smaller sized type, since VSX
22798 puts scalars in the upper 64-bits.
22799 (print_operand): Add support for PTImode.
22800 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
22801 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
22802 registers, but don't have arithmetic support.
22803 (rs6000_memory_move_cost): Add test for VSX.
22804 (rs6000_opt_masks): Add -mvsx-timode.
22806 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
22809 (VSr): Use wt constraint for TImode.
22810 (VSv): Drop TImode support.
22811 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
22812 (vsx_movti_64bit): Likewise.
22813 (vsx_movti_32bit): Likewise.
22814 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
22815 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
22816 one '?' on the appropriate output constraint. Do not allow TImode
22817 logical operations on 32-bit systems.
22818 (vsx_ior<mode>3): Likewise.
22819 (vsx_xor<mode>3): Likewise.
22820 (vsx_one_cmpl<mode>2): Likewise.
22821 (vsx_nor<mode>3): Likewise.
22822 (vsx_andc<mode>3): Likewise.
22823 (vsx_concat_<mode>): Likewise.
22824 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
22826 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
22827 OPTION_MASK_VSX_TIMODE.
22828 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
22829 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
22831 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
22832 (TI2 iterator): New iterator for TImode, PTImode.
22833 (wd mode attribute): Add values for vector types.
22834 (movti_string): Replace TI move operations with operations for TImode
22835 and PTImode. Add support for TImode being allowed in VSX registers.
22836 (mov<mode>_string, TImode/PTImode): Likewise.
22837 (movti_ppc64): Likewise.
22838 (mov<mode>_ppc64, TImode/PTImode): Likewise.
22839 (TI mode splitters): Likewise.
22841 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
22844 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
22846 PR tree-optimization/56355
22847 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
22848 Also handle integers with undefined overflow.
22850 2013-03-20 Catherine Moore <clm@codesourcery.com>
22851 Maciej W. Rozycki <macro@codesourcery.com>
22852 Tom de Vries <tom@codesourcery.com>
22853 Nathan Sidwell <nathan@codesourcery.com>
22854 Iain Sandoe <iain@codesourcery.com>
22855 Nathan Froyd <froydnj@codesourcery.com>
22856 Chao-ying Fu <fu@mips.com>
22858 * doc/extend.texi: (micromips, nomicromips, nocompression):
22859 Document new function attributes.
22860 * doc/invoke.texi (minterlink-compressed, mmicromips,
22861 m14k, m14ke, m14kec): Document new options.
22862 (minterlink-mips16): Update documentation.
22863 * doc/md.texi (ZC, ZD): Document new constraints.
22864 * configure.ac (gcc_cv_as_micromips): Check if linker
22865 supports the .set micromips directive.
22866 * configure: Regenerate.
22867 * config.in: Regenerate.
22868 * config/mips/mips-tables.opt: Regenerate.
22869 * config/mips/micromips.md: New file.
22870 * constraints.md (ZC, ZD): New constraints.
22871 * config/mips/predicates.md (movep_src_register): New predicate.
22872 (movep_src_operand): New predicate.
22873 (non_volatile_mem_operand): New predicate.
22874 * config/mips/mips.md (multimem): New type.
22875 (length): Differentiate between 17-bit and 18-bit branch offsets.
22876 (MOVEP1, MOVEP2): New mode iterator.
22877 (mov_<load>l): Use ZC constraint.
22878 (mov_<load>r): Likewise.
22879 (mov_<store>l): Likewise.
22880 (mov_<store>r): Likewise.
22881 (*branch_equality<mode>_inverted): Add microMIPS support.
22882 (*branch_equality<mode>): Likewise.
22883 (*jump_absolute): Likewise.
22884 (indirect_jump_<mode>): Likewise.
22885 (tablejump_<mode>): Likewise.
22886 (<optab>_internal): Likewise.
22887 (sibcall_internal): Likewise.
22888 (sibcall_value_internal): Likewise.
22889 (prefetch): Use constraint ZD.
22890 * config/mips/mips.opt (minterlink-compressed): New option.
22891 (minterlink-mips16): Now an alias for minterlink-compressed.
22892 (mmicromips): New option.
22893 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
22894 (compare_and_swap_12): Likewise.
22895 (sync_add<mode>): Likewise.
22896 (sync_<optab>_12): Likewise.
22897 (sync_old_<optab>_12): Likewise.
22898 (sync_new_<optab>_12): Likewise.
22899 (sync_nand_12): Likewise.
22900 (sync_old_nand_12): Likewise.
22901 (sync_new_nand_12): Likewise.
22902 (sync_sub<mode>): Likewise.
22903 (sync_old_add<mode>): Likewise.
22904 (sync_old_sub<mode>): Likewise.
22905 (sync_new_add<mode>): Likewise.
22906 (sync_new_sub<mode>): Likewise.
22907 (sync_<optab><mode>): Likewise.
22908 (sync_old_<optab><mode>): Likewise.
22909 (sync_new_<optab><mode>): Likewise.
22910 (sync_nand<mode>): Likewise.
22911 (sync_old_nand<mode>): Likewise.
22912 (sync_new_nand<mode>): Likewise.
22913 (sync_lock_test_and_set<mode>): Likewise.
22914 (test_and_set_12): Likewise.
22915 (atomic_compare_and_swap<mode>): Likewise.
22916 (atomic_exchange<mode>_llsc): Likewise.
22917 (atomic_fetch_add<mode>_llsc): Likewise.
22918 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
22919 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
22920 (umips_save_restore_pattern_p): Likewise.
22921 (umips_load_store_pair_p): Likewise.
22922 (umips_output_load_store_pair): Likewise.
22923 (umips_movep_target_p): Likewise.
22924 (umips_12bit_offset_address_p): Likewise.
22925 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
22926 (mips_base_mips16): Rename this...
22927 (mips_base_compression_flags): ...to this. Update all uses.
22928 (mips_attribute_table): Add micromips, nomicromips and nocompression.
22929 (mips_mips16_decl_p): Delete.
22930 (mips_nomips16_decl_p): Delete.
22931 (mips_get_compress_on_flags): New function.
22932 (mips_get_compress_off_flags): New function.
22933 (mips_get_compress_mode): New function.
22934 (mips_get_compress_on_name): New function.
22935 (mips_get_compress_off_name): New function.
22936 (mips_insert_attributes): Support multiple compression types.
22937 (mips_merge_decl_attributes): Likewise.
22938 (umips_12bit_offset_address_p): New function.
22939 (mips_start_function_definition): Emit .set micromips directive.
22940 (mips_call_may_need_jalx_p): New function.
22941 (mips_function_ok_for_sibcall): Add microMIPS support.
22942 (mips_print_operand_punctuation): Support short delay slots and
22944 (umips_swm_mask, umips_swm_encoding): New.
22945 (umips_build_save_restore): New function.
22946 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
22947 (was_mips16_p): Remove.
22948 (old_compression_mode): New.
22949 (mips_set_compression_mode): New function.
22950 (mips_set_current_function): Add microMIPS support.
22951 (mips_option_override): Likewise.
22952 (umips_save_restore_pattern_p): New function.
22953 (umips_output_save_restore): New function.
22954 (umips_load_store_pair_p_1): New function.
22955 (umips_load_store_pair_p): New function.
22956 (umips_output_load_store_pair_1): New function.
22957 (umips_output_load_store_pair): New function.
22958 (umips_movep_target_p) New function.
22959 (mips_prepare_pch_save): Add microMIPS support.
22960 * config/mips/mips.h (TARGET_COMPRESSION): New.
22961 (TARGET_CPU_CPP_BUILTINS): Update macro
22962 to use new compression flags and to support microMIPS.
22963 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
22964 (MIPS_ARCH_FLOAT_SPEC): Likewise.
22965 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
22966 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
22967 (ASM_SPEC): Support mmicromips and mno-micromips.
22968 (M16STORE_REG_P): New macro.
22969 (MIPS_CALL): Support TARGET_MICROMIPS.
22970 (MICROMIPS_J): New macro.
22971 (mips_base_mips16): Rename this...
22972 (mips_base_compression_flags): ...to this.
22973 (UMIPS_12BIT_OFFSET_P): New macro.
22974 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
22975 (MULTILIB_DIRNAMES): Likewise.
22976 2013-03-20 Richard Biener <rguenther@suse.de>
22978 PR tree-optimization/56661
22979 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
22980 the result does not have to be distinct.
22982 2013-03-20 Richard Biener <rguenther@suse.de>
22984 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
22987 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22988 Steven Bosscher <steven@gcc.gnu.org>
22990 PR rtl-optimization/56605
22991 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
22993 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22996 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
22997 that require movd instead of movq.
22999 2013-03-20 Richard Biener <rguenther@suse.de>
23001 * tree-ssa-structalias.c (struct variable_info): Add pointer
23002 to the first field of an aggregate with sub-vars. Make
23003 this and the pointer to the next subfield its ID.
23004 (vi_next): New function.
23005 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
23006 storedanything_id, integer_id): Increment by one.
23007 (new_var_info, get_call_vi, lookup_call_clobber_vi,
23008 get_call_clobber_vi): Adjust.
23009 (solution_set_expand): Simplify and speedup.
23010 (solution_set_add): Inline into ...
23011 (set_union_with_increment): ... this. Adjust accordingly.
23012 (do_sd_constraint): Likewise.
23013 (do_ds_constraint): Likewise.
23014 (do_complex_constraint): Simplify.
23015 (build_pred_graph): Adjust.
23016 (solve_graph): Likewise. Simplify and speedup.
23017 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
23018 get_constraint_for_component_ref, get_constraint_for_1,
23019 first_vi_for_offset, first_or_preceding_vi_for_offset,
23020 create_function_info_for, create_variable_info_for_1,
23021 create_variable_info_for, intra_create_variable_infos): Adjust.
23022 (init_base_vars): Push NULL for ID zero.
23023 (compute_points_to_sets): Adjust.
23025 2013-03-20 Richard Biener <rguenther@suse.de>
23027 * cfgloop.c (verify_loop_structure): Streamline and avoid
23028 ICEing on corrupt loop tree.
23029 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
23032 2013-03-20 Richard Biener <rguenther@suse.de>
23034 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
23035 check whether an SSA update is needed.
23037 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
23039 * config/mips/constraints.md (T): Rename to...
23043 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
23044 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
23046 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23048 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
23049 (*subsi3_carryin_uxtw): Likewise.
23051 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23053 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
23054 (*rorsi3_insn_uxtw): Likewise.
23056 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23058 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
23059 (*extrsi5_insn_uxtw): Likewise.
23061 2013-03-19 Richard Biener <rguenther@suse.de>
23063 PR tree-optimization/56273
23064 * passes.c (init_optimization_passes): Move second VRP after DOM.
23066 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
23068 * config/i386/i386.md (*movti_internal): Merge from
23069 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
23070 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
23071 nox64 isa attributes.
23073 2013-03-18 Richard Biener <rguenther@suse.de>
23075 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
23077 (merge_node_constraints): Likewise.
23078 (build_succ_graph): Likewise.
23079 (valid_graph_edge): Inline into single caller.
23080 (unify_nodes): Likewise. Use bitmap_set_bit return value
23082 (scc_visit): Fix formatting and variable use.
23083 (do_sd_constraint): Use gcc_checking_assert.
23084 (do_ds_constraint): Likewise.
23085 (do_complex_constraint): Likewise.
23086 (condense_visit): Likewise. Cleanup.
23087 (dump_pred_graph): New function.
23088 (perform_var_substitution): Dump the pred-graph before
23089 variable substitution.
23090 (find_equivalent_node): Use gcc_checking_assert.
23091 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
23093 2013-03-18 Richard Biener <rguenther@suse.de>
23095 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
23096 Remove cond_expr_stmt_list argument and do not gimplify the
23098 (vect_loop_versioning): Adjust.
23099 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
23100 Cleanup to use less temporaries.
23101 (vect_create_data_ref_ptr): Cleanup.
23103 2013-03-18 Jakub Jelinek <jakub@redhat.com>
23105 PR tree-optimization/56635
23106 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
23107 require types_compatible_p types.
23109 2013-03-18 Nick Clifton <nickc@redhat.com>
23111 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
23112 spurious backslash.
23114 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
23115 Add missing line to comment describing function.
23117 2013-03-18 Richard Biener <rguenther@suse.de>
23119 PR tree-optimization/56210
23120 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
23121 Handle string / character search functions.
23122 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
23124 2013-03-18 Richard Biener <rguenther@suse.de>
23126 PR middle-end/56483
23127 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
23128 and implement properly.
23129 * gimple.h (gimple_cond_single_var_p): Remove.
23131 2013-03-18 Richard Biener <rguenther@suse.de>
23133 * tree-data-ref.h (find_data_references_in_loop): Declare.
23134 * tree-data-ref.c (get_references_in_stmt): Use a stack
23135 vector pre-allocated in the callers.
23136 (find_data_references_in_stmt): Adjust.
23137 (graphite_find_data_references_in_stmt): Likewise.
23138 (create_rdg_vertices): Likewise.
23139 (find_data_references_in_loop): Export.
23140 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
23141 Compute dependences here...
23142 (vect_analyze_data_refs): ...not here. When we encounter
23143 a non-vectorizable data reference in basic-block vectorization
23144 truncate the data reference vector. Do not bother to
23145 fixup data-dependence information for gather loads.
23146 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
23147 of data references, as reported.
23149 2013-03-18 Richard Biener <rguenther@suse.de>
23151 PR tree-optimization/3713
23152 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
23153 has_constants and expr.
23154 (stmt_has_constants): Properly valueize SSA names when deciding
23155 whether the stmt has constants.
23157 2013-03-18 Richard Biener <rguenther@suse.de>
23159 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
23160 whole function when there is nothing to do.
23161 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
23162 * tree-vectorizer.c (vectorize_loops): Update virtual and
23163 loop-closed SSA once.
23164 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
23166 2013-03-18 Richard Biener <rguenther@suse.de>
23168 PR middle-end/56113
23169 * domwalk.c (bb_postorder): New global static.
23170 (cmp_bb_postorder): New function.
23171 (walk_dominator_tree): Replace scheme imposing an order for
23172 visiting dominator sons by one sorting them at the time they
23173 are pushed on the stack.
23175 2013-03-18 Richard Biener <rguenther@suse.de>
23177 PR tree-optimization/39326
23178 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
23179 (struct mem_ref): Replace mem member with ao_ref typed member.
23180 (MEM_ANALYZABLE): Adjust.
23181 (memref_eq): Likewise.
23182 (mem_ref_alloc): Likewise.
23183 (gather_mem_refs_stmt): Likewise.
23184 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
23185 (execute_sm_if_changed_flag_set): Adjust.
23186 (execute_sm): Likewise.
23187 (ref_always_accessed_p): Likewise.
23188 (refs_independent_p): Likewise.
23189 (can_sm_ref_p): Likewise.
23191 2013-03-18 Jakub Jelinek <jakub@redhat.com>
23194 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
23195 return 1 even for !unsignedp.
23197 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23199 * config/i386/i386.md (isa): Add x64 and nox64.
23200 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
23201 (*pushtf): Enable *roF alternative for x64 isa only.
23202 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
23203 mode attribute of integer alternatives to DImode for TARGET_64BIT.
23204 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
23205 (*movtf_internal): Merge from *movtf_internal_rex64 and
23206 *movtf_internal_sse. Use x64 and nox64 isa attributes.
23207 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
23208 nox64 isa attributes.
23209 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
23210 nox64 isa attributes.
23211 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
23213 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23215 * config/alpha/alpha.c (TARGET_LRA_P): New define.
23217 2013-03-17 Jakub Jelinek <jakub@redhat.com>
23220 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
23221 class names. Remove trailing comma after "ALL_REGS".
23223 2013-03-16 Jan Hubicka <jh@suse.cz>
23225 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
23226 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
23227 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
23228 of cgraph_get_create_node.
23229 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
23231 2013-03-16 Jason Merrill <jason@redhat.com>
23234 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
23235 with DW_AT_default_value.
23237 2013-03-16 Jakub Jelinek <jakub@redhat.com>
23239 * BASE-VER: Set to 4.9.0.
23241 2013-03-14 Andi Kleen <ak@linux.intel.com>
23244 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
23245 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
23246 Document _x* TSX intrinsics.
23248 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
23249 David Holsgrove <david.holsgrove@xilinx.com>
23251 * configure.ac: Add MicroBlaze TLS support detection.
23252 * configure: Regenerate.
23253 * config/microblaze/microblaze-protos.h
23254 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
23255 symbol_mentioned_p, label_mentioned_p): Add prototypes.
23256 * config/microblaze/microblaze.c (microblaze_address_type): Add
23257 ADDRESS_TLS and tls_reloc address types.
23258 (microblaze_address_info): Add tls_reloc.
23259 (TARGET_HAVE_TLS): Define.
23260 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
23261 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
23262 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
23263 load_tls_operand, microblaze_call_tls_get_addr,
23264 microblaze_legitimize_tls_address): New functions.
23265 (microblaze_classify_unspec): Handle UNSPEC_TLS.
23266 (get_base_reg): Use microblaze_tls_symbol_p.
23267 (microblaze_classify_address): Handle TLS.
23268 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
23269 label_mentioned_p and microblaze_tls_referenced_p.
23270 (microblaze_legitimize_address): Handle TLS.
23271 (microblaze_address_insns): Handle ADDRESS_TLS.
23272 (pic_address_needs_scratch): Handle TLS.
23273 (print_operand_address): Handle TLS.
23274 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
23275 (microblaze_expand_move): Handle TLS.
23276 (microblaze_legitimate_constant_p): Check
23277 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
23278 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
23279 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
23280 (PIC_OFFSET_TABLE_REGNUM): Set.
23281 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
23282 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
23283 (addsi3, movsi_internal2, movdf_internal): Update constraints
23284 * config/microblaze/predicates.md (arith_plus_operand): Define
23285 (move_operand): Redefine as move_src_operand,
23286 check microblaze_tls_referenced_p.
23288 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23290 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
23291 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
23293 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23295 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
23298 2013-03-14 Jakub Jelinek <jakub@redhat.com>
23300 PR tree-optimization/53265
23301 * common.opt (Waggressive-loop-optimizations): New option.
23302 * tree-ssa-loop-niter.c: Include tree-pass.h.
23303 (do_warn_aggressive_loop_optimizations): New function.
23304 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
23305 if number_of_latch_executions returned constant.
23306 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
23307 early. If number_of_latch_executions returned constant, set
23308 nb_iterations_upper_bound back to it.
23309 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
23311 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
23312 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
23314 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
23315 (MULTILIB_OSDIRNAMES): Set.
23316 * genmultilib: If defaultosdirname doesn't start with :: , set
23317 defaultosdirname2 instead, clear it and emit two . multilib_raw
23318 entries instead of just one.
23320 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
23322 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23323 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
23324 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23325 (SUBTARGET_OVERRIDE_OPTIONS): New.
23327 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
23330 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
23331 (musermode): Convert to Var(TARGET_USERMODE).
23332 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
23333 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
23334 * config/sh/sh.c (sh_option_override): Use
23335 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
23336 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
23338 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
23340 (udivsi3_i4_single, divsi3_i4_single): Use
23341 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
23343 2013-03-13 Dave Korn <dave.korn.cygwin@....>
23345 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
23348 2013-03-13 Richard Biener <rguenther@suse.de>
23350 PR tree-optimization/56608
23351 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
23352 calls when vectorizing basic-blocks.
23354 2013-03-13 Jakub Jelinek <jakub@redhat.com>
23357 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
23360 2013-03-12 Jakub Jelinek <jakub@redhat.com>
23362 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
23364 2013-03-11 Jan Hubicka <jh@suse.cz>
23367 * lto-streamer-out.c (output_symbol_p): Skip references from
23368 constructors of external variables.
23370 2013-03-11 Jan Hubicka <jh@suse.cz>
23372 PR middle-end/56571
23373 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
23375 * emit-rtl.c (verify_rtx_sharing): Likewise.
23376 (copy_insn_1): Likewise.
23377 * rtl.c (copy_rtx): Likewise.
23379 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
23382 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
23383 output_operand_lossage message.
23385 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
23388 * arm.c (shift_op): Validate RTL pattern on the fly.
23389 (arm_print_operand, case 'S'): Don't use shift_operator to validate
23392 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23395 * config/pa/pa.md (call_value): Check for calls to powf and direct to
23396 new call patterns that clobber %fr12.
23397 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
23398 split and postreload patterns.
23399 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
23400 registers %fr12 and %fr12R as call used.
23402 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
23404 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
23405 (canon_address, record_store, replace_read, check_mem_read_rtx,
23406 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
23407 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
23408 rest_of_handle_dse): Likewise.
23410 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
23412 PR middle-end/56524
23413 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
23415 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
23416 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
23417 (save_optabs_if_changed): Replace with...
23418 (init_tree_optimization_optabs): ...this.
23419 * optabs.c (save_optabs_if_changed): Rename to...
23420 (init_tree_optimization_optabs): ...this. Take the optimization node
23421 as argument. Do nothing if the base optabs are already correct.
23422 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
23423 to recompute optabs.
23424 * function.h (function): Remove optabs field.
23425 * function.c (invoke_set_current_function_hook): Call
23426 init_tree_optimization_optabs. Use the result to initialize
23429 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
23431 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
23432 if GTMA_HAS_NO_INSTRUMENTATION.
23433 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
23434 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
23435 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
23436 * gimple-pretty-print.c (dump_gimple_transaction): Handle
23437 GTMA_HAS_NO_INSTRUMENTATION.
23439 2013-03-08 Jakub Jelinek <jakub@redhat.com>
23441 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
23444 2013-03-08 Marek Polacek <polacek@redhat.com>
23445 Jakub Jelinek <jakub@redhat.com>
23447 PR tree-optimization/56478
23448 * predict.c (is_comparison_with_loop_invariant_p): Change the
23449 type of loop_step to tree.
23450 (predict_loops): Adjust.
23451 (predict_iv_comparison): Perform the computations on double_ints.
23453 2013-03-08 Richard Biener <rguenther@suse.de>
23455 PR tree-optimization/56570
23456 * tree-cfg.c (verify_expr_location_1): Verify locations for
23458 * tree-sra.c (create_access_replacement): Strip locations
23459 from DECL_DEBUG_EXPRs.
23461 2013-03-08 Richard Biener <rguenther@suse.de>
23463 * tree-inline.c (expand_call_inline): Do not associate
23464 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
23465 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
23467 2013-03-08 Richard Biener <rguenther@suse.de>
23469 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
23470 or block changes with -Og. Fix for location / block encoding
23471 changes and PHI arguments with locations.
23473 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
23475 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
23477 (struct output_info): Likewise.
23478 (register_overhead): Remove bad gcc_assert.
23479 (bitmap_find_bit): If there is only a single bitmap element, do not
23480 count a miss as a search.
23481 (print_statistics): Update for counter type changes.
23482 (dump_bitmap_statistics): Likewise. Print headers such that they
23483 are properly lined up with the printed counters.
23485 2013-03-07 Jakub Jelinek <jakub@redhat.com>
23487 PR tree-optimization/56559
23488 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
23489 check that it has only a single use.
23491 2013-03-07 Richard Biener <rguenther@suse.de>
23493 * doc/invoke.texi (fwhole-program): Discourage use in combination
23496 2013-03-06 Jakub Jelinek <jakub@redhat.com>
23498 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
23500 PR tree-optimization/56539
23501 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
23502 instead of GSI_CONTINUE_LINKING as last argument to
23503 force_gimple_operand_gsi. Adjust function comment.
23505 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
23508 PR middle-end/56548
23509 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
23510 promoted mode, convert the result back to the original mode.
23512 2013-03-06 Richard Biener <rguenther@suse.de>
23514 PR middle-end/56294
23515 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
23516 (insert_updated_phi_nodes_compare_uids): New function.
23517 (update_ssa): Sort symbols_to_rename after UID before
23518 traversing it to insert PHI nodes.
23520 2013-03-06 Richard Biener <rguenther@suse.de>
23522 PR middle-end/50494
23523 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
23524 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
23527 2013-02-13 Richard Biener <rguenther@suse.de>
23530 * varasm.c (output_constant_def_1): Get the decl representing
23531 the constant as argument.
23532 (output_constant_def): Wrap output_constant_def_1.
23533 (make_decl_rtl): Use output_constant_def_1 with the decl
23534 representing the constant.
23535 (build_constant_desc): Optionally re-use a decl already
23536 representing the constant.
23537 (tree_output_constant_def): Adjust.
23539 2013-03-06 Joey Ye <joey.ye@arm.com>
23542 * gcc.c (convert_white_space): New function.
23543 (main): Handles white space in function name.
23545 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
23548 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
23549 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
23550 to SH_DIV_CALL_TABLE for TARGET_SH2.
23551 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
23553 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
23554 call-table options.
23556 2013-03-05 Sterling Augustine <saugustine@google.com>
23557 Cary Coutant <ccoutant@google.com>
23560 * dwarf2out.c (resolve_addr): Don't call
23561 remove_loc_list_addr_table_entries a second time for the same
23564 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23567 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
23568 (avoid_complex_debug_insns): New function.
23569 (expand_debug_locations): Call it.
23571 PR rtl-optimization/56484
23572 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
23573 lifetimes of hard registers on small register class machines.
23575 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23577 * config/microblaze/microblaze-protos.h: Rename
23578 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
23579 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
23581 (microblaze_fast_interrupt_function_p): New function.
23582 (microblaze_is_interrupt_handler): Rename to
23583 microblaze_is_interrupt_variant and add fast_interrupt check.
23584 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
23585 (save_restore_insns): Likewise.
23586 (compute_frame_size): Likewise.
23587 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
23588 (microblaze_globalize_label): Likewise.
23589 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
23590 * config/microblaze/microblaze.md: Use wrapper
23591 microblaze_is_interrupt_variant.
23593 2013-03-05 Kai Tietz <ktietz@redhat.com>
23595 * sdbout.c (sdbout_one_type): Switch to current function's section
23596 supporting cold/hot.
23598 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23600 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
23603 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23605 PR middle-end/56461
23606 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
23607 if VALGRIND_GET_VBITS is defined, temporarily make object
23608 memory all defined, and restore previous valgrind addressability
23609 and definability afterwards. Free this_object at the end.
23611 PR middle-end/56461
23612 * lra.c (lra): Call lra_clear_live_ranges if live_p,
23613 right before calling lra_create_live_ranges, also call it
23614 when clearing live_p. Only call lra_clear_live_ranges
23615 at the end if live_p.
23617 PR middle-end/56461
23618 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
23620 2013-03-05 Richard Biener <rguenther@suse.de>
23622 PR tree-optimization/56521
23623 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
23626 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
23629 * except.h (remove_unreachable_eh_regions): New prototype.
23630 * except.c (remove_eh_handler_splicer): New function, split out
23631 of remove_eh_handler.
23632 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
23633 warning about running it on many EH regions one at a time.
23634 (remove_unreachable_eh_regions_worker): New function, walk the
23635 EH tree in depth-first order and remove non-marked regions.
23636 (remove_unreachable_eh_regions): New function.
23637 * tree-eh.c (mark_reachable_handlers): New function, split out
23638 from remove_unreachable_handlers.
23639 (remove_unreachable_handlers): Use mark_reachable_handlers and
23640 remove_unreachable_eh_regions.
23641 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
23642 and remove_unreachable_eh_regions.
23644 2013-03-05 Richard Biener <rguenther@suse.de>
23646 PR middle-end/56525
23647 * loop-init.c (fix_loop_structure): Remove loops in two stages,
23648 not freeing them until the end.
23650 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23652 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
23654 2013-03-05 Richard Biener <rguenther@suse.de>
23656 PR tree-optimization/56270
23657 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
23658 of loads after scheduling an SLP instance.
23660 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23662 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
23664 (check_gcc_parallelize): Run guality.exp as a separate job from
23665 vect.exp with unsorted.exp and $(dg_target_exps) separately from
23666 struct-layout-1.exp with stackalign.exp.
23668 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
23670 PR middle-end/56461
23671 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
23672 load_index sbitmap even if some bit in it isn't set.
23674 PR middle-end/56461
23675 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
23676 (discover_iteration_bound_by_body_walk): Change queues to
23677 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
23678 spelling in comment. Call safe_push on queues[bound_index] directly.
23679 Release queues[queue_index] in every iteration unconditionally.
23680 Release bounds vector.
23682 PR middle-end/56461
23683 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
23684 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
23685 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
23688 2013-03-05 Richard Biener <rguenther@suse.de>
23691 * tree-inline.c (remap_blocks_to_null): New function.
23692 (expand_call_inline): When expanding a call stmt without
23693 an associated block inline remap all callee blocks to NULL.
23695 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23697 PR rtl-optimization/56494
23698 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
23699 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
23700 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
23702 PR middle-end/56461
23703 * sel-sched-ir.c (free_sched_pools): Release
23704 succs_info_pool.stack[succs_info_pool.max_top] vectors too
23705 if succs_info_pool.max_top isn't -1.
23708 * opts.c (opts_obstack, opts_concat): Moved to...
23709 * opts-common.c (opts_obstack, opts_concat): ... here.
23711 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23713 PR middle-end/56461
23714 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
23716 2013-03-04 Martin Jambor <mjambor@suse.cz>
23718 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
23719 all appropriate places.
23721 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
23723 PR tree-optimization/56424
23724 * ipa-split.c (split_function): Do not set the RSO flag if result is
23725 not by reference and its type is a register type.
23727 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
23729 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
23730 (microblaze_legitimate_pic_operand): Likewise
23731 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
23732 new function microblaze_legitimate_pic_operand
23733 * config/microblaze/microblaze-protos.h
23734 (microblaze_legitimate_pic_operand): Declare.
23736 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23738 * config/microblaze/predicates.md (call_insn_simple_operand):
23739 New predicate for supported rtx code types.
23740 * config/microblaze/microblaze.md (call_internal1): Use
23741 call_insn_simple_operand predicate.
23743 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23745 PR middle-end/56461
23746 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
23747 partitions.ordered_remove.
23749 PR middle-end/56461
23750 * tree-vect-stmts.c (vectorizable_conversion): Don't call
23751 vec_oprnds0.create (1) for modifier == NONE.
23753 PR middle-end/56461
23754 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
23755 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
23756 vec_oprnds1 right before pushing anything to it for
23759 PR middle-end/56461
23760 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
23761 set nbbs to 0 instead of having separate code path.
23762 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
23763 instead of false as last argument if returning NULL.
23765 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
23767 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
23768 the attribute is now called "target" instead of "option".
23769 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
23770 * doc/tm.texi.in (Target Attributes): Likewise document the correct
23771 attribute/pragma name for TARGET_OPTION_VALID_P and
23772 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
23773 * doc/tm.texi: Regenerated.
23775 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
23777 * config/microblaze/microblaze.c:
23778 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
23779 * config/microblaze/microblaze.h: Add -mxl-reorder to
23781 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
23782 instructions emitted if TARGET_REORDER.
23783 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
23784 or 0 for -m/-mno case, but initialises as 2 to detect default use case
23787 2013-03-01 Xinliang David Li <davidxl@google.com>
23789 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
23792 2013-03-01 Jakub Jelinek <jakub@redhat.com>
23794 PR middle-end/56461
23795 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
23796 vector even when returning true. Fix up function comment formatting.
23798 PR middle-end/56461
23799 * ira-build.c (ira_loop_nodes_count): New variable.
23800 (create_loop_tree_nodes): Initialize it.
23801 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
23803 PR middle-end/56461
23804 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
23805 method on dr_chain and result_chain.
23806 * tree-vect-stmts.c (vectorizable_store): Only call
23807 result_chain.create if j == 0.
23809 PR middle-end/56461
23810 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
23811 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
23812 before overwriting it.
23814 2013-03-01 Tobias Burnus <burnus@net-b.de>
23816 * doc/extended.texi (C Extensions): Change order in @menu
23818 (Other MIPS Built-in Functions): Move last MIPS entry before
23819 "picoChip Built-in Functions".
23820 (SH Built-in Functions): Move after RX Built-in Functions.
23821 * doc/gcc.texi (Introduction): Change order in @menu
23823 * doc/md.texi (Constraints): Ditto.
23824 * gty.texi (Type Information): Ditto.
23825 (User-provided marking routines for template types): Make
23827 * doc/invoke.texi (AArch64 Options): Move before
23828 "Adapteva Epiphany Options".
23830 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
23831 Jakub Jelinek <jakub@redhat.com>
23834 * asan.c (gate_asan): Lookup no_sanitize_address instead of
23835 no_address_safety_analysis attribute.
23836 * doc/extend.texi (no_address_safety_attribute): Rename to
23837 no_sanitize_address attribute, mention no_address_safety_analysis
23838 attribute as deprecated alias.
23840 2013-02-28 Jakub Jelinek <jakub@redhat.com>
23842 PR middle-end/56461
23843 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
23844 type to vec<vec<tree> > *.
23845 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
23846 to be vec<tree> instead of vec<tree> *, set vec_defs
23847 to vNULL and call vec_defs.create (number_of_vects), adjust other
23849 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
23850 vectorizable_condition): Adjust vect_get_slp_defs callers.
23852 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23854 * config/aarch64/aarch64.c
23855 (aarch64_float_const_representable): Remove unused variable.
23857 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23859 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
23861 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23863 * config/aarch64/aarch64-builtins.c
23864 (aarch64_init_simd_builtins): Make static.
23866 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23868 * config/aarch64/aarch64.c
23869 (aarch64_simd_make_constant): Make static.
23871 2013-02-28 Martin Jambor <mjambor@suse.cz>
23873 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
23874 with no initialization to the RHS of debug statements.
23876 2013-02-28 Martin Jambor <mjambor@suse.cz>
23878 PR tree-optimization/56294
23879 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
23881 (get_access_replacement): Do not call create_access_replacement.
23882 Assert a replacement exists.
23883 (get_repl_default_def_ssa_name): Create the replacement declaration
23886 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23888 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
23889 final_end_function.
23891 2013-02-28 Marek Polacek <polacek@redhat.com>
23893 PR rtl-optimization/56466
23894 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
23895 if we're changing a loop.
23896 (peel_loops_completely): Likewise.
23898 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
23901 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
23903 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23906 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
23907 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
23908 INTX_FTYPE_FX, FX_FTYPE_INTX.
23909 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
23911 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23913 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
23914 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
23915 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
23916 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
23917 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
23918 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
23919 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
23920 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
23921 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
23922 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
23923 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
23924 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
23925 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
23926 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
23927 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
23928 (avrxmega6): Increase max flash segments from 5 to 6.
23929 * config/avr/t-multilib: Regenerate.
23930 * config/avr/avr-tables.opt: Regenerate.
23931 * doc/avr-mmcu.texi: Regenerate.
23933 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23935 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
23936 (avr_device_to_arch): Rename to avr_device_to_ld.
23937 (avr_device_to_as): New prototype.
23938 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
23939 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
23940 * config/avr/driver-avr.c (avr_device_to_as): New.
23941 (avr_device_to_arch): Rename to avr_device_to_ld.
23943 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23945 PR middle-end/56461
23946 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
23947 method on dr_chain and result_chain.
23949 PR middle-end/56461
23950 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
23951 pointer_set_destroy on not_executed_last_iteration.
23953 PR middle-end/56461
23954 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
23956 PR middle-end/56461
23957 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
23958 FOR_EACH_DEFINED_FUNCTION when freeing state.
23960 PR middle-end/56461
23961 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
23963 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
23966 PR middle-end/56461
23967 * ipa-cp.c (decide_whether_version_node): Call vec_free on
23968 known_aggs[i].items and release known_aggs vector.
23970 PR middle-end/56461
23971 * ipa-reference.c (propagate): Free node_info even for alias nodes.
23973 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23975 * config/microblaze/microblaze.c (microblaze_emit_compare):
23976 Use xor for EQ/NE comparisions.
23977 * config/microblaze/microblaze.md (cstoresf4): Add constraints
23978 (cbranchsf4): Adjust operator to comparison_operator.
23980 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23982 PR middle-end/56461
23983 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
23985 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
23986 vec_safe_push, always update *slot.
23987 (redirect_edge_var_map_clear): Use vec_free.
23988 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
23989 (free_var_map_entry): Use vec_free.
23990 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
23991 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
23993 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
23995 PR middle-end/45472
23996 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
23997 when the may_trap_p bit of the exprs being merged differs.
23998 Reorder tests for speculativeness in the logical and operator.
24000 2013-02-27 Jakub Jelinek <jakub@redhat.com>
24002 * incpath.c (add_standard_paths): Use reconcat instead of concat
24003 where appropriate and avoid leaking memory.
24005 * opts.h: Include obstack.h.
24006 (opts_concat): New prototype.
24007 (opts_obstack): New declaration.
24008 * opts.c (opts_concat): New function.
24009 (opts_obstack): New variable.
24010 (init_options_struct): Call gcc_init_obstack on opts_obstack.
24011 (finish_options): Use opts_concat instead of concat
24012 and XOBNEWVEC instead of XNEWVEC.
24013 * opts-common.c (generate_canonical_option, decode_cmdline_option,
24014 generate_option): Likewise.
24015 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
24016 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
24019 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
24020 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
24022 2013-02-26 Jakub Jelinek <jakub@redhat.com>
24024 PR middle-end/56461
24025 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
24027 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
24029 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
24030 (arm_block_move_unaligned_straight): Likewise.
24031 (arm_adjust_block_mem): Likewise.
24033 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
24036 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
24037 temp, cond and label.
24038 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
24041 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
24042 * config/c6x/c6x.h (dbx_register_map): Update declaration.
24045 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
24046 of prologue/epilogue functions.
24049 * config/tilegx/tilegx.c (tilegx_expand_prologue):
24050 Remove unused variable cfa_offset.
24051 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
24054 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
24055 type promotion to unsigned.
24058 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
24059 for HOST_WIDE_INT of 32 bit / same size as int.
24060 (arm_block_move_unaligned_straight): Likewise.
24061 (arm_adjust_block_mem): Likewise.
24064 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
24067 2013-02-26 Marek Polacek <polacek@redhat.com>
24069 PR tree-optimization/56426
24070 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
24072 2013-02-26 Richard Biener <rguenther@suse.de>
24075 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
24076 unused variable loops.
24078 2013-02-26 Jakub Jelinek <jakub@redhat.com>
24080 PR tree-optimization/56448
24081 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
24082 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
24083 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
24084 later operands of the references, or even first operand for
24085 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
24087 PR tree-optimization/56443
24088 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
24089 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
24090 to type_for_mode langhook.
24092 2013-02-25 Matt Turner <mattst88@gmail.com>
24094 * doc/invoke.texi: Document r4700.
24096 2013-02-25 Richard Biener <rguenther@suse.de>
24098 PR tree-optimization/56175
24099 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
24101 (simplify_bitwise_binary): ... here. Also guard the conversion
24102 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
24104 2013-02-25 Catherine Moore <clm@codesourcery.com>
24107 2013-02-24 Catherine Moore <clm@codesourcery.com>
24108 Maciej W. Rozycki <macro@codesourcery.com>
24109 Tom de Vries <tom@codesourcery.com>
24110 Nathan Sidwell <nathan@codesourcery.com>
24111 Iain Sandoe <iain@codesourcery.com>
24112 Nathan Froyd <froydnj@codesourcery.com>
24113 Chao-ying Fu <fu@mips.com>
24115 * doc/extend.texi: (micromips, nomicromips, nocompression):
24116 Document new function attributes.
24117 * doc/invoke.texi (minterlink-compressed, mmicromips,
24118 m14k, m14ke, m14kec): Document new options.
24119 (minterlink-mips16): Update documentation.
24120 * doc/md.texi (ZC, ZD): Document new constraints.
24121 * configure.ac (gcc_cv_as_micromips): Check if linker
24122 supports the .set micromips directive.
24123 * configure: Regenerate.
24124 * config.in: Regenerate.
24125 * config/mips/mips-tables.opt: Regenerate.
24126 * config/mips/micromips.md: New file.
24127 * constraints.md (ZC, AD): New constraints.
24128 * config/mips/predicates.md (movep_src_register): New predicate.
24129 (movep_src_operand): New predicate.
24130 (non_volatile_mem_operand): New predicate.
24131 * config/mips/mips.md (multimem): New type.
24132 (length): Differentiate between 17-bit and 18-bit branch offsets.
24133 (MOVEP1, MOVEP2): New mode iterator.
24134 (mov_<load>l): Use ZC constraint.
24135 (mov_<load>r): Likewise.
24136 (mov_<store>l): Likewise.
24137 (mov_<store>r): Likewise.
24138 (*branch_equality<mode>_inverted): Add microMIPS support.
24139 (*branch_equality<mode>): Likewise.
24140 (*jump_absolute): Likewise.
24141 (indirect_jump_<mode>): Likewise.
24142 (tablejump_<mode>): Likewise.
24143 (<optab>_internal): Likewise.
24144 (sibcall_internal): Likewise.
24145 (sibcall_value_internal): Likewise.
24146 (prefetch): Use constraint ZD.
24147 * config/mips/mips.opt (minterlink-compressed): New option.
24148 (minterlink-mips16): Now an alias for minterlink-compressed.
24149 (mmicromips): New option.
24150 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24151 (compare_and_swap_12): Likewise.
24152 (sync_add<mode>): Likewise.
24153 (sync_<optab>_12): Likewise.
24154 (sync_old_<optab>_12): Likewise.
24155 (sync_new_<optab>_12): Likewise.
24156 (sync_nand_12): Likewise.
24157 (sync_old_nand_12): Likewise.
24158 (sync_new_nand_12): Likewise.
24159 (sync_sub<mode>): Likewise.
24160 (sync_old_add<mode>): Likewise.
24161 (sync_old_sub<mode>): Likewise.
24162 (sync_new_add<mode>): Likewise.
24163 (sync_new_sub<mode>): Likewise.
24164 (sync_<optab><mode>): Likewise.
24165 (sync_old_<optab><mode>): Likewise.
24166 (sync_new_<optab><mode>): Likewise.
24167 (sync_nand<mode>): Likewise.
24168 (sync_old_nand<mode>): Likewise.
24169 (sync_new_nand<mode>): Likewise.
24170 (sync_lock_test_and_set<mode>): Likewise.
24171 (test_and_set_12): Likewise.
24172 (atomic_compare_and_swap<mode>): Likewise.
24173 (atomic_exchange<mode>_llsc): Likewise.
24174 (atomic_fetch_add<mode>_llsc): Likewise.
24175 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24176 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24177 (umips_save_restore_pattern_p): Likewise.
24178 (umips_load_store_pair_p): Likewise.
24179 (umips_output_load_store_pair): Likewise.
24180 (umips_movep_target_p): Likewise.
24181 (umips_12bit_offset_address_p): Likewise.
24182 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24183 (mips_base_mips16): Rename this...
24184 (mips_base_compression_flags): ...to this. Update all uses.
24185 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24186 (mips_mips16_decl_p): Delete.
24187 (mips_nomips16_decl_p): Delete.
24188 (mips_get_compress_on_flags): New function.
24189 (mips_get_compress_off_flags): New function.
24190 (mips_get_compress_mode): New function.
24191 (mips_get_compress_on_name): New function.
24192 (mips_get_compress_off_name): New function.
24193 (mips_insert_attributes): Support multiple compression types.
24194 (mips_merge_decl_attributes): Likewise.
24195 (umips_12bit_offset_address_p): New function.
24196 (mips_start_function_definition): Emit .set micromips directive.
24197 (mips_call_may_need_jalx_p): New function.
24198 (mips_function_ok_for_sibcall): Add microMIPS support.
24199 (mips_print_operand_punctuation): Support short delay slots and
24201 (umips_swm_mask, umips_swm_encoding): New.
24202 (umips_build_save_restore): New function.
24203 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24204 (was_mips16_p): Remove.
24205 (old_compression_mode): New.
24206 (mips_set_compression_mode): New function.
24207 (mips_set_current_function): Add microMIPS support.
24208 (mips_option_override): Likewise.
24209 (umips_save_restore_pattern_p): New function.
24210 (umips_output_save_restore): New function.
24211 (umips_load_store_pair_p_1): New function.
24212 (umips_load_store_pair_p): New function.
24213 (umips_output_load_store_pair_1): New function.
24214 (umips_output_load_store_pair): New function.
24215 (umips_movep_target_p) New function.
24216 (mips_prepare_pch_save): Add microMIPS support.
24217 * config/mips/mips.h (TARGET_COMPRESSION): New.
24218 (TARGET_CPU_CPP_BUILTINS): Update macro
24219 to use new compression flags and to support microMIPS.
24220 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24221 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24222 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24223 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24224 (ASM_SPEC): Support mmicromips and mno-micromips.
24225 (M16STORE_REG_P): New macro.
24226 (MIPS_CALL): Support TARGET_MICROMIPS.
24227 (MICROMIPS_J): New macro.
24228 (mips_base_mips16): Rename this...
24229 (mips_base_compression_flags): ...to this.
24230 (UMIPS_12BIT_OFFSET_P): New macro.
24231 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24232 (MULTILIB_DIRNAMES): Likewise.
24234 2013-02-25 Tom de Vries <tom@codesourcery.com>
24236 PR rtl-optimization/56131
24237 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
24238 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
24239 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
24241 2013-02-25 Tobias Burnus <burnus@net-b.de>
24243 * doc/invoke.texi (-fsanitize=): Move from optimization
24244 to debugging options.
24246 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
24248 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
24250 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
24251 Alexander Monakov <amonakov@ispras.ru>
24253 PR middle-end/56077
24254 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
24255 flush pending lists also on non-jumps. Adjust comment.
24257 2013-02-24 Catherine Moore <clm@codesourcery.com>
24258 Maciej W. Rozycki <macro@codesourcery.com>
24259 Tom de Vries <tom@codesourcery.com>
24260 Nathan Sidwell <nathan@codesourcery.com>
24261 Iain Sandoe <iain@codesourcery.com>
24262 Nathan Froyd <froydnj@codesourcery.com>
24263 Chao-ying Fu <fu@mips.com>
24265 * doc/extend.texi: (micromips, nomicromips, nocompression):
24266 Document new function attributes.
24267 * doc/invoke.texi (minterlink-compressed, mmicromips,
24268 m14k, m14ke, m14kec): Document new options.
24269 (minterlink-mips16): Update documentation.
24270 * doc/md.texi (ZC, ZD): Document new constraints.
24271 * configure.ac (gcc_cv_as_micromips): Check if linker
24272 supports the .set micromips directive.
24273 * configure: Regenerate.
24274 * config.in: Regenerate.
24275 * config/mips/mips-tables.opt: Regenerate.
24276 * config/mips/micromips.md: New file.
24277 * constraints.md (ZC, AD): New constraints.
24278 * config/mips/predicates.md (movep_src_register): New predicate.
24279 (movep_src_operand): New predicate.
24280 (non_volatile_mem_operand): New predicate.
24281 * config/mips/mips.md (multimem): New type.
24282 (length): Differentiate between 17-bit and 18-bit branch offsets.
24283 (MOVEP1, MOVEP2): New mode iterator.
24284 (mov_<load>l): Use ZC constraint.
24285 (mov_<load>r): Likewise.
24286 (mov_<store>l): Likewise.
24287 (mov_<store>r): Likewise.
24288 (*branch_equality<mode>_inverted): Add microMIPS support.
24289 (*branch_equality<mode>): Likewise.
24290 (*jump_absolute): Likewise.
24291 (indirect_jump_<mode>): Likewise.
24292 (tablejump_<mode>): Likewise.
24293 (<optab>_internal): Likewise.
24294 (sibcall_internal): Likewise.
24295 (sibcall_value_internal): Likewise.
24296 (prefetch): Use constraint ZD.
24297 * config/mips/mips.opt (minterlink-compressed): New option.
24298 (minterlink-mips16): Now an alias for minterlink-compressed.
24299 (mmicromips): New option.
24300 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24301 (compare_and_swap_12): Likewise.
24302 (sync_add<mode>): Likewise.
24303 (sync_<optab>_12): Likewise.
24304 (sync_old_<optab>_12): Likewise.
24305 (sync_new_<optab>_12): Likewise.
24306 (sync_nand_12): Likewise.
24307 (sync_old_nand_12): Likewise.
24308 (sync_new_nand_12): Likewise.
24309 (sync_sub<mode>): Likewise.
24310 (sync_old_add<mode>): Likewise.
24311 (sync_old_sub<mode>): Likewise.
24312 (sync_new_add<mode>): Likewise.
24313 (sync_new_sub<mode>): Likewise.
24314 (sync_<optab><mode>): Likewise.
24315 (sync_old_<optab><mode>): Likewise.
24316 (sync_new_<optab><mode>): Likewise.
24317 (sync_nand<mode>): Likewise.
24318 (sync_old_nand<mode>): Likewise.
24319 (sync_new_nand<mode>): Likewise.
24320 (sync_lock_test_and_set<mode>): Likewise.
24321 (test_and_set_12): Likewise.
24322 (atomic_compare_and_swap<mode>): Likewise.
24323 (atomic_exchange<mode>_llsc): Likewise.
24324 (atomic_fetch_add<mode>_llsc): Likewise.
24325 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24326 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24327 (umips_save_restore_pattern_p): Likewise.
24328 (umips_load_store_pair_p): Likewise.
24329 (umips_output_load_store_pair): Likewise.
24330 (umips_movep_target_p): Likewise.
24331 (umips_12bit_offset_address_p): Likewise.
24332 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24333 (mips_base_mips16): Rename this...
24334 (mips_base_compression_flags): ...to this. Update all uses.
24335 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24336 (mips_mips16_decl_p): Delete.
24337 (mips_nomips16_decl_p): Delete.
24338 (mips_get_compress_on_flags): New function.
24339 (mips_get_compress_off_flags): New function.
24340 (mips_get_compress_mode): New function.
24341 (mips_get_compress_on_name): New function.
24342 (mips_get_compress_off_name): New function.
24343 (mips_insert_attributes): Support multiple compression types.
24344 (mips_merge_decl_attributes): Likewise.
24345 (umips_12bit_offset_address_p): New function.
24346 (mips_start_function_definition): Emit .set micromips directive.
24347 (mips_call_may_need_jalx_p): New function.
24348 (mips_function_ok_for_sibcall): Add microMIPS support.
24349 (mips_print_operand_punctuation): Support short delay slots and
24351 (umips_swm_mask, umips_swm_encoding): New.
24352 (umips_build_save_restore): New function.
24353 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24354 (was_mips16_p): Remove.
24355 (old_compression_mode): New.
24356 (mips_set_compression_mode): New function.
24357 (mips_set_current_function): Add microMIPS support.
24358 (mips_option_override): Likewise.
24359 (umips_save_restore_pattern_p): New function.
24360 (umips_output_save_restore): New function.
24361 (umips_load_store_pair_p_1): New function.
24362 (umips_load_store_pair_p): New function.
24363 (umips_output_load_store_pair_1): New function.
24364 (umips_output_load_store_pair): New function.
24365 (umips_movep_target_p) New function.
24366 (mips_prepare_pch_save): Add microMIPS support.
24367 * config/mips/mips.h (TARGET_COMPRESSION): New.
24368 (TARGET_CPU_CPP_BUILTINS): Update macro
24369 to use new compression flags and to support microMIPS.
24370 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24371 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24372 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24373 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24374 (ASM_SPEC): Support mmicromips and mno-micromips.
24375 (M16STORE_REG_P): New macro.
24376 (MIPS_CALL): Support TARGET_MICROMIPS.
24377 (MICROMIPS_J): New macro.
24378 (mips_base_mips16): Rename this...
24379 (mips_base_compression_flags): ...to this.
24380 (UMIPS_12BIT_OFFSET_P): New macro.
24381 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24382 (MULTILIB_DIRNAMES): Likewise.
24384 2013-02-24 Jakub Jelinek <jakub@redhat.com>
24387 * target-globals.c (save_target_globals): For init_reg_sets and
24388 target_reinit remporarily set this_fn_optabs to this_target_optabs.
24390 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
24392 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
24393 * config/aarch64/t-aarch64
24394 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
24396 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
24398 PR inline-asm/56148
24399 * lra-constraints.c (process_alt_operands): Reload operand
24400 conflicting with earlier clobber only if no more other conflicting
24403 2013-02-22 Jakub Jelinek <jakub@redhat.com>
24406 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
24407 if not linking a shared library.
24409 2013-02-22 Seth LaForge <sethml@google.com>
24411 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
24413 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
24415 * config/arm/arm.md (split for extendsidi): Update condition.
24416 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
24417 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
24418 (qhs_zextenddi_cstr): Likewise.
24420 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24422 PR middle-end/56420
24423 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
24424 avoid signed wrapping.
24425 (expand_mult): Handle properly multiplication by
24426 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
24427 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
24428 in the compiler if coeff is HOST_WIDE_INT_MIN.
24429 (expand_divmod): Don't make ext_op1 static, change it's type to
24430 uhwi. Avoid undefined behavior in -INTVAL (op1).
24432 PR rtl-optimization/50339
24433 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
24435 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
24436 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
24437 into splitting_ashiftrt field.
24438 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
24440 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
24443 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
24445 PR middle-end/56108
24446 * trans-mem.c (execute_tm_mark): Do not expand transactions that
24447 are sure to go irrevocable.
24449 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
24451 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
24452 scalars are valid operands.
24454 2013-02-21 Martin Jambor <mjambor@suse.cz>
24456 PR tree-optimization/56310
24457 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
24458 only matching indices and non-negative final offsets.
24459 (intersect_aggregates_with_edge): Pass src_idx to
24460 agg_replacements_to_vector. Pass src_idx insstead of index to
24461 intersect_with_agg_replacements.
24463 2013-02-21 Martin Jambor <mjambor@suse.cz>
24465 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
24466 instead of hard-wired defaults.
24468 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
24470 * doc/invoke.texi (MIPS Options): Update documentation of the
24471 floating-point multiply-accumulate instruction restrictions.
24473 2013-02-21 Kostya Serebryany <kcc@google.com>
24475 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
24476 asan_shadow_offset on x86_64 linux.
24478 2013-02-21 Richard Biener <rguenther@suse.de>
24480 PR tree-optimization/56415
24482 2013-02-11 Richard Biener <rguenther@suse.de>
24484 PR tree-optimization/56273
24485 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24488 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24491 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
24494 PR inline-asm/56405
24495 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
24496 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
24498 2013-02-20 Jan Hubicka <jh@suse.cz>
24500 PR tree-optimization/56265
24501 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
24502 when target is referenced for first time.
24504 2013-02-20 Richard Biener <rguenther@suse.de>
24506 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
24507 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
24508 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
24509 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
24510 not return anything.
24511 (rename_ssa_copies): Do not remove unused locals.
24512 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
24513 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
24514 * passes.c (execute_function_todo): Do not schedule unused locals
24515 removal if cleanup_tree_cfg did something.
24516 * tree-ssa-live.c (remove_unused_locals): Dump statistics
24517 about the number of removed locals.
24519 2013-02-20 Richard Biener <rguenther@suse.de>
24521 PR tree-optimization/56398
24522 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
24524 2013-02-20 Martin Jambor <mjambor@suse.cz>
24526 PR tree-optimization/55334
24527 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
24528 restricted pointers to arrays.
24530 2013-02-20 Richard Biener <rguenther@suse.de>
24531 Jakub Jelinek <jakub@redhat.com>
24533 PR tree-optimization/56396
24534 * tree-ssa-ccp.c (n_const_val): New static variable.
24535 (get_value): Return NULL for SSA names we don't have a lattice
24537 (ccp_initialize): Initialize n_const_val.
24538 * tree-ssa-copy.c (n_copy_of): New static variable.
24539 (init_copy_prop): Initialize n_copy_of.
24540 (get_value): Return NULL_TREE for SSA names we don't have a
24543 2013-02-20 Martin Jambor <mjambor@suse.cz>
24545 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
24547 2013-02-20 Richard Biener <rguenther@suse.de>
24549 * genpreds.c (write_lookup_constraint): Do not compare first
24550 letter of the constraint again.
24552 2013-02-20 Richard Biener <rguenther@suse.de>
24554 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
24556 (get_use_iv_cost): Terminate hashtable walk when coming across
24559 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
24561 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
24562 reassociation for avx2 targets.
24564 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24566 * config/microblaze/microblaze.c: microblaze_has_clz = 0
24567 Add version check for v8.10.a to enable microblaze_has_clz
24568 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
24569 version and TARGET_PATTERN_COMPARE check
24570 * config/microblaze/microblaze.md: New clzsi2 instruction
24572 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24574 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
24575 function before branching.
24577 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24579 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
24581 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
24583 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24585 PR middle-end/55889
24586 * sel-sched.c: Include ira.h.
24587 (implicit_clobber_conflict_p): New function.
24588 (moveup_expr): Use it.
24589 * Makefile.in (sel-sched.o): Depend on ira.h.
24591 2013-02-19 Richard Biener <rguenther@suse.de>
24593 PR tree-optimization/56384
24594 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
24595 (vn_hash_type): Split out from ...
24596 (vn_hash_constant_with_type): ... here.
24597 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
24598 (vn_phi_eq): Compare types from vn_phi_s structure.
24599 (vn_phi_lookup): Populate vn_phi_s type.
24600 (vn_phi_insert): Likewise.
24602 2013-02-19 Jakub Jelinek <jakub@redhat.com>
24604 PR tree-optimization/56350
24605 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
24606 if haven't found reduction or nested cycle operand, rather than
24607 asserting we must find it.
24609 PR tree-optimization/56381
24610 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
24613 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
24614 Jakub Jelinek <jakub@redhat.com>
24617 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
24618 (swap_optab_enable): Same.
24619 (init_all_optabs): Use argument instead of global.
24620 * tree.h (struct tree_optimization_option): New field target_optabs.
24621 * expr.h (init_all_optabs): Add argument to prototype.
24622 (TREE_OPTIMIZATION_OPTABS): New.
24623 (save_optabs_if_changed): Protoize.
24624 * optabs.h: Declare this_fn_optabs.
24625 * optabs.c (save_optabs_if_changed): New.
24626 Declare this_fn_optabs.
24627 (init_optabs): Add argument to init_all_optabs() call.
24628 * function.c (invoke_set_current_function_hook): Handle per
24630 * function.h (struct function): New field optabs.
24631 * config/mips/mips.c (mips_set_mips16_mode): Handle when
24632 optimization_current_node has changed.
24633 * target-globals.h (save_target_globals_default_opts): Protoize.
24634 * target-globals.c (save_target_globals_default_opts): New.
24636 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24639 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
24640 registers %fr12 and %fr12R as call used.
24643 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
24644 and HImode, require all displacements to be an integer multiple of
24646 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
24647 only allow QImode and HImode when reload is in progress and strict is
24648 true. Likewise for symbolic addresses. Use base14_operand to check
24649 displacements in REG+BASE addresses.
24651 2013-02-18 Richard Biener <rguenther@suse.de>
24653 PR tree-optimization/56366
24654 * tree-vect-loop.c (get_initial_def_for_induction): Properly
24655 handle sign-conversion of outer-loop initial induction value.
24657 2013-02-18 Richard Biener <rguenther@suse.de>
24659 PR middle-end/56349
24660 * cfghooks.c (merge_blocks): If we merge a latch into another
24661 block adjust references to it.
24662 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
24663 (verify_loop_structure): Verify that a recorded latch is in fact
24666 2013-02-18 Richard Biener <rguenther@suse.de>
24668 PR tree-optimization/56321
24669 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
24670 order SSA name release and virtual operand unlinking.
24672 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24674 * config/microblaze/microblaze.md (save_stack_block): Define.
24675 (restore_stack_block): Likewise.
24677 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24679 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
24680 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
24681 * config/microblaze/microblaze.c (microblaze_option_override):
24682 Bail out early for PIC modes when target does not support PIC.
24684 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24686 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
24687 Replace with a microblaze version.
24688 (microblaze_trampoline_init): Adapt for microblaze.
24689 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
24692 2013-02-16 Jakub Jelinek <jakub@redhat.com>
24693 Dodji Seketeli <dodji@redhat.com>
24696 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
24697 (instrument_mem_region_access): Do not forget to always put
24698 instrumentation of the of 'base' and 'base + len' in a "if (len !=
24699 0) statement, even for cases where either 'base' or 'base + len'
24700 are not instrumented -- because they have been previously
24701 instrumented. Simplify the logic by putting all the statements
24702 instrument 'base + len' inside a sequence, and then insert that
24703 sequence right before the current insertion point. Then, to
24704 instrument 'base + len', just get an iterator on that statement.
24705 And do not forget to update the pointer to iterator the function
24706 received as argument.
24708 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
24710 PR rtl-optimization/56348
24711 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
24713 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
24715 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
24716 (clean_graph_dump_file): Pass base to start_graph_dump.
24718 2013-02-14 Richard Henderson <rth@redhat.com>
24721 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
24723 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
24725 * collect2-aix.h: Define F_LOADONLY.
24727 2013-02-14 Richard Biener <rguenther@suse.de>
24730 * varasm.c (output_constant_def_1): Get the decl representing
24731 the constant as argument.
24732 (output_constant_def): Wrap output_constant_def_1.
24733 (make_decl_rtl): Use output_constant_def_1 with the decl
24734 representing the constant.
24735 (build_constant_desc): Optionally re-use a decl already
24736 representing the constant.
24737 (tree_output_constant_def): Adjust.
24739 2013-02-14 Dodji Seketeli <dodji@redhat.com>
24742 * asan.c (instrument_builtin_call): Really put the length of the
24743 second source argument into src1_len.
24745 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24747 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
24748 argument. If it is false, don't create edge from then_bb to
24750 (insert_if_then_before_iter): Pass true to it.
24751 (build_check_stmt): Pass false to it.
24752 (transform_statements): Flush hash table only on extended basic
24753 block boundaries, rather than at the beginning of every bb.
24754 Don't flush hash table on nonfreeing_call_p calls.
24755 * tree-flow.h (nonfreeing_call_p): New prototype.
24756 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
24758 2013-02-13 David S. Miller <davem@davemloft.net>
24760 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
24762 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
24765 * ira.c (max_regno_before_ira): Move from ...
24767 (fix_reg_equiv_init): Use max_regno_before_ira instead of
24770 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24772 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
24774 2013-02-13 Richard Biener <rguenther@suse.de>
24777 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
24778 globals in MEM_REFs.
24780 2013-02-13 Richard Biener <rguenther@suse.de>
24782 * loop-init.c (loop_optimizer_init): Clear loop state when
24783 re-initializing preserved loops.
24784 * loop-unswitch.c (unswitch_single_loop): Return whether
24785 we unswitched the loop. Do not verify loop state here.
24786 (unswitch_loops): When we unswitched a loop discover new loops.
24788 2013-02-13 Kostya Serebryany <kcc@google.com>
24790 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
24792 * sanitizer.def: Rename __asan_init to __asan_init_v1.
24794 2013-02-12 Dodji Seketeli <dodji@redhat.com>
24796 Avoid instrumenting duplicated memory access in the same basic block
24797 * Makefile.in (asan.o): Add new dependency on hash-table.h
24798 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
24799 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
24800 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
24801 (free_mem_ref_resources, has_mem_ref_been_instrumented)
24802 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
24803 (get_mem_ref_of_assignment): New functions.
24804 (get_mem_refs_of_builtin_call): Extract from
24805 instrument_builtin_call and tweak a little bit to make it fit with
24807 (instrument_builtin_call): Use the new
24808 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
24809 of is_gimple_builtin_call.
24810 (instrument_derefs, instrument_mem_region_access): Insert the
24811 instrumented memory reference into the hash table.
24812 (maybe_instrument_assignment): Renamed instrument_assignment into
24813 this, and change it to advance the iterator when instrumentation
24814 actually happened and return true in that case. This makes it
24815 homogeneous with maybe_instrument_assignment, and thus give a
24816 chance to callers to be more 'regular'.
24817 (transform_statements): Clear the memory reference hash table
24818 whenever we enter a new BB, when we cross a function call, or when
24819 we are done transforming statements. Use
24820 maybe_instrument_assignment instead of instrumentation. No more
24821 need to special case maybe_instrument_assignment and advance the
24822 iterator after calling it; it's now handled just like
24823 maybe_instrument_call. Update comment.
24825 2013-02-13 Richard Biener <rguenther@suse.de>
24827 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
24828 Fix loop discovery code.
24830 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
24832 PR inline-asm/56148
24833 * lra-constraints.c (process_alt_operands): Match early clobber
24834 operand with itself. Check conflicts with earlyclobber only if
24835 the operand is not reloaded. Prefer to reload conflicting operand
24836 if earlyclobber and matching operands are the same.
24838 2013-02-12 Richard Biener <rguenther@suse.de>
24841 * lto-streamer-out.c (write_symbol): Do not output symbols
24842 for hard register variables.
24844 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
24847 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
24848 (umulsidi3_insn, mulsidi3_insn): New insns.
24850 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
24852 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
24853 (struct tune_params): Add vec_costs field.
24854 * config/arm/arm.c (arm_builtin_vectorization_cost)
24855 (arm_add_stmt_cost): New functions.
24856 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
24857 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
24858 (arm_default_vec_cost): New struct of type cpu_vec_costs.
24859 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
24860 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
24861 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
24862 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
24864 2013-02-12 Richard Biener <rguenther@suse.de>
24867 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
24868 decls again if possible.
24870 2013-02-12 Richard Biener <rguenther@suse.de>
24872 PR middle-end/56288
24873 * tree-ssa.c (verify_ssa_name): Fix check, move
24874 SSA_NAME_IN_FREE_LIST check up.
24876 2013-02-12 Jakub Jelinek <jakub@redhat.com>
24877 Steven Bosscher <steven@gcc.gnu.org>
24879 PR rtl-optimization/56151
24880 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
24881 equal to op0 or op1, and last_insn pattern is CODE operation
24882 with MEM dest and one of the operands matches that MEM.
24884 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
24886 * doc/extend.texi: Document Function Multiversioning and "default"
24887 parameter string to target attribute.
24888 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
24889 target attribute parameter is "default".
24890 (ix86_compare_version_priority): Remove checks for target attribute.
24891 (ix86_mangle_function_version_assembler_name): Change error to sorry.
24892 Remove check for target attribute equal to NULL. Add assert.
24893 (ix86_generate_version_dispatcher_body): Change error to sorry.
24895 2013-02-11 Iain Sandoe <iain@codesourcery.com>
24896 Jack Howarth <howarth@bromo.med.uc.edu>
24897 Patrick Marlier <patrick.marlier@gmail.com>
24900 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
24901 define ENDFILE_SPEC as TM_DESTRUCTOR.
24902 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
24904 2013-02-11 Alexander Potapenko <glider@google.com>
24905 Jack Howarth <howarth@bromo.med.uc.edu>
24906 Jakub Jelinek <jakub@redhat.com>
24909 * config/darwin.c (cdtor_record): Rename ctor_record.
24910 (sort_cdtor_records): Rename sort_ctor_records.
24911 (finalize_dtors): New routine to sort destructors by
24912 priority before use in assemble_integer.
24913 (machopic_asm_out_destructor): Use finalize_dtors if needed.
24915 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
24917 PR rtl-optimization/56275
24918 * simplify-rtx.c (avoid_constant_pool_reference): Check that
24919 offset is non-negative and less than cmode size before
24920 calling simplify_subreg.
24922 2013-02-11 Richard Biener <rguenther@suse.de>
24924 PR tree-optimization/56264
24925 * cfgloop.h (fix_loop_structure): Adjust prototype.
24926 * loop-init.c (fix_loop_structure): Return the number of
24927 newly discovered loops.
24928 * tree-cfgcleanup.c (repair_loop_structures): When new loops
24929 are discovered, do a full loop-closed SSA rewrite.
24931 2013-02-11 Richard Biener <rguenther@suse.de>
24933 PR tree-optimization/56273
24934 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24936 (check_array_ref): Fix missing newline in dumps.
24937 (search_for_addr_array): Likewise.
24939 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
24941 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
24943 2013-02-09 Jakub Jelinek <jakub@redhat.com>
24946 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
24948 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
24950 PR rtl-optimization/56246
24951 * lra-constraints.c (simplify_operand_subreg): Try to reuse
24953 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
24954 constraints are satisfied.
24956 2013-02-08 Jeff Law <law@redhat.com>
24959 * emit-rtl.c (reg_is_parm_p): New function.
24960 * regs.h (reg_is_parm_p): New prototype.
24961 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
24962 callee-clobbered registers.
24964 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
24967 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
24968 If there is no implicit builtin declaration, just return NULL.
24970 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24972 * config/i386/sse.md (FMAMODEM): New mode iterator.
24973 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
24974 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
24976 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24978 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
24979 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
24980 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
24982 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24984 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
24985 (microblaze*-*-elf): Likewise.
24986 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
24988 * config/microblaze/microblaze-c.c: Add builtin defines for
24989 _LITTLE_ENDIAN and _BIG_ENDIAN.
24990 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
24991 add to TARGET_DEFAULT flags.
24992 Expand ASM_SPEC and LINK_SPEC.
24993 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
24994 * config/microblaze/microblaze.md: Update extendsidi2 and
24995 movdi_internal instructions to use low-order / high-order reg
24997 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
24998 options and inversemask / mask of LITTLE_ENDIAN.
24999 * config/microblaze/t-microblaze: Expand multilib options to
25000 include mlittle-endian (le) and update exceptions patterns.
25002 2013-02-08 Jakub Jelinek <jakub@redhat.com>
25004 PR rtl-optimization/56195
25005 * lra-constraints.c (get_reload_reg): Don't reuse regs
25006 if they have smaller mode than requested, if they have
25007 wider mode than requested, try to return a SUBREG.
25009 PR tree-optimization/56250
25010 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
25011 if type is unsigned and code isn't MULT_EXPR.
25013 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
25015 PR tree-optimization/56064
25016 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
25017 bits according to mode.
25018 * fixed-value.h (fixed_from_double_int)
25019 (const_fixed_from_double_int): Adjust comments.
25021 2013-02-08 Richard Biener <rguenther@suse.de>
25024 * lto-streamer.h (struct data_in): Remove current_file, current_line
25025 and current_col members.
25026 * lto-streamer-out.c (lto_output_location): Stream changed bits
25027 en-block for efficiency.
25028 * lto-streamer-in.c (clear_line_info): Remove.
25029 (lto_input_location): Cache current file, line and column
25030 globally via local statics. Read changed bits en-block.
25031 (input_function): Do not call clear_line_info.
25032 (lto_read_body): Likewise.
25033 (lto_input_toplevel_asms): Likewise.
25035 2013-02-08 Michael Matz <matz@suse.de>
25037 PR tree-optimization/52448
25038 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
25039 (nt_call_phase): New static.
25040 (add_or_mark_expr): Only mark accesses with newer phase than any
25042 (nonfreeing_call_p): New.
25043 (nt_init_block): Update nt_call_phase, mark blocks as visited.
25044 (nt_fini_block): Keep blocks marked as visited.
25045 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
25047 2013-02-08 Richard Biener <rguenther@suse.de>
25049 * ira.c (ira): Free broken dominator information.
25051 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
25053 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
25055 2013-02-08 Marek Polacek <polacek@redhat.com>
25057 * cfgloop.c (verify_loop_structure): Add more checking of headers.
25059 2013-02-08 Richard Biener <rguenther@suse.de>
25061 PR middle-end/56181
25062 * cfgloop.h (flow_loops_find): Adjust.
25063 (bb_loop_header_p): Declare.
25064 * cfgloop.c (bb_loop_header_p): New function split out from ...
25065 (flow_loops_find): ... here. Adjust function signature,
25066 support incremental loop structure update.
25067 (verify_loop_structure): Cleanup. Verify a loop is a loop.
25068 * cfgloopmanip.c (fix_loop_structure): Move ...
25069 * loop-init.c (fix_loop_structure): ... here.
25070 (apply_loop_flags): Split out from ...
25071 (loop_optimizer_init): ... here.
25072 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
25073 in incremental mode, only remove dead loops here.
25075 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
25078 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
25079 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
25080 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
25081 (*round<mode>3.libgcc): New insns for fixed-modes.
25082 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
25083 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
25084 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
25085 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
25086 implementations. Define to __builtin_avr_absFX,
25087 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
25088 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
25089 __builtin_avr_countlsFX, respectively.
25090 * config/avr/avr-c.c (target.h): Include it.
25091 (enum avr_builtin_id): New enum.
25092 (avr_resolve_overloaded_builtin): New static function.
25093 (avr_register_target_pragmas): Use it to set
25094 targetm.resolve_overloaded_builtin.
25095 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
25096 tree nodes used by DEF_BUILTIN.
25097 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
25098 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
25099 <AVR_BUILTIN_xxBITS>: Same.
25101 2013-02-08 Richard Biener <rguenther@suse.de>
25103 * cfgloop.c (verify_loop_structure): Properly handle
25104 a loop exiting to another loop header.
25105 * ira-int.h (ira_loops): Remove.
25106 * ira.c (ira_loops): Remove.
25107 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
25108 (do_reload): Use loop_optimizer_finalize.
25109 * ira-build.c (create_loop_tree_nodes): Use get_loops and
25110 number_of_loops to access the loop tree.
25111 (more_one_region_p): Likewise.
25112 (finish_loop_tree_nodes): Likewise.
25113 (rebuild_regno_allocno_maps): Likewise.
25114 (mark_loops_for_removal): Likewise.
25115 (mark_all_loops_for_removal): Likewise.
25116 (remove_unnecessary_regions): Likewise.
25117 (ira_build): Likewise.
25118 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
25120 2013-02-08 Richard Biener <rguenther@suse.de>
25122 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
25123 * ipa-pure-const.c (analyze_function): Avoid calling
25124 mark_irreducible_loops twice.
25125 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
25127 2013-02-07 David S. Miller <davem@davemloft.net>
25129 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
25131 * var-tracking.c (vt_add_function_parameter): Test the presence of
25132 HAVE_window_save properly and do not remap argument registers when
25133 we have a leaf function.
25135 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
25138 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
25140 * config/i386/i386.c (ix86_print_operand): Ditto.
25142 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
25144 * lra-constraints.c (process_alt_operands): Fix recently added comment.
25146 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
25148 PR rtl-optimization/56225
25149 * lra-constraints.c (process_alt_operands): Check that reload hard
25150 reg can hold value for strict_low_part.
25152 2013-02-07 Jakub Jelinek <jakub@redhat.com>
25155 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
25156 dwarf2out_end_function.
25157 (in_first_function_p, maybe_at_text_label_p,
25158 first_loclabel_num_not_at_text_label): New variables.
25159 (dwarf2out_var_location): In the first function find out
25160 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
25161 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
25164 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
25166 PR rtl-optimization/56178
25167 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
25168 SUBREG of a register. Tidy up related block of code.
25169 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
25170 note if the source is a register or a SUBREG of a register.
25172 2013-02-07 Jakub Jelinek <jakub@redhat.com>
25175 * config/rs6000/rs6000.md (ptrm): New mode attr.
25176 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
25177 call_value_indirect_aix<pttrsize>,
25178 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
25181 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25183 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
25184 if -bnortl. Convert to strcmp and strncmp.
25186 2013-02-07 Alan Modra <amodra@gmail.com>
25189 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
25190 addresses won't wrap when offsetting.
25191 (rs6000_secondary_reload): Provide secondary reloads needed for
25192 wrapping LO_SUM addresses.
25194 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
25196 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
25197 MACH, just __MACH__.
25199 2013-02-06 Richard Biener <rguenther@suse.de>
25201 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
25202 instead of calling fix_loop_structure.
25204 2013-02-06 Jakub Jelinek <jakub@redhat.com>
25206 PR middle-end/56217
25207 * omp-low.c (use_pointer_for_field): Return false if
25208 lower_send_shared_vars doesn't generate any copy-out code.
25210 2013-02-06 Tom de Vries <tom@codesourcery.com>
25212 PR rtl-optimization/56131
25213 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
25214 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
25215 of the label is NULL. Add comment.
25217 2013-02-05 Jakub Jelinek <jakub@redhat.com>
25219 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
25222 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
25223 (STATIC_LIBTSAN_LIBS): Likewise.
25224 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
25225 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
25226 is defined, don't add anything else beyond that.
25227 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
25228 (LINK_COMMAND_SPEC): Use them.
25230 PR tree-optimization/56205
25231 * tree-stdarg.c (check_all_va_list_escapes): Return true if
25232 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
25233 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
25235 2013-02-05 Richard Biener <rguenther@suse.de>
25237 PR tree-optimization/53342
25238 PR tree-optimization/53185
25239 * tree-vectorizer.h (vect_check_strided_load): Remove.
25240 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
25241 not disallow peeling for vectorized strided loads.
25242 (vect_check_strided_load): Make static and simplify.
25243 (vect_analyze_data_refs): Adjust.
25244 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
25245 correctly when vectorizing strided loads.
25247 2013-02-05 Richard Biener <rguenther@suse.de>
25249 * doc/install.texi: Refer to ISL, not PPL.
25251 2013-02-05 Jan Hubicka <jh@suse.cz>
25253 PR tree-optimization/55789
25254 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
25256 2013-02-05 Jan Hubicka <jh@suse.cz>
25258 PR tree-optimization/55789
25259 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
25260 the dead call anyway.
25262 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
25265 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
25267 2013-02-04 Alexander Potapenko <glider@google.com>
25268 Jack Howarth <howarth@bromo.med.uc.edu>
25269 Jakub Jelinek <jakub@redhat.com>
25272 * config/darwin.c (sort_ctor_records): Stabilized qsort
25273 on constructor priority by using original position.
25274 (finalize_ctors): New routine to sort constructors by
25275 priority before use in assemble_integer.
25276 (machopic_asm_out_constructor): Use finalize_ctors if needed.
25278 2013-02-04 Jakub Jelinek <jakub@redhat.com>
25281 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
25282 about visibility on artificial decls.
25283 * config/sol2.c (solaris_assemble_visibility): Likewise.
25285 2013-02-04 Kai Tietz <ktietz@redhat.com>
25288 * config/i386/i386.c (function_value_ms_64): Add additional valtype
25289 argument and improve checking of return-argument types for 16-byte
25291 (ix86_function_value_1): Add additional valtype argument on call
25292 of function_value_64.
25293 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
25294 handling infunction_value_64 function.
25296 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
25298 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
25300 2013-02-04 Richard Biener <rguenther@suse.de>
25302 PR tree-optimization/56188
25303 * tree-ssa-structalias.c (label_visit): Consider case with
25304 initially non-empty points-to set.
25305 (perform_var_substitution): Dump node mapping and clean up.
25307 2013-02-04 Richard Guenther <rguenther@suse.de>
25310 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
25311 node prevail as last resort.
25312 (lto_symtab_merge_decls): Remove guard on LTRANS here.
25313 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
25315 2013-02-04 Richard Biener <rguenther@suse.de>
25317 PR tree-optimization/56113
25318 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
25320 (equiv_class_lookup_or_add): ... this.
25321 (label_visit): Adjust and fix error in previous patch.
25322 (perform_var_substitution): Adjust.
25324 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
25326 * config/sh/divtab.c: Fix formatting and comments throughout the file.
25327 * config/sh/sh4-300.md: Likewise.
25328 * config/sh/sh4a.md: Likewise.
25329 * config/sh/constraints.md: Likewise.
25330 * config/sh/sh.md: Likewise.
25331 * config/sh/netbsd-elf.h: Likewise.
25332 * config/sh/predicates.md: Likewise.
25333 * config/sh/sh-protos.h: Likewise.
25334 * config/sh/ushmedia.h: Likewise.
25335 * config/sh/linux.h: Likewise.
25336 * config/sh/sh.c: Likewise.
25337 * config/sh/superh.h: Likewise.
25338 * config/sh/elf.h: Likewise.
25339 * config/sh/sh4.md: Likewise.
25340 * config/sh/sh.h: Likewise.
25342 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25344 * config/pa/constraints.md: Adjust unused letters. Change "T"
25345 constraint to match_test floating_point_store_memory_operand().
25346 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
25347 (base14_operand): New.
25348 (floating_point_store_memory_operand): New.
25349 (integer_store_memory_operand): Revise to use base14_operand and
25350 reg_plus_base_memory_operand.
25351 (move_dest_operand): Allow symbolic_memory_operands.
25352 (symbolic_memory_operand): Check for LO_SOM.
25353 (symbolic_operand): Change default case to break.
25354 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
25355 CONST_DOUBLE values to be reloaded by putting them into memory when
25356 the destination is a floating point register.
25357 (movdf): Remove code to handle CONST_DOUBLE.
25359 (reload_indf_r1): New.
25360 (reload_insf_r1): New.
25361 Consistently use "Q" and "T" constraints with integer and floating
25362 point move instructions, respectively.
25363 (movdi): Remove FAIL.
25364 Change predicate for source operand unamed DImode move from
25365 general_operand to move_src_operand.
25366 (umulsidi3): Change predicate for destination operand to
25368 Likewise for similar unamed patterns.
25369 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
25370 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
25371 (hppa_legitimize_address): Simplify mask calculation.
25372 (pa_emit_move_sequence): Revised handling of secondary reloads from
25373 REG+D addresses for floating point loads and stores. Directly handle
25374 loading CONST0_RTX (mode) to a floating point register.
25375 (pa_secondary_reload): Handle reloading DF and SFmode constant values
25376 to floating point registers. Don't restrict secondary reloads to
25377 floating point registers to integer modes. Revise some comments and
25379 (TARGET_LEGITIMATE_ADDRESS_P): Define.
25380 (pa_legitimate_address_p): New.
25381 (pa_legitimize_reload_address): New.
25382 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
25383 (STRICT_REG_OK_FOR_BASE_P): New.
25384 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
25385 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
25387 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
25388 Andrew Dixie <andrewd@gentrack.com>
25390 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
25393 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
25395 * expmed.c (extract_bit_field_1): Pass the full width of the
25396 structure to get_best_reg_extraction_insn.
25398 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25401 * configure.ac (use_cxa_atexit): Add AIX.
25402 * configure: Regenerate.
25404 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
25406 2013-02-01 Jakub Jelinek <jakub@redhat.com>
25409 * final.c (need_profile_function): New variable.
25410 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
25411 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
25412 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
25413 notes, targetm.asm_out.function_prologue doesn't emit anything,
25414 HAVE_prologue and profiler should be emitted before prologue,
25415 set need_profile_function instead of emitting it.
25416 (final_scan_insn): If need_profile_function, emit
25417 profile_function on the first NOTE_INSN_BASIC_BLOCK or
25418 NOTE_INSN_FUNCTION_BEG note.
25420 2013-02-01 Richard Henderson <rth@redhat.com>
25422 * config/rs6000/rs6000.md (smulditi3): New.
25425 * config/alpha/alpha.md (umulditi3): New.
25427 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25429 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
25430 (ASM_OUTPUT_ALIGNED_LOCAL): New.
25432 2013-02-01 Richard Biener <rguenther@suse.de>
25434 PR tree-optimization/56113
25435 * tree-ssa-structalias.c (label_visit): Reduce work for
25436 single-predecessor nodes.
25438 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
25440 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
25441 range isn't testing for zero.
25443 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
25445 PR middle-end/56113
25446 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
25448 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
25449 Nick Clifton <nickc@redhat.com>
25451 * config/v850/constraints.md (Q): Define as a memory constraint.
25452 * config/v850/predicates.md (label_ref_operand): New predicate.
25453 (e3v5_shift_operand): New predicate.
25454 (ior_operator): New predicate.
25455 * config/v850/t-v850: Add e3v5 multilib.
25456 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
25457 (v850_gen_movdi): Prototype.
25458 * config/v850/v850.c: Add support for e3v5 architecture.
25459 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
25461 (construct_save_jarl): Add e3v5 long JARL support.
25462 (v850_adjust_insn_length): New function. Adjust length of call
25463 insns when using e3v5 instructions.
25464 (v850_gen_movdi): New function: Generate instructions to move a
25466 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
25467 (CPP_SPEC): Define __v850e3v5__ as appropriate.
25468 (TARGET_USE_FPU): Enable for e3v5.
25469 (CONST_OK_FOR_W): New macro.
25470 (ADJUST_INSN_LENGTH): Define.
25471 * config/v850/v850.md (UNSPEC_LOOP): Define.
25472 (attr cpu): Add v850e3v5.
25473 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
25474 (movdi): New pattern.
25475 (movdi_internal): New pattern.
25476 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
25477 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
25478 (cstoresf4): Likewise.
25479 (cstoredf4): Likewise.
25480 (insv): New pattern.
25481 (rotlso3_a): New pattern.
25482 (rotlsi3_b): New pattern
25483 (rotlsi3_v850e3v5): New pattern.
25484 (doloop_begin): New pattern.
25485 (fix_loop_counter): New pattern.
25486 (doloop_end): New pattern.
25487 (branch_normal): Add e3v5 long branch support.
25488 (branch_invert): Likewise.
25489 (branch_z_normal): Likewise.
25490 (branch_z_invert): Likewise.
25491 (branch_nz_normal): Likewise.
25492 (branch_nz_invert): Likewise.
25493 (call_internal_short): Add e3v5 register-indirect JARL support.
25494 (call_internal_long): Likewise.
25495 (call_value_internal_short): Likewise.
25496 (call_value_internal_long): Likewise.
25497 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
25498 (mloop): New option.
25499 * config.gcc: Add support for configuring v840e3v5 target.
25500 * doc/invoke.texi: Document new v850 specific command line options.
25502 2013-01-31 Paul Koning <ni1d@arrl.net>
25506 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
25507 children if parent is a class.
25508 (prune_unused_types_prune): Don't add DW_AT_declaration.
25510 2013-01-31 Richard Biener <rguenther@suse.de>
25512 PR tree-optimization/56157
25513 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
25514 match up operand with SLP child.
25516 2013-01-31 Jason Merrill <jason@redhat.com>
25519 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
25520 parameters the first time.
25521 (gen_scheduled_generic_parms_dies): Check completeness here.
25523 2013-01-31 Richard Biener <rguenther@suse.de>
25525 PR middle-end/53073
25526 * common.opt (faggressive-loop-optimizations): New flag,
25527 enabled by default.
25528 * doc/invoke.texi (faggressive-loop-optimizations): Document.
25529 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
25530 infer_loop_bounds_from_undefined by it.
25532 2013-01-31 Richard Biener <rguenther@suse.de>
25534 PR tree-optimization/56150
25535 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
25536 visit virtual operands.
25537 (find_uses_to_rename_bb): Likewise.
25539 2013-01-31 Richard Biener <rguenther@suse.de>
25541 PR tree-optimization/56150
25542 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
25543 mixed store non-store stmts.
25545 2013-01-30 Jakub Jelinek <jakub@redhat.com>
25548 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
25549 LIBASAN_EARLY_SPEC is defined.
25550 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
25551 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
25553 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
25556 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
25557 invalid args instead of ICEing on it.
25558 (ix86_valid_target_attribute_tree): Return error_mark_node if
25559 ix86_valid_target_attribute_inner_p failed.
25560 (ix86_valid_target_attribute_p): Return false only if
25561 ix86_valid_target_attribute_tree returned error_mark_node. Allow
25562 target("default") attribute.
25563 (sorted_attr_string): Change argument from const char * to tree,
25564 merge in all target attribute arguments rather than just one.
25565 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
25566 instead of free. Avoid using strcat.
25567 (ix86_mangle_function_version_assembler_name): Mangle
25568 target("default") as if no target attribute is present. Adjust
25569 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
25570 instead of xmalloc and XDELETEVEC instead of free.
25571 (ix86_function_versions): Don't return true if one of the decls
25572 doesn't have target attribute. If they don't and one of the decls
25573 is DECL_FUNCTION_VERSIONED, report an error. Adjust
25574 sorted_attr_string caller. Use XDELETEVEC instead of free.
25575 (ix86_supports_function_versions): Remove.
25576 (make_name): Fix up formatting.
25577 (make_dispatcher_decl): Remove resolver_name and its initialization.
25578 Avoid leaking memory.
25579 (is_function_default_version): Return true if there is
25580 target("default") attribute rather than no target attribute at all.
25581 (make_resolver_func): Avoid leaking memory.
25582 (ix86_generate_version_dispatcher_body): Likewise.
25583 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
25584 * target.def (supports_function_versions): Remove.
25585 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
25586 * doc/tm.texi: Regenerated.
25588 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
25590 PR rtl-optimization/56144
25591 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
25592 for values with side effects.
25594 2013-01-30 Richard Biener <rguenther@suse.de>
25596 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
25597 (sparseset_pop): Likewise.
25598 * cfganal.c (compute_idf): Likewise. Increase work-stack size
25599 to be able to use quick_push in the worker loop.
25601 2013-01-30 Marek Polacek <polacek@redhat.com>
25603 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
25605 2013-01-30 Richard Biener <rguenther@suse.de>
25608 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
25610 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
25612 PR tree-optimization/56064
25613 * fixed-value.c (fixed_from_double_int): New function.
25614 * fixed-value.h (fixed_from_double_int): New prototype.
25615 (const_fixed_from_double_int): New static inline function.
25616 * fold-const.c (native_interpret_fixed): New static function.
25617 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
25618 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
25619 (native_encode_fixed): New static function.
25620 (native_encode_expr) <FIXED_CST>: Use it.
25621 (native_interpret_int): Move double_int worker code to...
25622 * double-int.c (double_int::from_buffer): ...this new static method.
25623 * double-int.h (double_int::from_buffer): Prototype it.
25625 2013-01-30 Richard Biener <rguenther@suse.de>
25627 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
25628 New pointer-map and obstack.
25629 (init_alias_vars): Allocate pointer-map and obstack.
25630 (delete_points_to_sets): Free them.
25631 (find_what_var_points_to): Cache result.
25632 (find_what_p_points_to): Adjust for changed interface of
25633 find_what_var_points_to.
25634 (compute_points_to_sets): Likewise.
25635 (ipa_pta_execute): Likewise.
25637 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25639 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
25640 * configure: Regenerate.
25641 * config.in: Regenerate.
25642 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
25643 #nobits/#progbits if supported.
25645 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
25648 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
25649 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
25650 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
25652 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25654 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
25655 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
25657 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25659 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
25660 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
25662 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25664 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
25666 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
25667 * config/arm/cortex-a7.md: New bypasses using
25668 arm_mac_accumulator_is_result.
25670 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25672 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
25673 (cortex_a7_neon_mla): Likewise.
25674 (cortex_a7_fpfmad): New reservation.
25675 (cortex_a7_fpmacs): Use ffmas and update required units.
25676 (cortex_a7_fpmuld): Update required units and latency.
25677 (cortex_a7_fpmacd): Likewise.
25678 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
25679 (cortex_a7_neon). Likewise.
25680 (bypass) Update participating units.
25682 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25684 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
25685 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
25687 * config/arm/vfp11.md (vfp_farith): Use ffmas.
25688 (vfp_fmul): Use ffmad.
25689 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
25690 (cortex_r4_fmacd): Use ffmad.
25691 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
25692 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
25693 (cortex_a9_fmacd): Use ffmad.
25694 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
25695 (cortex_a8_vfp_macd): Use ffmad.
25696 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
25697 (cortex_a5_fpmacd): Use ffmad.
25698 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
25699 (cortex_a15_vfp_macd): Use ffmad.
25700 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
25702 2013-01-29 Jason Merrill <jason@redhat.com>
25705 * varasm.c (default_assemble_visibility): Don't warn about
25706 visibility on artificial decls.
25708 2013-01-29 Richard Biener <rguenther@suse.de>
25710 PR tree-optimization/56113
25711 * tree-ssa-structalias.c (equiv_class_lookup): Also return
25713 (label_visit): Free duplicate bitmaps and record the leader instead.
25714 (perform_var_substitution): Adjust.
25716 2013-01-29 Richard Biener <rguenther@suse.de>
25718 PR tree-optimization/55270
25719 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
25720 the CFG, schedule loops for fixup.
25722 2013-01-29 Nick Clifton <nickc@redhat.com>
25724 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
25727 2013-01-28 Leif Ekblad <leif@rdos.net>
25729 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
25730 * config/i386/i386.h (TARGET_RDOS): New macro.
25731 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
25732 * config/i386/i386.c (ix86_option_override_internal): For 64bit
25733 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
25734 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
25735 DEFAULT_LARGE_SECTION_THRESHOLD.
25736 * config/i386/i386.md (R14_REG, R15_REG): New constants.
25737 * config/i386/rdos.h: New file.
25738 * config/i386/rdos64.h: New file.
25740 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
25743 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
25746 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25748 PR rtl-optimization/56117
25749 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
25750 call cselib_lookup_from_insn on the MEM before calling
25751 add_insn_mem_dependence.
25753 2013-01-28 Richard Biener <rguenther@suse.de>
25755 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
25756 to a stmt that didn't have one.
25757 (copy_phis_for_bb): Likewise for PHI arguments.
25758 (copy_debug_stmt): Likewise for debug stmts.
25760 2013-01-28 Richard Biener <rguenther@suse.de>
25762 PR tree-optimization/56034
25763 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
25764 (partition_builtin_p): Adjust.
25765 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
25766 it is the last partition.
25767 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
25768 up the vertex for the definition.
25769 (classify_partition): Classify whether a partition is a
25770 PKIND_REDUCTION, thus has uses outside of the loop.
25771 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
25772 Merge all PKIND_REDUCTION partitions into the last partition.
25773 (tree_loop_distribution): Seed partitions from reductions as well.
25775 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25777 PR tree-optimization/56125
25778 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
25779 pow(x,c) into sqrt(x) * powi(x, n/2) or
25780 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
25781 optimizing for size.
25782 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
25783 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
25786 PR tree-optimization/56094
25787 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
25788 to UNKNOWN_LOCATION while gimplifying expr.
25790 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
25793 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
25794 operand 0 in movabs insn template for -masm=intel asm alternative.
25795 (*movabs<mode>_2): Ditto for operand 1.
25797 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
25800 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
25803 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25805 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
25808 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25810 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
25811 Undef to avoid warning.
25813 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25815 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
25816 * configure: Regenerate.
25818 2013-01-25 Jakub Jelinek <jakub@redhat.com>
25820 PR tree-optimization/56098
25821 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
25822 for stmts with volatile ops.
25823 (cond_store_replacement): Don't optimize if assign has volatile ops.
25824 (cond_if_else_store_replacement_1): Don't optimize if either
25825 then_assign or else_assign have volatile ops.
25826 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
25829 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25831 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
25833 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25835 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
25836 missing ':' in asm example.
25838 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
25840 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
25841 entries into lane and laneq entries.
25842 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
25843 Remove AdvSIMD scalar modes.
25844 (aarch64_sq<r>dmulh_laneq<mode>): New.
25845 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
25847 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
25848 builtin implementations to relfect changes in RTL in aarch64-simd.md.
25849 * config/aarch64/iterators.md (VCOND): New.
25852 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25855 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
25856 Add NULL LIBNAME argument to existing definitions.
25857 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
25858 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
25859 * config/avr/avr.c (DEF_BUILTIN): Same.
25860 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
25861 (avr_expand_builtin): Expand to a vanilla call if a libgcc
25862 implementation is available (DECL_ASSEMBLER_NAME is set).
25863 (avr_fold_absfx): New static function.
25864 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
25865 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
25866 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
25867 AVR_BUILTIN_ABSLLK.
25868 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
25869 (abshk, absk, abslk, absllk): Provide as static inline functions.
25871 2013-01-25 Marek Polacek <polacek@redhat.com>
25873 PR tree-optimization/56035
25874 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
25876 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
25878 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
25879 (*movtf_internal_rex64): Add (!o,C) alternative
25880 (*movxf_internal_rex64): Ditto.
25881 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
25883 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
25885 * doc/invoke.texi: fix typo.
25886 * doc/objc.texi: fix typo.
25888 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
25890 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
25891 for the first two alternatives.
25893 2013-01-24 Diego Novillo <dnovillo@google.com>
25895 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
25896 (ggc-zone.o): Remove.
25897 * configure.ac: Remove option --with-gc.
25898 * configure: Re-generate.
25899 * doc/install.texi: Remove documentation for --with-gc.
25900 * gengtype.c (write_enum_defn): Remove. Update all users.
25901 (write_Types_process_field): Remove generation of gt_e_* argument.
25902 (output_type_enum): Remove. Update all users.
25903 (write_enum_defn): Remove. Update all users.
25904 (enum alloc_zone): Remove. Update all users.
25905 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
25906 * ggc-common.c (ggc_splay_alloc): Remove first argument.
25907 Update all callers.
25908 (struct ptr_data): Remove field TYPE. Update all users.
25909 (gt_pch_note_object): Remove argument TYPE. Update all users.
25910 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
25912 * ggc-none.c (ggc_alloc_typed_stat): Remove.
25913 (struct alloc_zone): Remove.
25914 (ggc_internal_alloc_zone_stat): Remove.
25915 (ggc_internal_cleared_alloc_zone_stat): Remove.
25916 * ggc-page.c (ggc_alloc_typed_stat): Remove.
25917 (ggc_pch_count_object): Remove last argument. Update all users.
25918 (ggc_pch_alloc_object): Remove last argument. Update all users.
25919 (struct alloc_zone): Remove.
25920 * ggc-zone.c: Remove.
25921 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
25922 (struct alloc_zone): Remove.
25923 (ggc_alloc_typed_stat): Remove.
25924 (ggc_alloc_typed): Remove.
25925 (ggc_splay_alloc): Remove first argument.
25926 (rtl_zone): Remove. Update all users.
25927 (tree_zone): Remove. Update all users.
25928 (tree_id_zone): Remove. Update all users.
25929 (ggc_internal_zone_alloc_stat): Remove. Update all users.
25930 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
25931 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
25932 * tree-ssanames.c: Remove references to zone allocator in comments.
25934 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
25936 * config/avr/avr.c (avr_out_fract): Make register numbers that
25937 might be outside of source operand signed.
25939 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
25941 * config/i386/constraints.md (Yf): New constraint.
25942 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
25943 of f constraint to conditionaly disable x87 register preferences.
25944 (*movdf_internal): Ditto.
25945 (*movsf_internal): Ditto.
25947 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
25949 PR inline-asm/55934
25950 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
25951 that have operands with impossible constraints.
25952 Add a FIXME for a speed-up opportunity.
25953 * lra-constraints.c (process_alt_operands): Verify that a class
25954 selected from constraints on asms is valid for the operand mode.
25955 (curr_insn_transform): Remove incorrect comment.
25957 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
25959 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
25960 TOC operand is a valid symbol ref in the constant pool.
25962 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25964 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
25966 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
25969 * config/avr/stdfix.h: New file.
25970 * t-avr (stdfix-gcc.h): New rule to build it.
25971 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
25973 2013-01-23 Kostya Serebryany <kcc@google.com>
25975 * config/darwin.h: remove dependency on
25976 CoreFoundation (asan on Mac OS).
25978 2013-01-23 Jakub Jelinek <jakub@redhat.com>
25981 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
25982 instead of cmpdi_operand for first comparison operand.
25983 Don't assert that comparison operands aren't both constants.
25985 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25987 * doc/install.texi (Downloading the Source): Update references to
25988 downloading separate components.
25990 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25992 * doc/extend.texi (__int128): Improve grammar.
25994 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
25997 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
25998 alternative to (o,r).
25999 (*movdi_internal_rex64): Remove (!o,n) alternative.
26000 (DImode immediate->memory splitter): Remove.
26001 (DImode immediate->memory peephole2): Remove.
26002 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
26003 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
26004 alternative to (!o,*r).
26005 (*movtf_internal_sse): New pattern.
26006 (*movxf_internal_rex64): New pattern.
26007 (*movxf_internal): Disable for TARGET_64BIT.
26008 (*movdf_internal_rex64): Remove (!o,F) alternative.
26010 2013-01-22 Jakub Jelinek <jakub@redhat.com>
26012 PR middle-end/56074
26013 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
26014 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
26015 * tree-vect-loop-manip.c (find_loop_location): Also ignore
26016 stmt locations where LOCATION_LOCUS of the stmt location is
26017 UNKNOWN_LOCATION or BUILTINS_LOCATION.
26020 * config/i386/i386.md (UNSPEC_STOS): New.
26021 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
26022 *strsetqi_1): Add UNSPEC_STOS.
26024 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
26027 * doc/invoke.texi: Remove left over -Wsynth example.
26029 2013-01-21 Jakub Jelinek <jakub@redhat.com>
26031 PR tree-optimization/56051
26032 * fold-const.c (fold_binary_loc): Don't fold
26033 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
26034 a narrowing conversion, or widening conversion from signed
26037 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
26039 PR rtl-optimization/56023
26040 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
26041 dependent on debug instruction.
26043 2013-01-21 Martin Jambor <mjambor@suse.cz>
26045 PR middle-end/56022
26046 * function.c (allocate_struct_function): Call
26047 invoke_set_current_function_hook earlier.
26049 2013-01-21 Jakub Jelinek <jakub@redhat.com>
26051 * reload1.c (init_reload): Only initialize reload_obstack
26052 during the first call.
26054 2013-01-21 Marek Polacek <polacek@redhat.com>
26056 * cfgloop.c (verify_loop_structure): Fix up grammar.
26058 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
26060 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
26061 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
26063 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26066 * config/arm/marvell-pj4.md: Update copyright year.
26067 Fix up use of alu to alu_reg and simple_alu_imm.
26069 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
26071 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
26073 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
26076 * lra-constraints.c (curr_insn_transform): Don't reuse original
26077 insn for secondary memory move when memory mode should be different.
26079 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26081 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
26082 atomic_storedi_1): New patterns.
26084 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26086 btver2 pipeline descriptions.
26087 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
26089 * config/i386/i386.md (btver2_decode): New type attributes.
26090 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
26092 * config/i386/btver2.md: New file describing btver2 pipelines.
26094 2013-01-19 Andrew Pinski <apinski@cavium.com>
26096 PR tree-optimization/52631
26097 * tree-ssa-sccvn (visit_use): Before looking up the original
26098 statement, try looking up the simplified expression.
26100 2013-01-19 Anthony Green <green@moxielogic.com>
26102 * config/moxie/moxie.c (moxie_expand_prologue): Set
26103 current_function_static_stack_size.
26105 2013-01-18 Jakub Jelinek <jakub@redhat.com>
26107 PR tree-optimization/56029
26108 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
26109 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
26111 2013-01-18 Sharad Singhai <singhai@google.com>
26113 PR tree-optimization/55995
26114 * dumpfile.c (dump_loc): Print location only if available.
26115 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
26117 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
26120 * lra-constraints.c (curr_insn_transform): Reuse original insn for
26121 secondary memory move.
26122 (inherit_reload_reg): Use rclass instead of cl for
26123 check_secondary_memory_needed_p.
26125 2013-01-18 Jakub Jelinek <jakub@redhat.com>
26127 PR middle-end/56015
26128 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
26129 the case where writing real complex part of target modifies op1.
26131 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
26133 * config/aarch64/aarch64-simd.md
26134 (aarch64_vcond_internal<mode>): Handle unordered cases.
26135 * config/aarch64/iterators.md (v_cmp_result): New.
26137 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
26138 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26140 * config/arm/marvell-pj4.md: New file.
26141 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
26142 * config/arm/arm.md (generic_sched): Add marvell_pj4.
26143 (generic_vfp): Likewise.
26144 * config/arm/arm-cores.def: Add marvell-pj4.
26145 * config/arm/arm-tune.md: Regenerate.
26146 * config/arm/arm-tables.opt: Regenerate.
26147 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
26148 * doc/invoke.texi: Document marvell-pj4.
26150 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
26152 * config/aarch64/arm_neon.h: Map scalar types to standard types.
26154 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26159 * var-tracking.c (negative_power_of_two_p): New.
26160 (global_get_addr_cache, local_get_addr_cache): New.
26161 (get_addr_from_global_cache, get_addr_from_local_cache): New.
26162 (vt_canonicalize_addr): Rewrite using the above. Adjust the
26164 (vt_stack_offset_p): Remove.
26165 (vt_canon_true_dep): Always canonicalize loc's address.
26166 (clobber_overlapping_mems): Make sure we have a MEM.
26167 (local_get_addr_clear_given_value): New.
26168 (val_reset): Clear local cached entries.
26169 (compute_bb_dataflow): Create and release the local cache.
26170 Disable duplicate MEMs clobbering.
26171 (emit_notes_in_bb): Clobber MEMs likewise.
26172 (vt_emit_notes): Create and release the local cache.
26173 (vt_initialize, vt_finalize): Create and release the global
26174 cache, respectively.
26175 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
26177 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26179 PR libmudflap/53359
26180 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
26181 not found in the symtab.
26183 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26186 PR rtl-optimization/55547
26187 PR rtl-optimization/53827
26190 * alias.c (offset_overlap_p): New, factored out of...
26191 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
26192 the conservative special case for symbolic constants. Don't
26193 adjust zero sizes on alignment.
26195 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
26197 PR rtl-optimization/52573
26198 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
26199 REG_UNUSED for the same register.
26201 2013-01-17 Richard Biener <rguenther@suse.de>
26202 Marek Polacek <polacek@redhat.com>
26204 PR rtl-optimization/55833
26205 * loop-unswitch.c (unswitch_loops): Move loop verification...
26206 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
26207 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
26208 Set it to true when we're removing a loop from hierarchy tree in
26209 an irreducible region.
26210 (fix_bb_placements): Adjust caller.
26211 (fix_loop_placements): Likewise.
26213 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
26215 * config/avr/builtins.def (DEF_BUILTIN): Factor out
26216 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
26217 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
26218 Remove ID. Adjust comments.
26219 * config/avr/avr-c.c (avr_builtin_name): Remove.
26220 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
26221 * config/avr/avr.c (avr_tolower): New static function.
26222 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
26223 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
26224 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
26227 2013-01-17 Jan Hubicka <jh@suse.cz>
26229 PR tree-optimization/55273
26230 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
26232 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
26235 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
26236 store through atomic_store<mode>_1.
26237 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
26239 2013-01-17 Martin Jambor <mjambor@suse.cz>
26241 PR tree-optimizations/55264
26242 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
26243 for virtual methods.
26244 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
26245 virtual methods before inlining is over.
26246 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
26248 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
26251 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
26253 PR rtl-optimization/56005
26254 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
26255 pending reads for prefetch.
26257 2013-01-16 Ian Bolton <ian.bolton@arm.com>
26259 * config/aarch64/aarch64.md
26260 (*cstoresi_neg_uxtw): New pattern.
26261 (*cmovsi_insn_uxtw): New pattern.
26262 (*<optab>si3_uxtw): New pattern.
26263 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
26264 (*<optab>si3_insn_uxtw): New pattern.
26265 (*bswapsi2_uxtw): New pattern.
26267 2013-01-16 Richard Biener <rguenther@suse.de>
26269 * tree-inline.c (tree_function_versioning): Remove set but
26270 never used variable.
26272 2013-01-16 Richard Biener <rguenther@suse.de>
26274 PR tree-optimization/55964
26275 * tree-flow.h (rename_variables_in_loop): Remove.
26276 (rename_variables_in_bb): Likewise.
26277 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
26278 (copy_loop_before): Adjust and delete update-ssa status.
26279 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
26280 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
26281 (rename_variables_in_loop): Remove.
26282 (slpeel_update_phis_for_duplicate_loop): Likewise.
26283 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
26284 use available cfg machinery instead of duplicating it.
26285 Update PHI nodes and perform poor-mans SSA update here.
26286 (slpeel_tree_peel_loop_to_edge): Adjust.
26288 2013-01-16 Richard Biener <rguenther@suse.de>
26290 PR tree-optimization/54767
26291 PR tree-optimization/53465
26292 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
26293 (vrp_visit_phi_node): For PHI arguments coming via backedges
26294 drop all symbolical range information.
26295 (execute_vrp): Compute backedges.
26297 2013-01-16 Richard Biener <rguenther@suse.de>
26299 * doc/install.texi: Update CLooG and ISL requirements to
26302 2013-01-16 Christian Bruel <christian.bruel@st.com>
26305 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
26306 (broken_move): Handle UNSPECV_SP_SWITCH_B.
26307 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
26309 2013-01-16 DJ Delorie <dj@redhat.com>
26311 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
26312 (UNSPECV_SP_SWITCH_E): New.
26313 (sp_switch_1): Change to an unspec.
26314 (sp_switch_2): Change to an unspec. Don't use post-inc when we
26317 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
26319 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
26320 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
26321 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
26322 (expand_mem_thread_fence): Ditto.
26323 (expand_mem_signal_fence): Ditto.
26324 (expand_atomic_load): Ditto.
26325 (expand_atomic_store): Ditto.
26327 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
26329 PR rtl-optimization/55547
26330 PR rtl-optimization/53827
26333 * alias.c (memrefs_conflict_p): Set sizes to negative after
26336 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26339 * function.c (thread_prologue_and_epilogue_insns): Always
26340 add crtl->drap_reg to set_up_by_prologue.set, even if
26341 stack_realign_drap is false.
26343 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
26345 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
26346 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
26347 *call): Fix indention.
26349 2013-01-15 Tom de Vries <tom@codesourcery.com>
26352 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
26355 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
26357 PR rtl-optimization/55153
26358 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
26360 2013-01-15 Martin Jambor <mjambor@suse.cz>
26362 PR tree-optimization/55920
26363 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
26364 accesses as grp_to_be_debug_replaced.
26366 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26368 PR tree-optimization/55920
26369 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
26370 there is non-useless type conversion needed from debug rhs to lhs,
26371 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
26373 2013-01-15 Joseph Myers <joseph@codesourcery.com>
26374 Mikael Pettersson <mikpe@it.uu.se>
26377 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
26379 (ASM_OUTPUT_CASE_LABEL): Remove.
26380 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
26381 * final.c (shorten_branches): Update alignment of labels before
26382 jump tables if CASE_VECTOR_SHORTEN_MODE.
26384 2013-01-15 Richard Biener <rguenther@suse.de>
26387 * system.h: Do not include gmp.h for building host tools.
26389 2013-01-15 Richard Biener <rguenther@suse.de>
26391 PR middle-end/55882
26392 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
26393 account for bitpos when computing alignment.
26395 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
26397 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
26398 (ix86_target_macros_internal): Likewise.
26400 * config/i386/i386.c (m_CORE2I7): Removed.
26401 (m_CORE_HASWELL): New macro.
26402 (m_CORE_ALL): Likewise.
26403 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
26404 (initial_ix86_arch_features): Likewise.
26405 (processor_target_table): Initializations for Core avx2.
26406 (cpu_names): New names "core-avx2".
26407 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
26408 PROCESSOR_CORE_HASWELL.
26409 (ix86_issue_rate): New case.
26410 (ia32_multipass_dfa_lookahead): Likewise.
26411 (ix86_sched_init_global): Likewise.
26413 * config/i386/i386.h (TARGET_HASWELL): New macro.
26414 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
26415 (processor_type): New PROCESSOR_HASWELL.
26417 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26419 PR tree-optimization/55955
26420 * tree-vect-loop.c (vectorizable_reduction): Give up early on
26421 *SHIFT_EXPR and *ROTATE_EXPR codes.
26423 PR tree-optimization/48766
26424 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
26425 -ftrapv disable -fwrapv.
26427 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26430 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
26431 etc. to 1 and not to __flash.
26432 Use LL suffix for __INT24_MAX__ with -mint8.
26433 Use ULL suffix for __UINT24_MAX__ with -mint8.
26435 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26437 * config/avr/avr-arch.h
26438 (struct base_arch_s): Use typedef avr_arch_t instead.
26439 (struct arch_info_s): Use typedef avr_arch_info_t instead.
26440 (struct mcu_type_s): Use typedef avr_mcu_t instead.
26441 * config/avr/avr.c: Same.
26442 * config/avr/avr-devices.c: Same.
26443 * config/avr/driver-avr.c: Same.
26444 * config/avr/gen-avr-mmcu-texi.c: Same.
26445 * config/avr/avr-mcus.def: Adjust comment.
26447 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
26449 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
26450 * config/aarch64/iterators.md (VALLDI): New.
26452 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
26453 Andi Kleen <ak@linux.intel.com>
26456 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
26457 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
26460 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26462 * config/avr/avr-stdint.h: Remove trailing blanks.
26463 * config/avr/avr-log.h: Same.
26464 * config/avr/avr-arch.h: Same.
26465 * config/avr/avr-devices.c: Same.
26466 * config/avr/avr-dimode.md: Same.
26467 * config/avr/predicates.md: Same.
26468 * config/avr/avr-c.c: Same. And fix typo.
26470 * config/avr/avr-protos.h: Same. And:
26471 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
26472 (init_cumulative_args): Rename to avr_init_cumulative_args.
26473 (expand_prologue): Rename to avr_expand_prologue.
26474 (expand_epilogue): Rename to avr_expand_epilogue.
26475 (adjust_insn_length): Rename to avr_adjust_insn_length.
26476 (notice_update_cc): Rename to avr_notice_update_cc.
26477 (final_prescan_insn): Rename to avr_final_prescan_insn.
26478 * config/avr/avr.c: Same.
26479 * config/avr/avr.h: Same.
26480 * config/avr/avr.md: Remove trailing blanks.
26481 (prologue): Use avr_expand_prologue.
26482 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
26484 2013-01-14 Richard Biener <rguenther@suse.de>
26486 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
26487 verify_location, collect_subblocks): New functions.
26488 (verify_gimple_in_cfg): Verify that locations only reference
26489 BLOCKs in the functions BLOCK tree.
26491 2013-01-14 Richard Biener <rguenther@suse.de>
26493 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
26495 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
26497 (insert_out_of_ssa_copy_on_edge): Likewise.
26498 (rewrite_close_phi_out_of_ssa): Likewise.
26499 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
26501 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
26502 propagated constants.
26503 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
26506 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26508 * config/avr/avr-modes.def: Add GPL copyright notice.
26510 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
26512 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
26513 MEMMODEL_MASK to determine memory model.
26514 (atomic_store<mode>): Ditto from operands[2].
26515 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
26517 2013-01-13 Jakub Jelinek <jakub@redhat.com>
26520 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
26521 (fold_gimple_assign): Don't call unshare_expr here.
26522 (fold_ctor_reference): Call unshare_expr.
26524 2013-01-13 Terry Guo <terry.guo@arm.com>
26526 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
26527 * doc/fragments.texi: Document MULTILIB_REUSE.
26528 * gcc.c (multilib_reuse): New internal spec.
26529 (set_multilib_dir): Also search multilib from multilib_reuse.
26530 * genmultilib (tmpmultilib3): Refactor code.
26531 (tmpmultilib4): Ditto.
26532 (multilib_reuse): New multilib argument.
26534 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
26536 * Makefile.in: Update copyright.
26538 2013-01-12 Tom de Vries <tom@codesourcery.com>
26540 PR middle-end/55890
26541 * calls.c (expand_call): Check if arg_nr is valid.
26543 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
26545 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
26546 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
26547 documentation. Add missing '__' in front of
26548 __builtin_ia32_packssdw256.
26550 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26553 * config/s390/s390.c (s390_preferred_reload_class): Do not return
26554 NO_REGS for larl operands.
26555 (s390_reload_larl_operand): Use s390_load_address instead of
26558 2013-01-11 Richard Biener <rguenther@suse.de>
26560 * tree-cfg.c (verify_node_sharing_1): Split out from ...
26561 (verify_node_sharing): ... here.
26562 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
26564 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
26566 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
26567 Substitute TREECHECKING.
26568 * configure: Regenerate.
26569 * Makefile.in (TREECHECKING): New.
26571 2013-01-11 Richard Guenther <rguenther@suse.de>
26573 PR tree-optimization/44061
26574 * tree-vrp.c (extract_range_basic): Compute zero as
26575 value-range for __builtin_constant_p of function parameters.
26577 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
26579 Update copyright years.
26581 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
26583 PR rtl-optimization/55672
26584 * lra-eliminations.c (mark_not_eliminable): Permit addition with
26585 const to be eliminable.
26587 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
26589 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
26590 * configure: Regenerate.
26592 2013-01-10 Richard Biener <rguenther@suse.de>
26594 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
26596 2013-01-10 Richard Biener <rguenther@suse.de>
26599 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
26600 locations for virtual PHI arguments.
26601 (rewrite_update_phi_arguments): Likewise.
26603 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
26605 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
26608 2013-01-10 Jakub Jelinek <jakub@redhat.com>
26610 PR tree-optimization/55921
26611 * tree-complex.c (expand_complex_asm): New function.
26612 (expand_complex_operations_1): Call it for GIMPLE_ASM.
26614 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26617 * config/s390/s390.c (s390_symref_operand_p)
26618 (s390_loadrelative_operand_p): Merge the two functions.
26619 (s390_check_qrst_address, print_operand_address): Add parameters
26620 to s390_loadrelative_operand_p invokation.
26621 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
26622 (s390_reload_larl_operand, s390_secondary_reload): Use
26623 s390_loadrelative_operand_p instead of s390_symref_operand_p.
26624 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
26626 2013-01-09 Mike Stump <mikestump@comcast.net>
26628 * dse.c (record_store): Remove unnecessary assert.
26630 2013-01-09 Jan Hubicka <jh@suse.cz>
26632 PR tree-optimization/55569
26633 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
26634 * cfgloop.h (scale_loop_profile): Likewise.
26636 2013-01-09 Jan Hubicka <jh@suse.cz>
26639 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
26641 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
26643 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
26645 PR middle-end/55114
26646 * expr.h (maybe_emit_group_store): Declare.
26647 * expr.c (maybe_emit_group_store): New function.
26648 * builtins.c (expand_builtin_int_roundingfn): Call it.
26649 (expand_builtin_int_roundingfn_2): Likewise.
26651 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
26653 PR rtl-optimization/55829
26654 * lra-constraints.c (match_reload): Add code for absent output.
26655 (curr_insn_transform): Add code for reloads of matched inputs
26658 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
26660 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
26661 attribute of movddup insn to DF.
26662 (*vec_interleave_lowv2df): Ditto.
26663 (vec_dupv2df): Ditto.
26665 2013-01-09 Jan Hubicka <jh@suse.cz>
26667 PR tree-optimiation/55875
26668 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
26669 EVERY_ITERATION parameter.
26670 (number_of_iterations_exit): Check if exit is executed every iteration.
26671 (idx_infer_loop_bounds): Similarly here.
26672 (n_of_executions_at_most): Simplify
26673 to only test for cases where statement is dominated by the
26674 particular bound; handle correctly the "postdominance" test.
26675 (scev_probably_wraps_p): Use max loop iterations info
26676 as a global bound first.
26678 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
26679 Nick Clifton <nickc@redhat.com>
26681 * config/v850/v850.md (cbranchsf4): New pattern.
26682 (cstoresf4): New pattern.
26683 (cbranchdf4): New pattern.
26684 (cstoredf4): New pattern.
26685 (movsicc): Disallow floating point comparisons.
26686 (cmpsf_le_insn): Fix order of operators.
26687 (cmpsf_lt_insn): Likewise.
26688 (cmpsf_eq_insn): Likewise.
26689 (cmpdf_le_insn): Likewise.
26690 (cmpdf_lt_insn): Likewise.
26691 (cmpdf_eq_insn): Likewise.
26692 (cmpsf_ge_insn): Use LE comparison.
26693 (cmpdf_ge_insn): Likewise.
26694 (cmpsf_gt_insn): Use LT comparison.
26695 (cmpdf_gt_insn): Likewise.
26696 (cmpsf_ne_insn): Delete pattern.
26697 (cmpdf_ne_insn): Delete pattern.
26698 * config/v850/v850.c (v850_gen_float_compare): Use
26699 gen_cmpdf_eq_insn for NE comparison.
26700 (v850_float_z_comparison_operator)
26701 (v850_float_nz_comparison_operator): Move from here ...
26702 * config/v850/predicates.md: ... to here. Move GT and GE
26703 comparisons into v850_float_z_comparison_operator.
26704 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
26706 (v850_float_nz_comparison_operator): Likewise.
26708 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26710 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
26711 with calls to gen_insvsi/gen_insvdi.
26713 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26715 * config/i386/i386.c (initial_ix86_tune_features): Set up
26716 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
26718 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
26719 Jakub Jelinek <jakub@redhat.com>
26721 PR tree-optimization/48189
26722 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
26723 If nitercst is 0, don't predict the exit edge.
26725 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26727 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
26728 in asm_fprintf with reg_names.
26729 (aarch64_print_operand_address): Likewise.
26730 (aarch64_return_addr): Likewise.
26731 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
26733 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26735 * config/pa/pa.h (VAL_U6_BITS_P): Define.
26736 (INT_U6_BITS): Likewise.
26737 * config/pa/predicates.md (uint6_operand): New predicate.
26738 (shift5_operand, shift6_operand): Likewise.
26739 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
26741 (lshrdi3): Use shift6_operand.
26742 (shrpsi4, shrpdi4): New insn patterns.
26743 (extzv): Delete expander.
26744 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
26745 predicates in unamed zero extract patterns. Tighten common constraint.
26746 (extv): Delete expander.
26747 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
26748 predicates in unamed sign extract patterns. Tighten common constraint.
26749 (insv): Delete expander.
26750 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
26751 predicates in unamed insert patterns. Tighten common constraint.
26752 Change uint32_operand predicate to uint6_operand predicate in unamed
26753 DImode pattern to insert constant values of type 1...1xxxx.
26755 2013-01-04 Jan Hubicka <jh@suse.cz>
26757 PR tree-optimization/55823
26758 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
26761 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26762 Uros Bizjak <ubizjak@gmail.com>
26764 PR rtl-optimization/55845
26765 * df-problems.c (can_move_insns_across): Stop scanning at
26766 volatile_insn_p source instruction or give up if
26767 across_from .. across_to range contains any volatile_insn_p
26770 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26772 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
26773 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
26775 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
26776 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
26778 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26781 * asan.c (asan_clear_shadow): New function.
26782 (asan_emit_stack_protection): Use it.
26784 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26786 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
26787 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
26788 with tab instead of space.
26790 2013-01-08 Nick Clifton <nickc@redhat.com>
26792 * config/rl78/rl78.c (rl78_expand_prologue): Always select
26793 register bank 0 at the start of an interrupt handler.
26794 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
26797 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26799 * config/aarch64/aarch64-simd.md
26800 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
26801 (aarch64_simd_bsl): Likewise.
26802 (aarch64_vcond_internal<mode>): Likewise.
26803 (vcond<mode><mode>): Likewise.
26804 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
26805 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
26807 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26809 * config/aarch64/aarch64-builtins.c
26810 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
26812 2013-01-08 Martin Jambor <mjambor@suse.cz>
26815 * tree-sra.c (analyze_access_subtree): Return true also after
26816 potentially creating a debug-only replacement.
26818 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26820 PR middle-end/55890
26821 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
26823 PR tree-optimization/54120
26824 * tree-vrp.c (range_fits_type_p): Don't allow
26825 src_precision < precision from signed vr to unsigned_p
26826 if vr->min or vr->max is negative.
26827 (simplify_float_conversion_using_ranges): Test can_float_p
26828 against CODE_FOR_nothing.
26830 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26831 Richard Biener <rguenther@suse.de>
26833 PR middle-end/55851
26834 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
26835 types instead of just INTEGER_TYPE types.
26837 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
26839 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
26842 2013-01-07 Steve Ellcey <sellcey@mips.com>
26845 * config/mips/mips.opt: Change mad to mmad to match documentation.
26847 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26850 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
26851 .progmemx.data now.
26853 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26856 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
26857 (avr_addrspace_t): Add .section_name field.
26858 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
26860 (avr_addrspace): Same. Initialize .section_name. Remove last
26861 NULL entry. Put __memx into .progmemx.data.
26862 (progmem_section_prefix): Remove.
26863 (avr_asm_init_sections): No need to initialize progmem_section.
26864 (avr_asm_named_section): Use avr_addrspace[].section_name to get
26865 section name prefix.
26866 (avr_asm_select_section): Ditto. And use get_unnamed_section to
26867 retrieve the progmem section.
26868 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
26869 boundary to run over avr_addrspace[].
26870 (avr_register_target_pragmas): Ditto.
26872 2013-01-06 Jakub Jelinek <jakub@redhat.com>
26874 * varasm.c (output_constant_def_contents): For asan_protect_global
26875 protected strings, adjust DECL_ALIGN if needed, before testing for
26877 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
26878 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
26880 (output_object_block): For asan protected decls, emit asan padding
26881 after their contents.
26882 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
26883 (asan_finish_file): Test it here instead.
26885 2013-01-07 Nick Clifton <nickc@redhat.com>
26886 Matthias Klose <doko@debian.org>
26887 Doug Kwan <dougkwan@google.com>
26888 H.J. Lu <hongjiu.lu@intel.com>
26891 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
26893 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
26895 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
26897 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
26899 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
26901 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26904 * doc/install.texi (Cross-Compiler-Specific Options): Document
26907 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
26909 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
26910 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
26911 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
26912 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
26913 vqmovun_high_s64): Fix source operand number and update copyright.
26915 2013-01-07 Richard Biener <rguenther@suse.de>
26917 PR middle-end/55890
26918 * gimple.h (gimple_call_builtin_p): New overload.
26919 * gimple.c (validate_call): New function.
26920 (gimple_call_builtin_p): Likewise.
26921 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26922 Use gimple_call_builtin_p.
26923 (find_func_clobbers): Likewise.
26924 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
26925 (strlen_optimize_stmt): Likewise.
26927 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26929 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
26930 (vld1q_dup_*): Likewise.
26931 (vld1_*): Likewise.
26932 (vld1q_*): Likewise.
26933 (vld1_lane_*): Likewise.
26934 (vld1q_lane_*): Likewise.
26936 2013-01-07 Richard Biener <rguenther@suse.de>
26938 * lto-streamer.h (LTO_minor_version): Bump to 2.
26940 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26942 * config/aarch64/aarch64-protos.h
26943 (aarch64_const_double_zero_rtx_p): Rename to...
26944 (aarch64_float_const_zero_rtx_p): ...this.
26945 (aarch64_float_const_representable_p): New.
26946 (aarch64_output_simd_mov_immediate): Likewise.
26947 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
26948 move immediate case.
26949 * config/aarch64/aarch64.c
26950 (aarch64_const_double_zero_rtx_p): Rename to...
26951 (aarch64_float_const_zero_rtx_p): ...this.
26952 (aarch64_print_operand): Allow printing of new constants.
26953 (aarch64_valid_floating_const): New.
26954 (aarch64_legitimate_constant_p): Check for valid floating-point
26956 (aarch64_simd_valid_immediate): Likewise.
26957 (aarch64_vect_float_const_representable_p): New.
26958 (aarch64_float_const_representable_p): Likewise.
26959 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
26960 (aarch64_output_simd_mov_immediate): New.
26961 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
26962 (*movdf_aarch64): Likewise.
26963 * config/aarch64/constraints.md (Ufc): New.
26964 (Y): call aarch64_float_const_zero_rtx.
26965 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
26967 2013-01-07 Richard Biener <rguenther@suse.de>
26969 PR tree-optimization/55888
26970 PR tree-optimization/55862
26971 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
26972 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
26973 not if it is contained therein.
26975 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26977 * config/avr/t-avr: Typo.
26979 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26982 * config/avr/t-avr: Don't automatically rebuild
26983 $(srcdir)/config/avr/t-multilib
26984 $(srcdir)/config/avr/avr-tables.opt
26985 $(srcdir)/doc/avr-mmcu.texi
26986 (avr-mcus): New phony target to build them on request.
26987 (s-avr-mlib, s-avr-mmcu-texi): Remove.
26988 * avr/avr-mcus.def: Adjust comments.
26990 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
26992 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
26994 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26996 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
26998 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
27000 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
27002 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
27004 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
27005 to generate profiling.
27006 * config/rs6000/aix64.h (LIB_SPEC): Same.
27008 2013-01-04 Andrew Pinski <apinski@cavium.com>
27010 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
27012 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
27014 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
27016 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
27018 (ix86_expand_move): Ditto.
27019 (ix86_zero_extend_to_Pmode): Ditto.
27020 (ix86_expand_call): Ditto.
27021 (ix86_expand_special_args_builtin): Ditto.
27022 (ix86_expand_builtin): Ditto.
27024 2013-01-04 Richard Biener <rguenther@suse.de>
27026 PR tree-optimization/55862
27027 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
27028 translating them through PHI nodes.
27030 2013-01-04 Martin Jambor <mjambor@suse.cz>
27032 PR tree-optimization/55755
27033 * tree-sra.c (sra_modify_assign): Do not check that an access has no
27034 children when trying to avoid producing a VIEW_CONVERT_EXPR.
27036 2013-01-04 Marek Polacek <polacek@redhat.com>
27038 PR middle-end/55859
27039 * opts.c (default_options_optimization): Clarify error message.
27041 2013-01-04 Richard Biener <rguenther@suse.de>
27043 PR middle-end/55863
27044 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
27047 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27050 * config/pa/pa.md (movsi): Revert previous change.
27051 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
27054 2013-01-03 Richard Henderson <rth@redhat.com>
27056 * config/i386/i386.c (ix86_expand_move): Always assign to op1
27057 after eliminating TLS symbols.
27059 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
27062 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
27063 * graphite-poly.c (debug_gmp_value): Likewise.
27065 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
27068 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
27069 selected code model, define __code_mode_small__, __code_model_medium__,
27070 __code_model_large__, __code_model_32__ or __code_model_kernel__.
27071 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
27072 xchg temporary register with %k. Declare temporary register as
27074 [__x86_64__]: For medium and large code models, preserve %rbx register.
27076 2013-01-03 Richard Biener <rguenther@suse.de>
27078 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
27079 (dump_subscript): Adjust.
27080 (finalize_ddr_dependent): Do not dump redundant info.
27081 (analyze_siv_subscript): Adjust.
27082 (subscript_dependence_tester): Likewise.
27083 (compute_affine_dependence): Likewise.
27085 2013-01-03 Richard Biener <rguenther@suse.de>
27088 2013-01-03 Richard Biener <rguenther@suse.de>
27090 PR tree-optimization/55857
27091 * tree-vect-stmts.c (vectorizable_load): Do not setup
27092 re-alignment for invariant loads.
27094 2013-01-02 Richard Biener <rguenther@suse.de>
27096 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
27097 invariant load do not generate a vector load from the scalar location.
27099 2013-01-03 Richard Biener <rguenther@suse.de>
27101 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
27102 for not vectorizing.
27103 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
27104 not build INDIRECT_REFs, call get_name once only.
27105 (vect_create_data_ref_ptr): Likewise. Dump base object kind
27106 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
27108 2013-01-03 Richard Biener <rguenther@suse.de>
27110 PR tree-optimization/55857
27111 * tree-vect-stmts.c (vectorizable_load): Do not setup
27112 re-alignment for invariant loads.
27114 2013-01-03 Richard Biener <rguenther@suse.de>
27117 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
27118 prefer a built-in decl.
27120 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27122 * gcc.c (process_command): Update copyright notice dates.
27123 * gcov.c (print_version): Likewise.
27124 * gcov-dump.c (print_version): Likewise.
27126 PR rtl-optimization/55838
27127 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
27128 iv0.step, iv1.step and step.
27130 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27131 Marc Glisse <marc.glisse@inria.fr>
27133 PR tree-optimization/55832
27134 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
27135 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
27136 integer_{one,zero}_node.
27138 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27141 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
27142 * var-tracking.c (reverse_op): Don't add reverse ops to
27143 VALUEs that have already
27144 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
27146 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
27148 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
27150 2013-01-02 Teresa Johnson <tejohnson@google.com>
27152 * dumpfile.c (dump_loc): Print filename with location.
27153 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
27154 new location_t parameter to emit complete unroll message with
27155 new dump framework.
27156 (canonicalize_loop_induction_variables): Compute loops location
27157 and pass to try_unroll_loop_completely.
27158 * loop-unroll.c (report_unroll_peel): New function.
27159 (peel_loops_completely): Use new dump format with location
27160 for main dumpfile message, and invoke report_unroll_peel on success.
27161 (decide_unrolling_and_peeling): Ditto.
27162 (decide_peel_once_rolling): Remove old dumpfile message subsumed
27163 by report_unroll_peel.
27164 (decide_peel_completely): Ditto.
27165 (decide_unroll_constant_iterations): Ditto.
27166 (decide_unroll_runtime_iterations): Ditto.
27167 (decide_peel_simple): Ditto.
27168 (decide_unroll_stupid): Ditto.
27169 * cfgloop.c (get_loop_location): New function.
27170 * cfgloop.h (get_loop_location): Declare.
27172 2013-01-02 Sriraman Tallam <tmsriram@google.com>
27174 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
27177 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27179 PR middle-end/55198
27180 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
27181 BLKmode objects when EXPAND_MEMORY is specified.
27183 2013-01-02 Sriraman Tallam <tmsriram@google.com>
27185 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
27187 (fold_builtin_cpu): Do not share cpu model decls across statements.
27189 2013-01-02 Jason Merrill <jason@redhat.com>
27192 * tree.c (build_array_type_1): Revert earlier change.
27194 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
27196 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
27198 * config/aarch64/aarch64-tune.md: Re-generate.
27200 2013-01-02 Richard Biener <rguenther@suse.de>
27202 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
27203 invariant load do not generate a vector load from the scalar location.
27205 2013-01-02 Richard Biener <rguenther@suse.de>
27208 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
27209 * configure: Regenerate.
27211 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27213 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
27214 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
27215 (expand_builtin_int_roundingfn_2): Keep the original target around
27216 for the fallback case.
27218 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27220 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
27221 to be clear for sign changes.
27223 2013-01-01 Jan Hubicka <jh@suse.cz>
27225 * ipa-inline-analysis.c: Fix formatting.
27227 2013-01-01 Jakub Jelinek <jakub@redhat.com>
27229 PR tree-optimization/55831
27230 * tree-vect-loop.c (get_initial_def_for_induction): Use
27231 gsi_after_labels instead of gsi_start_bb.
27233 Copyright (C) 2013 Free Software Foundation, Inc.
27235 Copying and distribution of this file, with or without modification,
27236 are permitted in any medium without royalty provided the copyright
27237 notice and this notice are preserved.